diff options
524 files changed, 16655 insertions, 7540 deletions
diff --git a/.github/CODEOWNERS b/.github/CODEOWNERS new file mode 100644 index 000000000..ab5908649 --- /dev/null +++ b/.github/CODEOWNERS @@ -0,0 +1,18 @@ +# Pull requests concerning the listed files will automatically invite the respective maintainers as reviewers. +# This file is not used for denoting any kind of ownership, but is merely a tool for handling notifications. +# +# Merge permissions are required for maintaining an entry in this file. +# For documentation on this mechanism, see https://help.github.com/articles/about-codeowners/ + +# Default reviewers if nothing else matches +* @edolstra + +# This file +.github/CODEOWNERS @edolstra + +# Public documentation +/doc @fricklerhandwerk +*.md @fricklerhandwerk + +# Libstore layer +/src/libstore @thufschmitt diff --git a/.github/ISSUE_TEMPLATE/bug_report.md b/.github/ISSUE_TEMPLATE/bug_report.md index e6d346bc1..984f9a9ea 100644 --- a/.github/ISSUE_TEMPLATE/bug_report.md +++ b/.github/ISSUE_TEMPLATE/bug_report.md @@ -30,3 +30,7 @@ A clear and concise description of what you expected to happen. **Additional context** Add any other context about the problem here. + +**Priorities** + +Add :+1: to [issues you find important](https://github.com/NixOS/nix/issues?q=is%3Aissue+is%3Aopen+sort%3Areactions-%2B1-desc). diff --git a/.github/ISSUE_TEMPLATE/feature_request.md b/.github/ISSUE_TEMPLATE/feature_request.md index 392ed30c6..42c658b52 100644 --- a/.github/ISSUE_TEMPLATE/feature_request.md +++ b/.github/ISSUE_TEMPLATE/feature_request.md @@ -2,7 +2,7 @@ name: Feature request about: Suggest an idea for this project title: '' -labels: improvement +labels: feature assignees: '' --- @@ -18,3 +18,7 @@ A clear and concise description of any alternative solutions or features you've **Additional context** Add any other context or screenshots about the feature request here. + +**Priorities** + +Add :+1: to [issues you find important](https://github.com/NixOS/nix/issues?q=is%3Aissue+is%3Aopen+sort%3Areactions-%2B1-desc). diff --git a/.github/ISSUE_TEMPLATE/installer.md b/.github/ISSUE_TEMPLATE/installer.md new file mode 100644 index 000000000..3768a49c9 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/installer.md @@ -0,0 +1,36 @@ +--- +name: Installer issue +about: Report problems with installation +title: '' +labels: installer +assignees: '' + +--- + +## Platform + +<!-- select the platform on which you tried to install Nix --> + +- [ ] Linux: <!-- state your distribution, e.g. Arch Linux, Ubuntu, ... --> +- [ ] macOS +- [ ] WSL + +## Additional information + +<!-- state special circumstances on your system or additional steps you have taken prior to installation --> + +## Output + +<details><summary>Output</summary> + +```log + +<!-- paste console output here and remove this comment --> + +``` + +</details> + +## Priorities + +Add :+1: to [issues you find important](https://github.com/NixOS/nix/issues?q=is%3Aissue+is%3Aopen+sort%3Areactions-%2B1-desc). diff --git a/.github/ISSUE_TEMPLATE/missing_documentation.md b/.github/ISSUE_TEMPLATE/missing_documentation.md new file mode 100644 index 000000000..942d7a971 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/missing_documentation.md @@ -0,0 +1,31 @@ +--- +name: Missing or incorrect documentation +about: Help us improve the reference manual +title: '' +labels: documentation +assignees: '' + +--- + +## Problem + +<!-- describe your problem --> + +## Checklist + +<!-- make sure this issue is not redundant or obsolete --> + +- [ ] checked [latest Nix manual] \([source]) +- [ ] checked [open documentation issues and pull requests] for possible duplicates + +[latest Nix manual]: https://nixos.org/manual/nix/unstable/ +[source]: https://github.com/NixOS/nix/tree/master/doc/manual/src +[open documentation issues and pull requests]: https://github.com/NixOS/nix/labels/documentation + +## Proposal + +<!-- propose a solution --> + +## Priorities + +Add :+1: to [issues you find important](https://github.com/NixOS/nix/issues?q=is%3Aissue+is%3Aopen+sort%3Areactions-%2B1-desc). diff --git a/.github/PULL_REQUEST_TEMPLATE.md b/.github/PULL_REQUEST_TEMPLATE.md new file mode 100644 index 000000000..60742cf6a --- /dev/null +++ b/.github/PULL_REQUEST_TEMPLATE.md @@ -0,0 +1,32 @@ +# Motivation +<!-- Briefly explain what the change is about and why it is desirable. --> + +# Context +<!-- Provide context. Reference open issues if available. --> + +<!-- Non-trivial change: Briefly outline the implementation strategy. --> + +<!-- Invasive change: Discuss alternative designs or approaches you considered. --> + +<!-- Large change: Provide instructions to reviewers how to read the diff. --> + +# Checklist for maintainers + +<!-- Contributors: please leave this as is --> + +Maintainers: tick if completed or explain if not relevant + + - [ ] agreed on idea + - [ ] agreed on implementation strategy + - [ ] tests, as appropriate + - functional tests - `tests/**.sh` + - unit tests - `src/*/tests` + - integration tests - `tests/nixos/*` + - [ ] documentation in the manual + - [ ] code and comments are self-explanatory + - [ ] commit message explains why the change was made + - [ ] new feature or incompatible change: updated release notes + +# Priorities + +Add :+1: to [pull requests you find important](https://github.com/NixOS/nix/pulls?q=is%3Aopen+sort%3Areactions-%2B1-desc). diff --git a/.github/PULL_REQUEST_TEMPLATE/pull_request_template.md b/.github/PULL_REQUEST_TEMPLATE/pull_request_template.md deleted file mode 100644 index 537aa0909..000000000 --- a/.github/PULL_REQUEST_TEMPLATE/pull_request_template.md +++ /dev/null @@ -1,7 +0,0 @@ -**Release Notes** -Please include relevant [release notes](https://github.com/NixOS/nix/blob/master/doc/manual/src/release-notes/rl-next.md) as needed. - - -**Testing** - -If this issue is a regression or something that should block release, please consider including a test either in the [testsuite](https://github.com/NixOS/nix/tree/master/tests) or as a [hydraJob]( https://github.com/NixOS/nix/blob/master/flake.nix#L396) so that it can be part of the [automatic checks](https://hydra.nixos.org/jobset/nix/master). diff --git a/.github/labeler.yml b/.github/labeler.yml new file mode 100644 index 000000000..dc502b6d5 --- /dev/null +++ b/.github/labeler.yml @@ -0,0 +1,6 @@ +"documentation": + - doc/manual/* + - src/nix/**/*.md + +"tests": + - tests/**/* diff --git a/.github/workflows/backport.yml b/.github/workflows/backport.yml index 3a2d4de0e..b04723b95 100644 --- a/.github/workflows/backport.yml +++ b/.github/workflows/backport.yml @@ -2,9 +2,15 @@ name: Backport on: pull_request_target: types: [closed, labeled] +permissions: + contents: read jobs: backport: name: Backport Pull Request + permissions: + # for zeebe-io/backport-action + contents: write + pull-requests: write if: github.repository_owner == 'NixOS' && github.event.pull_request.merged == true && (github.event_name != 'labeled' || startsWith('backport', github.event.label.name)) runs-on: ubuntu-latest steps: @@ -15,12 +21,12 @@ jobs: fetch-depth: 0 - name: Create backport PRs # should be kept in sync with `version` - uses: zeebe-io/backport-action@v0.0.8 + uses: zeebe-io/backport-action@v1.2.0 with: # Config README: https://github.com/zeebe-io/backport-action#backport-action github_token: ${{ secrets.GITHUB_TOKEN }} github_workspace: ${{ github.workspace }} pull_description: |- - Bot-based backport to `${target_branch}`, triggered by a label in #${pull_number}. + Automatic backport to `${target_branch}`, triggered by a label in #${pull_number}. # should be kept in sync with `uses` version: v0.0.5 diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 956f81684..158eb057c 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -9,7 +9,7 @@ permissions: read-all jobs: tests: - needs: [check_cachix] + needs: [check_secrets] strategy: matrix: os: [ubuntu-latest, macos-latest] @@ -19,33 +19,37 @@ jobs: - uses: actions/checkout@v3 with: fetch-depth: 0 - - uses: cachix/install-nix-action@v17 + - uses: cachix/install-nix-action@v20 - run: echo CACHIX_NAME="$(echo $GITHUB_REPOSITORY-install-tests | tr "[A-Z]/" "[a-z]-")" >> $GITHUB_ENV - - uses: cachix/cachix-action@v10 - if: needs.check_cachix.outputs.secret == 'true' + - uses: cachix/cachix-action@v12 + if: needs.check_secrets.outputs.cachix == 'true' with: name: '${{ env.CACHIX_NAME }}' signingKey: '${{ secrets.CACHIX_SIGNING_KEY }}' authToken: '${{ secrets.CACHIX_AUTH_TOKEN }}' - run: nix --experimental-features 'nix-command flakes' flake check -L - check_cachix: + check_secrets: permissions: contents: none - name: Cachix secret present for installer tests + name: Check Cachix and Docker secrets present for installer tests runs-on: ubuntu-latest outputs: - secret: ${{ steps.secret.outputs.secret }} + cachix: ${{ steps.secret.outputs.cachix }} + docker: ${{ steps.secret.outputs.docker }} steps: - - name: Check for Cachix secret + - name: Check for secrets id: secret env: _CACHIX_SECRETS: ${{ secrets.CACHIX_SIGNING_KEY }}${{ secrets.CACHIX_AUTH_TOKEN }} - run: echo "::set-output name=secret::${{ env._CACHIX_SECRETS != '' }}" + _DOCKER_SECRETS: ${{ secrets.DOCKERHUB_USERNAME }}${{ secrets.DOCKERHUB_TOKEN }} + run: | + echo "::set-output name=cachix::${{ env._CACHIX_SECRETS != '' }}" + echo "::set-output name=docker::${{ env._DOCKER_SECRETS != '' }}" installer: - needs: [tests, check_cachix] - if: github.event_name == 'push' && needs.check_cachix.outputs.secret == 'true' + needs: [tests, check_secrets] + if: github.event_name == 'push' && needs.check_secrets.outputs.cachix == 'true' runs-on: ubuntu-latest outputs: installerURL: ${{ steps.prepare-installer.outputs.installerURL }} @@ -54,8 +58,10 @@ jobs: with: fetch-depth: 0 - run: echo CACHIX_NAME="$(echo $GITHUB_REPOSITORY-install-tests | tr "[A-Z]/" "[a-z]-")" >> $GITHUB_ENV - - uses: cachix/install-nix-action@v17 - - uses: cachix/cachix-action@v10 + - uses: cachix/install-nix-action@v19 + with: + install_url: https://releases.nixos.org/nix/nix-2.13.3/install + - uses: cachix/cachix-action@v12 with: name: '${{ env.CACHIX_NAME }}' signingKey: '${{ secrets.CACHIX_SIGNING_KEY }}' @@ -64,8 +70,8 @@ jobs: run: scripts/prepare-installer-for-github-actions installer_test: - needs: [installer, check_cachix] - if: github.event_name == 'push' && needs.check_cachix.outputs.secret == 'true' + needs: [installer, check_secrets] + if: github.event_name == 'push' && needs.check_secrets.outputs.cachix == 'true' strategy: matrix: os: [ubuntu-latest, macos-latest] @@ -73,28 +79,38 @@ jobs: steps: - uses: actions/checkout@v3 - run: echo CACHIX_NAME="$(echo $GITHUB_REPOSITORY-install-tests | tr "[A-Z]/" "[a-z]-")" >> $GITHUB_ENV - - uses: cachix/install-nix-action@v17 + - uses: cachix/install-nix-action@v19 with: install_url: '${{needs.installer.outputs.installerURL}}' install_options: "--tarball-url-prefix https://${{ env.CACHIX_NAME }}.cachix.org/serve" - - run: nix-instantiate -E 'builtins.currentTime' --eval + - run: sudo apt install fish zsh + if: matrix.os == 'ubuntu-latest' + - run: brew install fish + if: matrix.os == 'macos-latest' + - run: exec bash -c "nix-instantiate -E 'builtins.currentTime' --eval" + - run: exec sh -c "nix-instantiate -E 'builtins.currentTime' --eval" + - run: exec zsh -c "nix-instantiate -E 'builtins.currentTime' --eval" + - run: exec fish -c "nix-instantiate -E 'builtins.currentTime' --eval" docker_push_image: - needs: [check_cachix, tests] + needs: [check_secrets, tests] if: >- github.event_name == 'push' && github.ref_name == 'master' && - needs.check_cachix.outputs.secret == 'true' + needs.check_secrets.outputs.cachix == 'true' && + needs.check_secrets.outputs.docker == 'true' runs-on: ubuntu-latest steps: - uses: actions/checkout@v3 with: fetch-depth: 0 - - uses: cachix/install-nix-action@v17 + - uses: cachix/install-nix-action@v19 + with: + install_url: https://releases.nixos.org/nix/nix-2.13.3/install - run: echo CACHIX_NAME="$(echo $GITHUB_REPOSITORY-install-tests | tr "[A-Z]/" "[a-z]-")" >> $GITHUB_ENV - run: echo NIX_VERSION="$(nix --experimental-features 'nix-command flakes' eval .\#default.version | tr -d \")" >> $GITHUB_ENV - - uses: cachix/cachix-action@v10 - if: needs.check_cachix.outputs.secret == 'true' + - uses: cachix/cachix-action@v12 + if: needs.check_secrets.outputs.cachix == 'true' with: name: '${{ env.CACHIX_NAME }}' signingKey: '${{ secrets.CACHIX_SIGNING_KEY }}' diff --git a/.github/workflows/labels.yml b/.github/workflows/labels.yml new file mode 100644 index 000000000..5f949ddc5 --- /dev/null +++ b/.github/workflows/labels.yml @@ -0,0 +1,24 @@ +name: "Label PR" + +on: + pull_request_target: + types: [edited, opened, synchronize, reopened] + +# WARNING: +# When extending this action, be aware that $GITHUB_TOKEN allows some write +# access to the GitHub API. This means that it should not evaluate user input in +# a way that allows code injection. + +permissions: + contents: read + pull-requests: write + +jobs: + labels: + runs-on: ubuntu-latest + if: github.repository_owner == 'NixOS' + steps: + - uses: actions/labeler@v4 + with: + repo-token: ${{ secrets.GITHUB_TOKEN }} + sync-labels: true diff --git a/.gitignore b/.gitignore index ba8e95191..e326966d6 100644 --- a/.gitignore +++ b/.gitignore @@ -22,11 +22,13 @@ perl/Makefile.config /doc/manual/src/SUMMARY.md /doc/manual/src/command-ref/new-cli /doc/manual/src/command-ref/conf-file.md -/doc/manual/src/expressions/builtins.md +/doc/manual/src/language/builtins.md # /scripts/ /scripts/nix-profile.sh /scripts/nix-profile-daemon.sh +/scripts/nix-profile.fish +/scripts/nix-profile-daemon.fish # /src/libexpr/ /src/libexpr/lexer-tab.cc @@ -35,14 +37,14 @@ perl/Makefile.config /src/libexpr/parser-tab.hh /src/libexpr/parser-tab.output /src/libexpr/nix.tbl -/src/libexpr/tests/libexpr-tests +/src/libexpr/tests/libnixexpr-tests # /src/libstore/ *.gen.* -/src/libstore/tests/libstore-tests +/src/libstore/tests/libnixstore-tests # /src/libutil/ -/src/libutil/tests/libutil-tests +/src/libutil/tests/libnixutil-tests /src/nix/nix @@ -73,7 +75,7 @@ perl/Makefile.config # /tests/ /tests/test-tmp -/tests/common.sh +/tests/common/vars-and-functions.sh /tests/result* /tests/restricted-innocent /tests/shell @@ -101,6 +103,7 @@ outputs/ *.a *.o +*.o.tmp *.so *.dylib *.dll @@ -1 +1 @@ -2.10.0
\ No newline at end of file +2.15.0
\ No newline at end of file diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 000000000..1b0ecaf36 --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,61 @@ +# Contributing to Nix + +Welcome and thank you for your interest in contributing to Nix! +We appreciate your support. + +Reading and following these guidelines will help us make the contribution process easy and effective for everyone involved. + + +## Report a bug + +1. Check on the [GitHub issue tracker](https://github.com/NixOS/nix/issues) if your bug was already reported. + +2. If you were not able to find the bug or feature [open a new issue](https://github.com/NixOS/nix/issues/new/choose) + +3. The issue templates will guide you in specifying your issue. + The more complete the information you provide, the more likely it can be found by others and the more useful it is in the future. + Make sure reported bugs can be reproduced easily. + +4. Once submitted, do not expect issues to be picked up or solved right away. + The only way to ensure this, is to [work on the issue yourself](#making-changes-to-nix). + +## Report a security vulnerability + +Check out the [security policy](https://github.com/NixOS/nix/security/policy). + +## Making changes to Nix + +1. Check for [pull requests](https://github.com/NixOS/nix/pulls) that might already cover the contribution you are about to make. + There are many open pull requests that might already do what you intent to work on. + You can use [labels](https://github.com/NixOS/nix/labels) to filter for relevant topics. + +2. Search for related issues that cover what you're going to work on. It could help to mention there that you will work on the issue. + +3. Check the [Nix reference manual](https://nixos.org/manual/nix/unstable/contributing/hacking.html) for information on building Nix and running its tests. + + For contributions to the command line interface, please check the [CLI guidelines](https://nixos.org/manual/nix/unstable/contributing/cli-guideline.html). + +4. Make your changes! + +5. [Create a pull request](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/proposing-changes-to-your-work-with-pull-requests/creating-a-pull-request) for your changes. + * [Mark the pull request as draft](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/proposing-changes-to-your-work-with-pull-requests/changing-the-stage-of-a-pull-request) if you're not done with the changes. + * Make sure to have [a clean history of commits on your branch by using rebase](https://www.digitalocean.com/community/tutorials/how-to-rebase-and-update-a-pull-request). + * Link related issues in your pull request to inform interested parties and future contributors about your change. + If your pull request closes one or multiple issues, note that in the description using `Closes: #<number>`, as it will then happen automatically when your change is merged. + +6. Do not expect your pull request to be reviewed immediately. + Nix maintainers follow a [structured process for reviews and design decisions](https://github.com/NixOS/nix/tree/master/maintainers#project-board-protocol), which may or may not prioritise your work. + +7. If you need additional feedback or help to getting pull request into shape, ask other contributors using [@mentions](https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax#mentioning-people-and-teams). + +## Making changes to the Nix manual + +The Nix reference manual is hosted on https://nixos.org/manual/nix. +The underlying source files are located in [`doc/manual/src`](./doc/manual/src). +For small changes you can [use GitHub to edit these files](https://docs.github.com/en/repositories/working-with-files/managing-files/editing-files) +For larger changes see the [Nix reference manual](https://nixos.org/manual/nix/unstable/contributing/hacking.html). + +## Getting help + +Whenever you're stuck or do not know how to proceed, you can always ask for help. +The appropriate channels to do so can be found on the [NixOS Community](https://nixos.org/community/) page. @@ -2,13 +2,10 @@ makefiles = \ mk/precompiled-headers.mk \ local.mk \ src/libutil/local.mk \ - src/libutil/tests/local.mk \ src/libstore/local.mk \ - src/libstore/tests/local.mk \ src/libfetchers/local.mk \ src/libmain/local.mk \ src/libexpr/local.mk \ - src/libexpr/tests/local.mk \ src/libcmd/local.mk \ src/nix/local.mk \ src/resolve-system-dependencies/local.mk \ @@ -20,11 +17,22 @@ makefiles = \ misc/launchd/local.mk \ misc/upstart/local.mk \ doc/manual/local.mk \ - tests/local.mk \ - tests/plugins/local.mk + doc/internal-api/local.mk -include Makefile.config +ifeq ($(tests), yes) +makefiles += \ + src/libutil/tests/local.mk \ + src/libstore/tests/local.mk \ + src/libexpr/tests/local.mk \ + tests/local.mk \ + tests/plugins/local.mk +else +makefiles += \ + mk/disable-tests.mk +endif + OPTIMIZE = 1 ifeq ($(OPTIMIZE), 1) @@ -36,4 +44,4 @@ endif include mk/lib.mk -GLOBAL_CXXFLAGS += -g -Wall -include config.h -std=c++17 -I src +GLOBAL_CXXFLAGS += -g -Wall -include config.h -std=c++2a -I src diff --git a/Makefile.config.in b/Makefile.config.in index 1c5405c6d..707cfe0e3 100644 --- a/Makefile.config.in +++ b/Makefile.config.in @@ -22,6 +22,7 @@ LOWDOWN_LIBS = @LOWDOWN_LIBS@ OPENSSL_LIBS = @OPENSSL_LIBS@ PACKAGE_NAME = @PACKAGE_NAME@ PACKAGE_VERSION = @PACKAGE_VERSION@ +RAPIDCHECK_HEADERS = @RAPIDCHECK_HEADERS@ SHELL = @bash@ SODIUM_LIBS = @SODIUM_LIBS@ SQLITE3_LIBS = @SQLITE3_LIBS@ @@ -45,3 +46,5 @@ sandbox_shell = @sandbox_shell@ storedir = @storedir@ sysconfdir = @sysconfdir@ system = @system@ +tests = @tests@ +internal_api_docs = @internal_api_docs@ @@ -20,8 +20,8 @@ Information on additional installation methods is available on the [Nix download ## Building And Developing -See our [Hacking guide](https://hydra.nixos.org/job/nix/master/build.x86_64-linux/latest/download-by-type/doc/manual/contributing/hacking.html) in our manual for instruction on how to -build nix from source with nix-build or how to get a development environment. +See our [Hacking guide](https://nixos.org/manual/nix/unstable/contributing/hacking.html) in our manual for instruction on how to +to set up a development environment and build Nix from source. ## Additional Resources diff --git a/boehmgc-coroutine-sp-fallback.diff b/boehmgc-coroutine-sp-fallback.diff index e659bf470..2826486fb 100644 --- a/boehmgc-coroutine-sp-fallback.diff +++ b/boehmgc-coroutine-sp-fallback.diff @@ -1,17 +1,49 @@ +diff --git a/darwin_stop_world.c b/darwin_stop_world.c +index 3dbaa3fb..36a1d1f7 100644 +--- a/darwin_stop_world.c ++++ b/darwin_stop_world.c +@@ -352,6 +352,7 @@ GC_INNER void GC_push_all_stacks(void) + int nthreads = 0; + word total_size = 0; + mach_msg_type_number_t listcount = (mach_msg_type_number_t)THREAD_TABLE_SZ; ++ size_t stack_limit; + if (!EXPECT(GC_thr_initialized, TRUE)) + GC_thr_init(); + +@@ -407,6 +408,19 @@ GC_INNER void GC_push_all_stacks(void) + GC_push_all_stack_sections(lo, hi, p->traced_stack_sect); + } + if (altstack_lo) { ++ // When a thread goes into a coroutine, we lose its original sp until ++ // control flow returns to the thread. ++ // While in the coroutine, the sp points outside the thread stack, ++ // so we can detect this and push the entire thread stack instead, ++ // as an approximation. ++ // We assume that the coroutine has similarly added its entire stack. ++ // This could be made accurate by cooperating with the application ++ // via new functions and/or callbacks. ++ stack_limit = pthread_get_stacksize_np(p->id); ++ if (altstack_lo >= altstack_hi || altstack_lo < altstack_hi - stack_limit) { // sp outside stack ++ altstack_lo = altstack_hi - stack_limit; ++ } ++ + total_size += altstack_hi - altstack_lo; + GC_push_all_stack(altstack_lo, altstack_hi); + } diff --git a/pthread_stop_world.c b/pthread_stop_world.c -index 4b2c429..1fb4c52 100644 +index b5d71e62..aed7b0bf 100644 --- a/pthread_stop_world.c +++ b/pthread_stop_world.c -@@ -673,6 +673,8 @@ GC_INNER void GC_push_all_stacks(void) - struct GC_traced_stack_sect_s *traced_stack_sect; - pthread_t self = pthread_self(); - word total_size = 0; +@@ -768,6 +768,8 @@ STATIC void GC_restart_handler(int sig) + /* world is stopped. Should not fail if it isn't. */ + GC_INNER void GC_push_all_stacks(void) + { + size_t stack_limit; + pthread_attr_t pattr; - - if (!EXPECT(GC_thr_initialized, TRUE)) - GC_thr_init(); -@@ -722,6 +724,31 @@ GC_INNER void GC_push_all_stacks(void) + GC_bool found_me = FALSE; + size_t nthreads = 0; + int i; +@@ -851,6 +853,31 @@ GC_INNER void GC_push_all_stacks(void) hi = p->altstack + p->altstack_size; /* FIXME: Need to scan the normal stack too, but how ? */ /* FIXME: Assume stack grows down */ diff --git a/configure.ac b/configure.ac index f0210ab78..f1f45f868 100644 --- a/configure.ac +++ b/configure.ac @@ -41,8 +41,6 @@ AC_DEFINE_UNQUOTED(SYSTEM, ["$system"], [platform identifier ('cpu-os')]) test "$localstatedir" = '${prefix}/var' && localstatedir=/nix/var -CFLAGS= -CXXFLAGS= AC_PROG_CC AC_PROG_CXX AC_PROG_CPP @@ -147,6 +145,18 @@ if test "x$GCC_ATOMIC_BUILTINS_NEED_LIBATOMIC" = xyes; then LDFLAGS="-latomic $LDFLAGS" fi +# Building without tests is useful for bootstrapping with a smaller footprint +# or running the tests in a separate derivation. Otherwise, we do compile and +# run them. +AC_ARG_ENABLE(tests, AS_HELP_STRING([--disable-tests],[Do not build the tests]), + tests=$enableval, tests=yes) +AC_SUBST(tests) + +# Building without API docs is the default as Nix' C++ interfaces are internal and unstable. +AC_ARG_ENABLE(internal_api_docs, AS_HELP_STRING([--enable-internal-api-docs],[Build API docs for Nix's internal unstable C++ interfaces]), + internal_api_docs=$enableval, internal_api_docs=no) +AC_SUBST(internal_api_docs) + # LTO is currently broken with clang for unknown reasons; ld segfaults in the llvm plugin AC_ARG_ENABLE(lto, AS_HELP_STRING([--enable-lto],[Enable LTO (only supported with GCC) [default=no]]), lto=$enableval, lto=no) @@ -177,7 +187,7 @@ fi PKG_CHECK_MODULES([OPENSSL], [libcrypto], [CXXFLAGS="$OPENSSL_CFLAGS $CXXFLAGS"]) -# Checks for libarchive +# Look for libarchive. PKG_CHECK_MODULES([LIBARCHIVE], [libarchive >= 3.1.2], [CXXFLAGS="$LIBARCHIVE_CFLAGS $CXXFLAGS"]) # Workaround until https://github.com/libarchive/libarchive/issues/1446 is fixed if test "$shared" != yes; then @@ -272,10 +282,24 @@ if test "$gc" = yes; then fi +if test "$tests" = yes; then + # Look for gtest. PKG_CHECK_MODULES([GTEST], [gtest_main]) +# Look for rapidcheck. +# No pkg-config yet, https://github.com/emil-e/rapidcheck/issues/302 +AC_LANG_PUSH(C++) +AC_SUBST(RAPIDCHECK_HEADERS) +[CXXFLAGS="-I $RAPIDCHECK_HEADERS $CXXFLAGS"] +AC_CHECK_HEADERS([rapidcheck/gtest.h], [], [], [#include <gtest/gtest.h>]) +dnl No good for C++ libs with mangled symbols +dnl AC_CHECK_LIB([rapidcheck], []) +AC_LANG_POP(C++) + +fi + # Look for nlohmann/json. PKG_CHECK_MODULES([NLOHMANN_JSON], [nlohmann_json >= 3.9]) @@ -296,15 +320,6 @@ AC_CHECK_FUNCS([setresuid setreuid lchown]) AC_CHECK_FUNCS([strsignal posix_fallocate sysconf]) -# This is needed if bzip2 is a static library, and the Nix libraries -# are dynamic. -case "${host_os}" in - darwin*) - LDFLAGS="-all_load $LDFLAGS" - ;; -esac - - AC_ARG_WITH(sandbox-shell, AS_HELP_STRING([--with-sandbox-shell=PATH],[path of a statically-linked shell to use as /bin/sh in sandboxes]), sandbox_shell=$withval) AC_SUBST(sandbox_shell) diff --git a/default.nix b/default.nix index 00ec5b617..2cccff28d 100644 --- a/default.nix +++ b/default.nix @@ -1,3 +1,10 @@ -(import (fetchTarball "https://github.com/edolstra/flake-compat/archive/master.tar.gz") { - src = ./.; -}).defaultNix +(import + ( + let lock = builtins.fromJSON (builtins.readFile ./flake.lock); in + fetchTarball { + url = "https://github.com/edolstra/flake-compat/archive/${lock.nodes.flake-compat.locked.rev}.tar.gz"; + sha256 = lock.nodes.flake-compat.locked.narHash; + } + ) + { src = ./.; } +).defaultNix diff --git a/doc/internal-api/.gitignore b/doc/internal-api/.gitignore new file mode 100644 index 000000000..dab28b6b0 --- /dev/null +++ b/doc/internal-api/.gitignore @@ -0,0 +1,3 @@ +/doxygen.cfg +/html +/latex diff --git a/doc/internal-api/doxygen.cfg.in b/doc/internal-api/doxygen.cfg.in new file mode 100644 index 000000000..8f526536d --- /dev/null +++ b/doc/internal-api/doxygen.cfg.in @@ -0,0 +1,63 @@ +# Doxyfile 1.9.5 + +# The PROJECT_NAME tag is a single word (or a sequence of words surrounded by +# double-quotes, unless you are using Doxywizard) that should identify the +# project for which the documentation is generated. This name is used in the +# title of most generated pages and in a few other places. +# The default value is: My Project. + +PROJECT_NAME = "Nix" + +# The PROJECT_NUMBER tag can be used to enter a project or revision number. This +# could be handy for archiving the generated documentation or if some version +# control system is used. + +PROJECT_NUMBER = @PACKAGE_VERSION@ + +# Using the PROJECT_BRIEF tag one can provide an optional one line description +# for a project that appears at the top of each page and should give viewer a +# quick idea about the purpose of the project. Keep the description short. + +PROJECT_BRIEF = "Nix, the purely functional package manager; unstable internal interfaces" + +# If the GENERATE_LATEX tag is set to YES, doxygen will generate LaTeX output. +# The default value is: YES. + +GENERATE_LATEX = NO + +# The INPUT tag is used to specify the files and/or directories that contain +# documented source files. You may enter file names like myfile.cpp or +# directories like /usr/src/myproject. Separate the files or directories with +# spaces. See also FILE_PATTERNS and EXTENSION_MAPPING +# Note: If this tag is empty the current directory is searched. + +# FIXME Make this list more maintainable somehow. We could maybe generate this +# in the Makefile, but we would need to change how `.in` files are preprocessed +# so they can expand variables despite configure variables. + +INPUT = \ + src/libcmd \ + src/libexpr \ + src/libexpr/flake \ + src/libexpr/tests \ + src/libexpr/tests/value \ + src/libexpr/value \ + src/libfetchers \ + src/libmain \ + src/libstore \ + src/libstore/build \ + src/libstore/builtins \ + src/libstore/tests \ + src/libutil \ + src/libutil/tests \ + src/nix \ + src/nix-env \ + src/nix-store + +# The INCLUDE_PATH tag can be used to specify one or more directories that +# contain include files that are not input files but should be processed by the +# preprocessor. Note that the INCLUDE_PATH is not recursive, so the setting of +# RECURSIVE has no effect here. +# This tag requires that the tag SEARCH_INCLUDES is set to YES. + +INCLUDE_PATH = @RAPIDCHECK_HEADERS@ diff --git a/doc/internal-api/local.mk b/doc/internal-api/local.mk new file mode 100644 index 000000000..890f341b7 --- /dev/null +++ b/doc/internal-api/local.mk @@ -0,0 +1,19 @@ +.PHONY: internal-api-html + +ifeq ($(internal_api_docs), yes) + +$(docdir)/internal-api/html/index.html $(docdir)/internal-api/latex: $(d)/doxygen.cfg + mkdir -p $(docdir)/internal-api + { cat $< ; echo "OUTPUT_DIRECTORY=$(docdir)/internal-api" ; } | doxygen - + +# Generate the HTML API docs for Nix's unstable internal interfaces. +internal-api-html: $(docdir)/internal-api/html/index.html + +else + +# Make a nicer error message +internal-api-html: + @echo "Internal API docs are disabled. Configure with '--enable-internal-api-docs', or avoid calling 'make internal-api-html'." + @exit 1 + +endif diff --git a/doc/manual/book.toml b/doc/manual/book.toml index 5f78a7614..73fb7e75e 100644 --- a/doc/manual/book.toml +++ b/doc/manual/book.toml @@ -1,7 +1,21 @@ +[book] +title = "Nix Reference Manual" + [output.html] additional-css = ["custom.css"] additional-js = ["redirects.js"] +edit-url-template = "https://github.com/NixOS/nix/tree/master/doc/manual/{path}" +git-repository-url = "https://github.com/NixOS/nix" [preprocessor.anchors] renderers = ["html"] command = "jq --from-file doc/manual/anchors.jq" + +[output.linkcheck] +# no Internet during the build (in the sandbox) +follow-web-links = false + +# mdbook-linkcheck does not understand [foo]{#bar} style links, resulting in +# excessive "Potential incomplete link" warnings. No other kind of warning was +# produced at the time of writing. +warning-policy = "ignore" diff --git a/doc/manual/custom.css b/doc/manual/custom.css index 69d48d4a7..b90f5423f 100644 --- a/doc/manual/custom.css +++ b/doc/manual/custom.css @@ -5,3 +5,7 @@ h1:not(:first-of-type) { h2 { margin-top: 1em; } + +.hljs-meta { + user-select: none; +} diff --git a/doc/manual/generate-builtins.nix b/doc/manual/generate-builtins.nix index 6c8b88da2..115bb3f94 100644 --- a/doc/manual/generate-builtins.nix +++ b/doc/manual/generate-builtins.nix @@ -1,16 +1,20 @@ -with builtins; -with import ./utils.nix; +builtinsDump: +let + showBuiltin = name: + let + inherit (builtinsDump.${name}) doc args; + in + '' + <dt id="builtins-${name}"> + <a href="#builtins-${name}"><code>${name} ${listArgs args}</code></a> + </dt> + <dd> -builtins: + ${doc} + + </dd> + ''; + listArgs = args: builtins.concatStringsSep " " (map (s: "<var>${s}</var>") args); +in +with builtins; concatStringsSep "\n" (map showBuiltin (attrNames builtinsDump)) -concatStrings (map - (name: - let builtin = builtins.${name}; in - "<dt id=\"builtins-${name}\"><a href=\"#builtins-${name}\"><code>${name} " - + concatStringsSep " " (map (s: "<var>${s}</var>") builtin.args) - + "</code></a></dt>" - + "<dd>\n\n" - + builtin.doc - + "\n\n</dd>" - ) - (attrNames builtins)) diff --git a/doc/manual/generate-manpage.nix b/doc/manual/generate-manpage.nix index 244cfa0c2..bc19d0303 100644 --- a/doc/manual/generate-manpage.nix +++ b/doc/manual/generate-manpage.nix @@ -1,99 +1,127 @@ -{ command, renderLinks ? false }: +{ toplevel }: with builtins; with import ./utils.nix; let - showCommand = - { command, def, filename }: - '' - **Warning**: This program is **experimental** and its interface is subject to change. - '' - + "# Name\n\n" - + "`${command}` - ${def.description}\n\n" - + "# Synopsis\n\n" - + showSynopsis { inherit command; args = def.args; } - + (if def.commands or {} != {} - then - let - categories = sort (x: y: x.id < y.id) (unique (map (cmd: cmd.category) (attrValues def.commands))); - listCommands = cmds: - concatStrings (map (name: - "* " - + (if renderLinks - then "[`${command} ${name}`](./${appendName filename name}.md)" - else "`${command} ${name}`") - + " - ${cmds.${name}.description}\n") - (attrNames cmds)); - in - "where *subcommand* is one of the following:\n\n" - # FIXME: group by category - + (if length categories > 1 - then - concatStrings (map - (cat: - "**${toString cat.description}:**\n\n" - + listCommands (filterAttrs (n: v: v.category == cat) def.commands) - + "\n" - ) categories) - + "\n" - else - listCommands def.commands - + "\n") - else "") - + (if def ? doc - then def.doc + "\n\n" - else "") - + (let s = showOptions def.flags; in - if s != "" - then "# Options\n\n${s}" - else "") - ; + showCommand = { command, details, filename, toplevel }: + let + + result = '' + > **Warning** \ + > This program is **experimental** and its interface is subject to change. + + # Name + + `${command}` - ${details.description} + + # Synopsis + + ${showSynopsis command details.args} + + ${maybeSubcommands} + + ${maybeDocumentation} + + ${maybeOptions} + ''; + + showSynopsis = command: args: + let + showArgument = arg: "*${arg.label}*" + (if arg ? arity then "" else "..."); + arguments = concatStringsSep " " (map showArgument args); + in '' + `${command}` [*option*...] ${arguments} + ''; + + maybeSubcommands = if details ? commands && details.commands != {} + then '' + where *subcommand* is one of the following: + + ${subcommands} + '' + else ""; + + subcommands = if length categories > 1 + then listCategories + else listSubcommands details.commands; + + categories = sort (x: y: x.id < y.id) (unique (map (cmd: cmd.category) (attrValues details.commands))); + + listCategories = concatStrings (map showCategory categories); + + showCategory = cat: '' + **${toString cat.description}:** + + ${listSubcommands (filterAttrs (n: v: v.category == cat) details.commands)} + ''; + + listSubcommands = cmds: concatStrings (attrValues (mapAttrs showSubcommand cmds)); + + showSubcommand = name: subcmd: '' + * [`${command} ${name}`](./${appendName filename name}.md) - ${subcmd.description} + ''; + + maybeDocumentation = if details ? doc then details.doc else ""; + + maybeOptions = if details.flags == {} then "" else '' + # Options + + ${showOptions details.flags toplevel.flags} + ''; + + showOptions = options: commonOptions: + let + allOptions = options // commonOptions; + showCategory = cat: '' + ${if cat != "" then "**${cat}:**" else ""} + + ${listOptions (filterAttrs (n: v: v.category == cat) allOptions)} + ''; + listOptions = opts: concatStringsSep "\n" (attrValues (mapAttrs showOption opts)); + showOption = name: option: + let + shortName = if option ? shortName then "/ `-${option.shortName}`" else ""; + labels = if option ? labels then (concatStringsSep " " (map (s: "*${s}*") option.labels)) else ""; + in trim '' + - `--${name}` ${shortName} ${labels} + + ${option.description} + ''; + categories = sort builtins.lessThan (unique (map (cmd: cmd.category) (attrValues allOptions))); + in concatStrings (map showCategory categories); + in squash result; appendName = filename: name: (if filename == "nix" then "nix3" else filename) + "-" + name; - showOptions = flags: + processCommand = { command, details, filename, toplevel }: let - categories = sort builtins.lessThan (unique (map (cmd: cmd.category) (attrValues flags))); - in - concatStrings (map - (cat: - (if cat != "" - then "**${cat}:**\n\n" - else "") - + concatStrings - (map (longName: - let - flag = flags.${longName}; - in - " - `--${longName}`" - + (if flag ? shortName then " / `-${flag.shortName}`" else "") - + (if flag ? labels then " " + (concatStringsSep " " (map (s: "*${s}*") flag.labels)) else "") - + " \n" - + " " + flag.description + "\n\n" - ) (attrNames (filterAttrs (n: v: v.category == cat) flags)))) - categories); - - showSynopsis = - { command, args }: - "`${command}` [*option*...] ${concatStringsSep " " - (map (arg: "*${arg.label}*" + (if arg ? arity then "" else "...")) args)}\n\n"; - - processCommand = { command, def, filename }: - [ { name = filename + ".md"; value = showCommand { inherit command def filename; }; inherit command; } ] - ++ concatMap - (name: processCommand { - filename = appendName filename name; - command = command + " " + name; - def = def.commands.${name}; - }) - (attrNames def.commands or {}); - -in + cmd = { + inherit command; + name = filename + ".md"; + value = showCommand { inherit command details filename toplevel; }; + }; + subcommand = subCmd: processCommand { + command = command + " " + subCmd; + details = details.commands.${subCmd}; + filename = appendName filename subCmd; + inherit toplevel; + }; + in [ cmd ] ++ concatMap subcommand (attrNames details.commands or {}); -let - manpages = processCommand { filename = "nix"; command = "nix"; def = builtins.fromJSON command; }; - summary = concatStrings (map (manpage: " - [${manpage.command}](command-ref/new-cli/${manpage.name})\n") manpages); -in -(listToAttrs manpages) // { "SUMMARY.md" = summary; } + parsedToplevel = builtins.fromJSON toplevel; + + manpages = processCommand { + command = "nix"; + details = parsedToplevel; + filename = "nix"; + toplevel = parsedToplevel; + }; + + tableOfContents = let + showEntry = page: + " - [${page.command}](command-ref/new-cli/${page.name})"; + in concatStringsSep "\n" (map showEntry manpages) + "\n"; + +in (listToAttrs manpages) // { "SUMMARY.md" = tableOfContents; } diff --git a/doc/manual/generate-options.nix b/doc/manual/generate-options.nix index 2d586fa1b..a4ec36477 100644 --- a/doc/manual/generate-options.nix +++ b/doc/manual/generate-options.nix @@ -1,29 +1,41 @@ -with builtins; -with import ./utils.nix; +let + inherit (builtins) attrNames concatStringsSep isAttrs isBool; + inherit (import ./utils.nix) concatStrings squash splitLines; +in -options: +optionsInfo: +let + showOption = name: + let + inherit (optionsInfo.${name}) description documentDefault defaultValue aliases; + result = squash '' + - <span id="conf-${name}">[`${name}`](#conf-${name})</span> -concatStrings (map - (name: - let option = options.${name}; in - " - [`${name}`](#conf-${name})" - + "<p id=\"conf-${name}\"></p>\n\n" - + concatStrings (map (s: " ${s}\n") (splitLines option.description)) + "\n\n" - + (if option.documentDefault - then " **Default:** " + ( - if option.value == "" || option.value == [] - then "*empty*" - else if isBool option.value - then (if option.value then "`true`" else "`false`") - else - # n.b. a StringMap value type is specified as a string, but - # this shows the value type. The empty stringmap is "null" in - # JSON, but that converts to "{ }" here. - (if isAttrs option.value then "`\"\"`" - else "`" + toString option.value + "`")) + "\n\n" - else " **Default:** *machine-specific*\n") - + (if option.aliases != [] - then " **Deprecated alias:** " + (concatStringsSep ", " (map (s: "`${s}`") option.aliases)) + "\n\n" - else "") - ) - (attrNames options)) + ${indent " " body} + ''; + # separate body to cleanly handle indentation + body = '' + ${description} + + **Default:** ${showDefault documentDefault defaultValue} + + ${showAliases aliases} + ''; + showDefault = documentDefault: defaultValue: + if documentDefault then + # a StringMap value type is specified as a string, but + # this shows the value type. The empty stringmap is `null` in + # JSON, but that converts to `{ }` here. + if defaultValue == "" || defaultValue == [] || isAttrs defaultValue + then "*empty*" + else if isBool defaultValue then + if defaultValue then "`true`" else "`false`" + else "`${toString defaultValue}`" + else "*machine-specific*"; + showAliases = aliases: + if aliases == [] then "" else + "**Deprecated alias:** ${(concatStringsSep ", " (map (s: "`${s}`") aliases))}"; + indent = prefix: s: + concatStringsSep "\n" (map (x: if x == "" then x else "${prefix}${x}") (splitLines s)); + in result; +in concatStrings (map showOption (attrNames optionsInfo)) diff --git a/doc/manual/local.mk b/doc/manual/local.mk index 371ed6f21..531168e3f 100644 --- a/doc/manual/local.mk +++ b/doc/manual/local.mk @@ -1,13 +1,17 @@ ifeq ($(doc_generate),yes) -# Generate man pages. +MANUAL_SRCS := \ + $(call rwildcard, $(d)/src, *.md) \ + $(call rwildcard, $(d)/src, */*.md) + man-pages := $(foreach n, \ - nix-env.1 nix-build.1 nix-shell.1 nix-store.1 nix-instantiate.1 \ - nix-collect-garbage.1 \ - nix-prefetch-url.1 nix-channel.1 \ - nix-hash.1 nix-copy-closure.1 \ - nix.conf.5 nix-daemon.8, \ - $(d)/$(n)) + nix-env.1 nix-store.1 \ + nix-build.1 nix-shell.1 nix-instantiate.1 \ + nix-collect-garbage.1 \ + nix-prefetch-url.1 nix-channel.1 \ + nix-hash.1 nix-copy-closure.1 \ + nix.conf.5 nix-daemon.8 \ +, $(d)/$(n)) clean-files += $(d)/*.1 $(d)/*.5 $(d)/*.8 @@ -22,35 +26,47 @@ dummy-env = env -i \ nix-eval = $(dummy-env) $(bindir)/nix eval --experimental-features nix-command -I nix/corepkgs=corepkgs --store dummy:// --impure --raw +# re-implement mdBook's include directive to make it usable for terminal output and for proper @docroot@ substitution +define process-includes + while read -r line; do \ + filename=$$(sed 's/{{#include \(.*\)}}/\1/'<<< $$line); \ + matchline=$$(sed 's|/|\\/|g' <<< $$line); \ + sed -i "/$$matchline/r $$(dirname $(2))/$$filename" $(2); \ + sed -i "s/$$matchline//" $(2); \ + done < <(grep '{{#include' $(1)) +endef + $(d)/%.1: $(d)/src/command-ref/%.md @printf "Title: %s\n\n" "$$(basename $@ .1)" > $^.tmp @cat $^ >> $^.tmp - $(trace-gen) lowdown -sT man -M section=1 $^.tmp -o $@ + @$(call process-includes,$^,$^.tmp) + $(trace-gen) lowdown -sT man --nroff-nolinks -M section=1 $^.tmp -o $@ @rm $^.tmp $(d)/%.8: $(d)/src/command-ref/%.md @printf "Title: %s\n\n" "$$(basename $@ .8)" > $^.tmp @cat $^ >> $^.tmp - $(trace-gen) lowdown -sT man -M section=8 $^.tmp -o $@ + $(trace-gen) lowdown -sT man --nroff-nolinks -M section=8 $^.tmp -o $@ @rm $^.tmp $(d)/nix.conf.5: $(d)/src/command-ref/conf-file.md @printf "Title: %s\n\n" "$$(basename $@ .5)" > $^.tmp @cat $^ >> $^.tmp - $(trace-gen) lowdown -sT man -M section=5 $^.tmp -o $@ + $(trace-gen) lowdown -sT man --nroff-nolinks -M section=5 $^.tmp -o $@ @rm $^.tmp $(d)/src/SUMMARY.md: $(d)/src/SUMMARY.md.in $(d)/src/command-ref/new-cli - $(trace-gen) cat doc/manual/src/SUMMARY.md.in | while IFS= read line; do if [[ $$line = @manpages@ ]]; then cat doc/manual/src/command-ref/new-cli/SUMMARY.md; else echo "$$line"; fi; done > $@.tmp - @mv $@.tmp $@ + @cp $< $@ + @$(call process-includes,$@,$@) $(d)/src/command-ref/new-cli: $(d)/nix.json $(d)/generate-manpage.nix $(bindir)/nix @rm -rf $@ - $(trace-gen) $(nix-eval) --write-to $@ --expr 'import doc/manual/generate-manpage.nix { command = builtins.readFile $<; renderLinks = true; }' + $(trace-gen) $(nix-eval) --write-to $@.tmp --expr 'import doc/manual/generate-manpage.nix { toplevel = builtins.readFile $<; }' + @mv $@.tmp $@ $(d)/src/command-ref/conf-file.md: $(d)/conf-file.json $(d)/generate-options.nix $(d)/src/command-ref/conf-file-prefix.md $(bindir)/nix @cat doc/manual/src/command-ref/conf-file-prefix.md > $@.tmp - $(trace-gen) $(nix-eval) --expr 'import doc/manual/generate-options.nix (builtins.fromJSON (builtins.readFile $<))' >> $@.tmp + $(trace-gen) $(nix-eval) --expr 'import doc/manual/generate-options.nix (builtins.fromJSON (builtins.readFile $<))' >> $@.tmp; @mv $@.tmp $@ $(d)/nix.json: $(bindir)/nix @@ -61,10 +77,10 @@ $(d)/conf-file.json: $(bindir)/nix $(trace-gen) $(dummy-env) $(bindir)/nix show-config --json --experimental-features nix-command > $@.tmp @mv $@.tmp $@ -$(d)/src/expressions/builtins.md: $(d)/builtins.json $(d)/generate-builtins.nix $(d)/src/expressions/builtins-prefix.md $(bindir)/nix - @cat doc/manual/src/expressions/builtins-prefix.md > $@.tmp - $(trace-gen) $(nix-eval) --expr 'import doc/manual/generate-builtins.nix (builtins.fromJSON (builtins.readFile $<))' >> $@.tmp - @cat doc/manual/src/expressions/builtins-suffix.md >> $@.tmp +$(d)/src/language/builtins.md: $(d)/builtins.json $(d)/generate-builtins.nix $(d)/src/language/builtins-prefix.md $(bindir)/nix + @cat doc/manual/src/language/builtins-prefix.md > $@.tmp + $(trace-gen) $(nix-eval) --expr 'import doc/manual/generate-builtins.nix (builtins.fromJSON (builtins.readFile $<))' >> $@.tmp; + @cat doc/manual/src/language/builtins-suffix.md >> $@.tmp @mv $@.tmp $@ $(d)/builtins.json: $(bindir)/nix @@ -72,7 +88,8 @@ $(d)/builtins.json: $(bindir)/nix @mv $@.tmp $@ # Generate the HTML manual. -html: $(docdir)/manual/index.html +.PHONY: manual-html +manual-html: $(docdir)/manual/index.html install: $(docdir)/manual/index.html # Generate 'nix' manpages. @@ -80,6 +97,8 @@ install: $(mandir)/man1/nix3-manpages man: doc/manual/generated/man1/nix3-manpages all: doc/manual/generated/man1/nix3-manpages +# FIXME: unify with how the other man pages are generated. +# this one works differently and does not use any of the amenities provided by `/mk/lib.mk`. $(mandir)/man1/nix3-manpages: doc/manual/generated/man1/nix3-manpages @mkdir -p $(DESTDIR)$$(dirname $@) $(trace-install) install -m 0644 $$(dirname $<)/* $(DESTDIR)$$(dirname $@) @@ -87,17 +106,31 @@ $(mandir)/man1/nix3-manpages: doc/manual/generated/man1/nix3-manpages doc/manual/generated/man1/nix3-manpages: $(d)/src/command-ref/new-cli @mkdir -p $(DESTDIR)$$(dirname $@) $(trace-gen) for i in doc/manual/src/command-ref/new-cli/*.md; do \ - name=$$(basename $$i .md); \ - tmpFile=$$(mktemp); \ - if [[ $$name = SUMMARY ]]; then continue; fi; \ - printf "Title: %s\n\n" "$$name" > $$tmpFile; \ - cat $$i >> $$tmpFile; \ - lowdown -sT man -M section=1 $$tmpFile -o $(DESTDIR)$$(dirname $@)/$$name.1; \ - rm $$tmpFile; \ + name=$$(basename $$i .md); \ + tmpFile=$$(mktemp); \ + if [[ $$name = SUMMARY ]]; then continue; fi; \ + printf "Title: %s\n\n" "$$name" > $$tmpFile; \ + cat $$i >> $$tmpFile; \ + lowdown -sT man --nroff-nolinks -M section=1 $$tmpFile -o $(DESTDIR)$$(dirname $@)/$$name.1; \ + rm $$tmpFile; \ done @touch $@ -$(docdir)/manual/index.html: $(MANUAL_SRCS) $(d)/book.toml $(d)/anchors.jq $(d)/custom.css $(d)/src/SUMMARY.md $(d)/src/command-ref/new-cli $(d)/src/command-ref/conf-file.md $(d)/src/expressions/builtins.md $(call rwildcard, $(d)/src, *.md) - $(trace-gen) RUST_LOG=warn mdbook build doc/manual -d $(DESTDIR)$(docdir)/manual +$(docdir)/manual/index.html: $(MANUAL_SRCS) $(d)/book.toml $(d)/anchors.jq $(d)/custom.css $(d)/src/SUMMARY.md $(d)/src/command-ref/new-cli $(d)/src/command-ref/conf-file.md $(d)/src/language/builtins.md + $(trace-gen) \ + tmp="$$(mktemp -d)"; \ + cp -r doc/manual "$$tmp"; \ + find "$$tmp" -name '*.md' | while read -r file; do \ + $(call process-includes,$$file,$$file); \ + docroot="$$(realpath --relative-to="$$(dirname "$$file")" $$tmp/manual/src)"; \ + sed -i "s,@docroot@,$$docroot,g" "$$file"; \ + done; \ + set -euo pipefail; \ + RUST_LOG=warn mdbook build "$$tmp/manual" -d $(DESTDIR)$(docdir)/manual.tmp 2>&1 \ + | { grep -Fv "because fragment resolution isn't implemented" || :; }; \ + rm -rf "$$tmp/manual" + @rm -rf $(DESTDIR)$(docdir)/manual + @mv $(DESTDIR)$(docdir)/manual.tmp/html $(DESTDIR)$(docdir)/manual + @rm -rf $(DESTDIR)$(docdir)/manual.tmp endif diff --git a/doc/manual/redirects.js b/doc/manual/redirects.js index 19f928c7e..69f75d3a0 100644 --- a/doc/manual/redirects.js +++ b/doc/manual/redirects.js @@ -1,337 +1,421 @@ -// Redirects from old DocBook manual. -var redirects = { - "#part-advanced-topics": "advanced-topics/advanced-topics.html", - "#chap-tuning-cores-and-jobs": "advanced-topics/cores-vs-jobs.html", - "#chap-diff-hook": "advanced-topics/diff-hook.html", - "#check-dirs-are-unregistered": "advanced-topics/diff-hook.html#check-dirs-are-unregistered", - "#chap-distributed-builds": "advanced-topics/distributed-builds.html", - "#chap-post-build-hook": "advanced-topics/post-build-hook.html", - "#chap-post-build-hook-caveats": "advanced-topics/post-build-hook.html#implementation-caveats", - "#part-command-ref": "command-ref/command-ref.html", - "#conf-allow-import-from-derivation": "command-ref/conf-file.html#conf-allow-import-from-derivation", - "#conf-allow-new-privileges": "command-ref/conf-file.html#conf-allow-new-privileges", - "#conf-allowed-uris": "command-ref/conf-file.html#conf-allowed-uris", - "#conf-allowed-users": "command-ref/conf-file.html#conf-allowed-users", - "#conf-auto-optimise-store": "command-ref/conf-file.html#conf-auto-optimise-store", - "#conf-binary-cache-public-keys": "command-ref/conf-file.html#conf-binary-cache-public-keys", - "#conf-binary-caches": "command-ref/conf-file.html#conf-binary-caches", - "#conf-build-compress-log": "command-ref/conf-file.html#conf-build-compress-log", - "#conf-build-cores": "command-ref/conf-file.html#conf-build-cores", - "#conf-build-extra-chroot-dirs": "command-ref/conf-file.html#conf-build-extra-chroot-dirs", - "#conf-build-extra-sandbox-paths": "command-ref/conf-file.html#conf-build-extra-sandbox-paths", - "#conf-build-fallback": "command-ref/conf-file.html#conf-build-fallback", - "#conf-build-max-jobs": "command-ref/conf-file.html#conf-build-max-jobs", - "#conf-build-max-log-size": "command-ref/conf-file.html#conf-build-max-log-size", - "#conf-build-max-silent-time": "command-ref/conf-file.html#conf-build-max-silent-time", - "#conf-build-repeat": "command-ref/conf-file.html#conf-build-repeat", - "#conf-build-timeout": "command-ref/conf-file.html#conf-build-timeout", - "#conf-build-use-chroot": "command-ref/conf-file.html#conf-build-use-chroot", - "#conf-build-use-sandbox": "command-ref/conf-file.html#conf-build-use-sandbox", - "#conf-build-use-substitutes": "command-ref/conf-file.html#conf-build-use-substitutes", - "#conf-build-users-group": "command-ref/conf-file.html#conf-build-users-group", - "#conf-builders": "command-ref/conf-file.html#conf-builders", - "#conf-builders-use-substitutes": "command-ref/conf-file.html#conf-builders-use-substitutes", - "#conf-compress-build-log": "command-ref/conf-file.html#conf-compress-build-log", - "#conf-connect-timeout": "command-ref/conf-file.html#conf-connect-timeout", - "#conf-cores": "command-ref/conf-file.html#conf-cores", - "#conf-diff-hook": "command-ref/conf-file.html#conf-diff-hook", - "#conf-enforce-determinism": "command-ref/conf-file.html#conf-enforce-determinism", - "#conf-env-keep-derivations": "command-ref/conf-file.html#conf-env-keep-derivations", - "#conf-extra-binary-caches": "command-ref/conf-file.html#conf-extra-binary-caches", - "#conf-extra-platforms": "command-ref/conf-file.html#conf-extra-platforms", - "#conf-extra-sandbox-paths": "command-ref/conf-file.html#conf-extra-sandbox-paths", - "#conf-extra-substituters": "command-ref/conf-file.html#conf-extra-substituters", - "#conf-fallback": "command-ref/conf-file.html#conf-fallback", - "#conf-fsync-metadata": "command-ref/conf-file.html#conf-fsync-metadata", - "#conf-gc-keep-derivations": "command-ref/conf-file.html#conf-gc-keep-derivations", - "#conf-gc-keep-outputs": "command-ref/conf-file.html#conf-gc-keep-outputs", - "#conf-hashed-mirrors": "command-ref/conf-file.html#conf-hashed-mirrors", - "#conf-http-connections": "command-ref/conf-file.html#conf-http-connections", - "#conf-keep-build-log": "command-ref/conf-file.html#conf-keep-build-log", - "#conf-keep-derivations": "command-ref/conf-file.html#conf-keep-derivations", - "#conf-keep-env-derivations": "command-ref/conf-file.html#conf-keep-env-derivations", - "#conf-keep-outputs": "command-ref/conf-file.html#conf-keep-outputs", - "#conf-max-build-log-size": "command-ref/conf-file.html#conf-max-build-log-size", - "#conf-max-free": "command-ref/conf-file.html#conf-max-free", - "#conf-max-jobs": "command-ref/conf-file.html#conf-max-jobs", - "#conf-max-silent-time": "command-ref/conf-file.html#conf-max-silent-time", - "#conf-min-free": "command-ref/conf-file.html#conf-min-free", - "#conf-narinfo-cache-negative-ttl": "command-ref/conf-file.html#conf-narinfo-cache-negative-ttl", - "#conf-narinfo-cache-positive-ttl": "command-ref/conf-file.html#conf-narinfo-cache-positive-ttl", - "#conf-netrc-file": "command-ref/conf-file.html#conf-netrc-file", - "#conf-plugin-files": "command-ref/conf-file.html#conf-plugin-files", - "#conf-post-build-hook": "command-ref/conf-file.html#conf-post-build-hook", - "#conf-pre-build-hook": "command-ref/conf-file.html#conf-pre-build-hook", - "#conf-repeat": "command-ref/conf-file.html#conf-repeat", - "#conf-require-sigs": "command-ref/conf-file.html#conf-require-sigs", - "#conf-restrict-eval": "command-ref/conf-file.html#conf-restrict-eval", - "#conf-run-diff-hook": "command-ref/conf-file.html#conf-run-diff-hook", - "#conf-sandbox": "command-ref/conf-file.html#conf-sandbox", - "#conf-sandbox-dev-shm-size": "command-ref/conf-file.html#conf-sandbox-dev-shm-size", - "#conf-sandbox-paths": "command-ref/conf-file.html#conf-sandbox-paths", - "#conf-secret-key-files": "command-ref/conf-file.html#conf-secret-key-files", - "#conf-show-trace": "command-ref/conf-file.html#conf-show-trace", - "#conf-stalled-download-timeout": "command-ref/conf-file.html#conf-stalled-download-timeout", - "#conf-substitute": "command-ref/conf-file.html#conf-substitute", - "#conf-substituters": "command-ref/conf-file.html#conf-substituters", - "#conf-system": "command-ref/conf-file.html#conf-system", - "#conf-system-features": "command-ref/conf-file.html#conf-system-features", - "#conf-tarball-ttl": "command-ref/conf-file.html#conf-tarball-ttl", - "#conf-timeout": "command-ref/conf-file.html#conf-timeout", - "#conf-trace-function-calls": "command-ref/conf-file.html#conf-trace-function-calls", - "#conf-trusted-binary-caches": "command-ref/conf-file.html#conf-trusted-binary-caches", - "#conf-trusted-public-keys": "command-ref/conf-file.html#conf-trusted-public-keys", - "#conf-trusted-substituters": "command-ref/conf-file.html#conf-trusted-substituters", - "#conf-trusted-users": "command-ref/conf-file.html#conf-trusted-users", - "#extra-sandbox-paths": "command-ref/conf-file.html#extra-sandbox-paths", - "#sec-conf-file": "command-ref/conf-file.html", - "#env-NIX_PATH": "command-ref/env-common.html#env-NIX_PATH", - "#env-common": "command-ref/env-common.html", - "#envar-remote": "command-ref/env-common.html#env-NIX_REMOTE", - "#sec-common-env": "command-ref/env-common.html", - "#ch-files": "command-ref/files.html", - "#ch-main-commands": "command-ref/main-commands.html", - "#opt-out-link": "command-ref/nix-build.html#opt-out-link", - "#sec-nix-build": "command-ref/nix-build.html", - "#sec-nix-channel": "command-ref/nix-channel.html", - "#sec-nix-collect-garbage": "command-ref/nix-collect-garbage.html", - "#sec-nix-copy-closure": "command-ref/nix-copy-closure.html", - "#sec-nix-daemon": "command-ref/nix-daemon.html", - "#refsec-nix-env-install-examples": "command-ref/nix-env.html#examples", - "#rsec-nix-env-install": "command-ref/nix-env.html#operation---install", - "#rsec-nix-env-set": "command-ref/nix-env.html#operation---set", - "#rsec-nix-env-set-flag": "command-ref/nix-env.html#operation---set-flag", - "#rsec-nix-env-upgrade": "command-ref/nix-env.html#operation---upgrade", - "#sec-nix-env": "command-ref/nix-env.html", - "#ssec-version-comparisons": "command-ref/nix-env.html#versions", - "#sec-nix-hash": "command-ref/nix-hash.html", - "#sec-nix-instantiate": "command-ref/nix-instantiate.html", - "#sec-nix-prefetch-url": "command-ref/nix-prefetch-url.html", - "#sec-nix-shell": "command-ref/nix-shell.html", - "#ssec-nix-shell-shebang": "command-ref/nix-shell.html#use-as-a--interpreter", - "#nixref-queries": "command-ref/nix-store.html#queries", - "#opt-add-root": "command-ref/nix-store.html#opt-add-root", - "#refsec-nix-store-dump": "command-ref/nix-store.html#operation---dump", - "#refsec-nix-store-export": "command-ref/nix-store.html#operation---export", - "#refsec-nix-store-import": "command-ref/nix-store.html#operation---import", - "#refsec-nix-store-query": "command-ref/nix-store.html#operation---query", - "#refsec-nix-store-verify": "command-ref/nix-store.html#operation---verify", - "#rsec-nix-store-gc": "command-ref/nix-store.html#operation---gc", - "#rsec-nix-store-generate-binary-cache-key": "command-ref/nix-store.html#operation---generate-binary-cache-key", - "#rsec-nix-store-realise": "command-ref/nix-store.html#operation---realise", - "#rsec-nix-store-serve": "command-ref/nix-store.html#operation---serve", - "#sec-nix-store": "command-ref/nix-store.html", - "#opt-I": "command-ref/opt-common.html#opt-I", - "#opt-attr": "command-ref/opt-common.html#opt-attr", - "#opt-common": "command-ref/opt-common.html", - "#opt-cores": "command-ref/opt-common.html#opt-cores", - "#opt-log-format": "command-ref/opt-common.html#opt-log-format", - "#opt-max-jobs": "command-ref/opt-common.html#opt-max-jobs", - "#opt-max-silent-time": "command-ref/opt-common.html#opt-max-silent-time", - "#opt-timeout": "command-ref/opt-common.html#opt-timeout", - "#sec-common-options": "command-ref/opt-common.html", - "#ch-utilities": "command-ref/utilities.html", - "#chap-hacking": "contributing/hacking.html", - "#adv-attr-allowSubstitutes": "expressions/advanced-attributes.html#adv-attr-allowSubstitutes", - "#adv-attr-allowedReferences": "expressions/advanced-attributes.html#adv-attr-allowedReferences", - "#adv-attr-allowedRequisites": "expressions/advanced-attributes.html#adv-attr-allowedRequisites", - "#adv-attr-disallowedReferences": "expressions/advanced-attributes.html#adv-attr-disallowedReferences", - "#adv-attr-disallowedRequisites": "expressions/advanced-attributes.html#adv-attr-disallowedRequisites", - "#adv-attr-exportReferencesGraph": "expressions/advanced-attributes.html#adv-attr-exportReferencesGraph", - "#adv-attr-impureEnvVars": "expressions/advanced-attributes.html#adv-attr-impureEnvVars", - "#adv-attr-outputHash": "expressions/advanced-attributes.html#adv-attr-outputHash", - "#adv-attr-outputHashAlgo": "expressions/advanced-attributes.html#adv-attr-outputHashAlgo", - "#adv-attr-outputHashMode": "expressions/advanced-attributes.html#adv-attr-outputHashMode", - "#adv-attr-passAsFile": "expressions/advanced-attributes.html#adv-attr-passAsFile", - "#adv-attr-preferLocalBuild": "expressions/advanced-attributes.html#adv-attr-preferLocalBuild", - "#fixed-output-drvs": "expressions/advanced-attributes.html#adv-attr-outputHash", - "#sec-advanced-attributes": "expressions/advanced-attributes.html", - "#sec-arguments": "expressions/arguments-variables.html", - "#sec-build-script": "expressions/build-script.html", - "#builtin-abort": "expressions/builtins.html#builtins-abort", - "#builtin-add": "expressions/builtins.html#builtins-add", - "#builtin-all": "expressions/builtins.html#builtins-all", - "#builtin-any": "expressions/builtins.html#builtins-any", - "#builtin-attrNames": "expressions/builtins.html#builtins-attrNames", - "#builtin-attrValues": "expressions/builtins.html#builtins-attrValues", - "#builtin-baseNameOf": "expressions/builtins.html#builtins-baseNameOf", - "#builtin-bitAnd": "expressions/builtins.html#builtins-bitAnd", - "#builtin-bitOr": "expressions/builtins.html#builtins-bitOr", - "#builtin-bitXor": "expressions/builtins.html#builtins-bitXor", - "#builtin-builtins": "expressions/builtins.html#builtins-builtins", - "#builtin-compareVersions": "expressions/builtins.html#builtins-compareVersions", - "#builtin-concatLists": "expressions/builtins.html#builtins-concatLists", - "#builtin-concatStringsSep": "expressions/builtins.html#builtins-concatStringsSep", - "#builtin-currentSystem": "expressions/builtins.html#builtins-currentSystem", - "#builtin-deepSeq": "expressions/builtins.html#builtins-deepSeq", - "#builtin-derivation": "expressions/builtins.html#builtins-derivation", - "#builtin-dirOf": "expressions/builtins.html#builtins-dirOf", - "#builtin-div": "expressions/builtins.html#builtins-div", - "#builtin-elem": "expressions/builtins.html#builtins-elem", - "#builtin-elemAt": "expressions/builtins.html#builtins-elemAt", - "#builtin-fetchGit": "expressions/builtins.html#builtins-fetchGit", - "#builtin-fetchTarball": "expressions/builtins.html#builtins-fetchTarball", - "#builtin-fetchurl": "expressions/builtins.html#builtins-fetchurl", - "#builtin-filterSource": "expressions/builtins.html#builtins-filterSource", - "#builtin-foldl-prime": "expressions/builtins.html#builtins-foldl-prime", - "#builtin-fromJSON": "expressions/builtins.html#builtins-fromJSON", - "#builtin-functionArgs": "expressions/builtins.html#builtins-functionArgs", - "#builtin-genList": "expressions/builtins.html#builtins-genList", - "#builtin-getAttr": "expressions/builtins.html#builtins-getAttr", - "#builtin-getEnv": "expressions/builtins.html#builtins-getEnv", - "#builtin-hasAttr": "expressions/builtins.html#builtins-hasAttr", - "#builtin-hashFile": "expressions/builtins.html#builtins-hashFile", - "#builtin-hashString": "expressions/builtins.html#builtins-hashString", - "#builtin-head": "expressions/builtins.html#builtins-head", - "#builtin-import": "expressions/builtins.html#builtins-import", - "#builtin-intersectAttrs": "expressions/builtins.html#builtins-intersectAttrs", - "#builtin-isAttrs": "expressions/builtins.html#builtins-isAttrs", - "#builtin-isBool": "expressions/builtins.html#builtins-isBool", - "#builtin-isFloat": "expressions/builtins.html#builtins-isFloat", - "#builtin-isFunction": "expressions/builtins.html#builtins-isFunction", - "#builtin-isInt": "expressions/builtins.html#builtins-isInt", - "#builtin-isList": "expressions/builtins.html#builtins-isList", - "#builtin-isNull": "expressions/builtins.html#builtins-isNull", - "#builtin-isString": "expressions/builtins.html#builtins-isString", - "#builtin-length": "expressions/builtins.html#builtins-length", - "#builtin-lessThan": "expressions/builtins.html#builtins-lessThan", - "#builtin-listToAttrs": "expressions/builtins.html#builtins-listToAttrs", - "#builtin-map": "expressions/builtins.html#builtins-map", - "#builtin-match": "expressions/builtins.html#builtins-match", - "#builtin-mul": "expressions/builtins.html#builtins-mul", - "#builtin-parseDrvName": "expressions/builtins.html#builtins-parseDrvName", - "#builtin-path": "expressions/builtins.html#builtins-path", - "#builtin-pathExists": "expressions/builtins.html#builtins-pathExists", - "#builtin-placeholder": "expressions/builtins.html#builtins-placeholder", - "#builtin-readDir": "expressions/builtins.html#builtins-readDir", - "#builtin-readFile": "expressions/builtins.html#builtins-readFile", - "#builtin-removeAttrs": "expressions/builtins.html#builtins-removeAttrs", - "#builtin-replaceStrings": "expressions/builtins.html#builtins-replaceStrings", - "#builtin-seq": "expressions/builtins.html#builtins-seq", - "#builtin-sort": "expressions/builtins.html#builtins-sort", - "#builtin-split": "expressions/builtins.html#builtins-split", - "#builtin-splitVersion": "expressions/builtins.html#builtins-splitVersion", - "#builtin-stringLength": "expressions/builtins.html#builtins-stringLength", - "#builtin-sub": "expressions/builtins.html#builtins-sub", - "#builtin-substring": "expressions/builtins.html#builtins-substring", - "#builtin-tail": "expressions/builtins.html#builtins-tail", - "#builtin-throw": "expressions/builtins.html#builtins-throw", - "#builtin-toFile": "expressions/builtins.html#builtins-toFile", - "#builtin-toJSON": "expressions/builtins.html#builtins-toJSON", - "#builtin-toPath": "expressions/builtins.html#builtins-toPath", - "#builtin-toString": "expressions/builtins.html#builtins-toString", - "#builtin-toXML": "expressions/builtins.html#builtins-toXML", - "#builtin-trace": "expressions/builtins.html#builtins-trace", - "#builtin-tryEval": "expressions/builtins.html#builtins-tryEval", - "#builtin-typeOf": "expressions/builtins.html#builtins-typeOf", - "#ssec-builtins": "expressions/builtins.html", - "#attr-system": "expressions/derivations.html#attr-system", - "#ssec-derivation": "expressions/derivations.html", - "#ch-expression-language": "expressions/expression-language.html", - "#sec-expression-syntax": "expressions/expression-syntax.html", - "#sec-generic-builder": "expressions/generic-builder.html", - "#sec-constructs": "expressions/language-constructs.html", - "#sect-let-expressions": "expressions/language-constructs.html#let-expressions", - "#ss-functions": "expressions/language-constructs.html#functions", - "#sec-language-operators": "expressions/language-operators.html", - "#table-operators": "expressions/language-operators.html", - "#ssec-values": "expressions/language-values.html", - "#sec-building-simple": "expressions/simple-building-testing.html", - "#ch-simple-expression": "expressions/simple-expression.html", - "#chap-writing-nix-expressions": "expressions/writing-nix-expressions.html", - "#gloss-closure": "glossary.html#gloss-closure", - "#gloss-derivation": "glossary.html#gloss-derivation", - "#gloss-deriver": "glossary.html#gloss-deriver", - "#gloss-nar": "glossary.html#gloss-nar", - "#gloss-output-path": "glossary.html#gloss-output-path", - "#gloss-profile": "glossary.html#gloss-profile", - "#gloss-reachable": "glossary.html#gloss-reachable", - "#gloss-reference": "glossary.html#gloss-reference", - "#gloss-substitute": "glossary.html#gloss-substitute", - "#gloss-user-env": "glossary.html#gloss-user-env", - "#gloss-validity": "glossary.html#gloss-validity", - "#part-glossary": "glossary.html", - "#sec-building-source": "installation/building-source.html", - "#ch-env-variables": "installation/env-variables.html", - "#sec-installer-proxy-settings": "installation/env-variables.html#proxy-environment-variables", - "#sec-nix-ssl-cert-file": "installation/env-variables.html#nix_ssl_cert_file", - "#sec-nix-ssl-cert-file-with-nix-daemon-and-macos": "installation/env-variables.html#nix_ssl_cert_file-with-macos-and-the-nix-daemon", - "#chap-installation": "installation/installation.html", - "#ch-installing-binary": "installation/installing-binary.html", - "#sect-macos-installation": "installation/installing-binary.html#macos-installation", - "#sect-macos-installation-change-store-prefix": "installation/installing-binary.html#macos-installation", - "#sect-macos-installation-encrypted-volume": "installation/installing-binary.html#macos-installation", - "#sect-macos-installation-recommended-notes": "installation/installing-binary.html#macos-installation", - "#sect-macos-installation-symlink": "installation/installing-binary.html#macos-installation", - "#sect-multi-user-installation": "installation/installing-binary.html#multi-user-installation", - "#sect-nix-install-binary-tarball": "installation/installing-binary.html#installing-from-a-binary-tarball", - "#sect-nix-install-pinned-version-url": "installation/installing-binary.html#installing-a-pinned-nix-version-from-a-url", - "#sect-single-user-installation": "installation/installing-binary.html#single-user-installation", - "#ch-installing-source": "installation/installing-source.html", - "#ssec-multi-user": "installation/multi-user.html", - "#ch-nix-security": "installation/nix-security.html", - "#sec-obtaining-source": "installation/obtaining-source.html", - "#sec-prerequisites-source": "installation/prerequisites-source.html", - "#sec-single-user": "installation/single-user.html", - "#ch-supported-platforms": "installation/supported-platforms.html", - "#ch-upgrading-nix": "installation/upgrading.html", - "#ch-about-nix": "introduction.html", - "#chap-introduction": "introduction.html", - "#ch-basic-package-mgmt": "package-management/basic-package-mgmt.html", - "#ssec-binary-cache-substituter": "package-management/binary-cache-substituter.html", - "#sec-channels": "package-management/channels.html", - "#ssec-copy-closure": "package-management/copy-closure.html", - "#sec-garbage-collection": "package-management/garbage-collection.html", - "#ssec-gc-roots": "package-management/garbage-collector-roots.html", - "#chap-package-management": "package-management/package-management.html", - "#sec-profiles": "package-management/profiles.html", - "#ssec-s3-substituter": "package-management/s3-substituter.html", - "#ssec-s3-substituter-anonymous-reads": "package-management/s3-substituter.html#anonymous-reads-to-your-s3-compatible-binary-cache", - "#ssec-s3-substituter-authenticated-reads": "package-management/s3-substituter.html#authenticated-reads-to-your-s3-binary-cache", - "#ssec-s3-substituter-authenticated-writes": "package-management/s3-substituter.html#authenticated-writes-to-your-s3-compatible-binary-cache", - "#sec-sharing-packages": "package-management/sharing-packages.html", - "#ssec-ssh-substituter": "package-management/ssh-substituter.html", - "#chap-quick-start": "quick-start.html", - "#sec-relnotes": "release-notes/release-notes.html", - "#ch-relnotes-0.10.1": "release-notes/rl-0.10.1.html", - "#ch-relnotes-0.10": "release-notes/rl-0.10.html", - "#ssec-relnotes-0.11": "release-notes/rl-0.11.html", - "#ssec-relnotes-0.12": "release-notes/rl-0.12.html", - "#ssec-relnotes-0.13": "release-notes/rl-0.13.html", - "#ssec-relnotes-0.14": "release-notes/rl-0.14.html", - "#ssec-relnotes-0.15": "release-notes/rl-0.15.html", - "#ssec-relnotes-0.16": "release-notes/rl-0.16.html", - "#ch-relnotes-0.5": "release-notes/rl-0.5.html", - "#ch-relnotes-0.6": "release-notes/rl-0.6.html", - "#ch-relnotes-0.7": "release-notes/rl-0.7.html", - "#ch-relnotes-0.8.1": "release-notes/rl-0.8.1.html", - "#ch-relnotes-0.8": "release-notes/rl-0.8.html", - "#ch-relnotes-0.9.1": "release-notes/rl-0.9.1.html", - "#ch-relnotes-0.9.2": "release-notes/rl-0.9.2.html", - "#ch-relnotes-0.9": "release-notes/rl-0.9.html", - "#ssec-relnotes-1.0": "release-notes/rl-1.0.html", - "#ssec-relnotes-1.1": "release-notes/rl-1.1.html", - "#ssec-relnotes-1.10": "release-notes/rl-1.10.html", - "#ssec-relnotes-1.11.10": "release-notes/rl-1.11.10.html", - "#ssec-relnotes-1.11": "release-notes/rl-1.11.html", - "#ssec-relnotes-1.2": "release-notes/rl-1.2.html", - "#ssec-relnotes-1.3": "release-notes/rl-1.3.html", - "#ssec-relnotes-1.4": "release-notes/rl-1.4.html", - "#ssec-relnotes-1.5.1": "release-notes/rl-1.5.1.html", - "#ssec-relnotes-1.5.2": "release-notes/rl-1.5.2.html", - "#ssec-relnotes-1.5": "release-notes/rl-1.5.html", - "#ssec-relnotes-1.6.1": "release-notes/rl-1.6.1.html", - "#ssec-relnotes-1.6.0": "release-notes/rl-1.6.html", - "#ssec-relnotes-1.7": "release-notes/rl-1.7.html", - "#ssec-relnotes-1.8": "release-notes/rl-1.8.html", - "#ssec-relnotes-1.9": "release-notes/rl-1.9.html", - "#ssec-relnotes-2.0": "release-notes/rl-2.0.html", - "#ssec-relnotes-2.1": "release-notes/rl-2.1.html", - "#ssec-relnotes-2.2": "release-notes/rl-2.2.html", - "#ssec-relnotes-2.3": "release-notes/rl-2.3.html" +// redirect rules for anchors ensure backwards compatibility of URLs. +// this must be done on the client side, as web servers do not see the anchor part of the URL. + +// redirections are declared as follows: +// each entry has as its key a path matching the requested URL path, relative to the mdBook document root. +// +// IMPORTANT: it must specify the full path with file name and suffix +// +// each entry is itself a set of key-value pairs, where +// - keys are anchors on the matched path. +// - values are redirection targets relative to the current path. + +const redirects = { + "index.html": { + "part-advanced-topics": "advanced-topics/advanced-topics.html", + "chap-tuning-cores-and-jobs": "advanced-topics/cores-vs-jobs.html", + "chap-diff-hook": "advanced-topics/diff-hook.html", + "check-dirs-are-unregistered": "advanced-topics/diff-hook.html#check-dirs-are-unregistered", + "chap-distributed-builds": "advanced-topics/distributed-builds.html", + "chap-post-build-hook": "advanced-topics/post-build-hook.html", + "chap-post-build-hook-caveats": "advanced-topics/post-build-hook.html#implementation-caveats", + "part-command-ref": "command-ref/command-ref.html", + "conf-allow-import-from-derivation": "command-ref/conf-file.html#conf-allow-import-from-derivation", + "conf-allow-new-privileges": "command-ref/conf-file.html#conf-allow-new-privileges", + "conf-allowed-uris": "command-ref/conf-file.html#conf-allowed-uris", + "conf-allowed-users": "command-ref/conf-file.html#conf-allowed-users", + "conf-auto-optimise-store": "command-ref/conf-file.html#conf-auto-optimise-store", + "conf-binary-cache-public-keys": "command-ref/conf-file.html#conf-binary-cache-public-keys", + "conf-binary-caches": "command-ref/conf-file.html#conf-binary-caches", + "conf-build-compress-log": "command-ref/conf-file.html#conf-build-compress-log", + "conf-build-cores": "command-ref/conf-file.html#conf-build-cores", + "conf-build-extra-chroot-dirs": "command-ref/conf-file.html#conf-build-extra-chroot-dirs", + "conf-build-extra-sandbox-paths": "command-ref/conf-file.html#conf-build-extra-sandbox-paths", + "conf-build-fallback": "command-ref/conf-file.html#conf-build-fallback", + "conf-build-max-jobs": "command-ref/conf-file.html#conf-build-max-jobs", + "conf-build-max-log-size": "command-ref/conf-file.html#conf-build-max-log-size", + "conf-build-max-silent-time": "command-ref/conf-file.html#conf-build-max-silent-time", + "conf-build-timeout": "command-ref/conf-file.html#conf-build-timeout", + "conf-build-use-chroot": "command-ref/conf-file.html#conf-build-use-chroot", + "conf-build-use-sandbox": "command-ref/conf-file.html#conf-build-use-sandbox", + "conf-build-use-substitutes": "command-ref/conf-file.html#conf-build-use-substitutes", + "conf-build-users-group": "command-ref/conf-file.html#conf-build-users-group", + "conf-builders": "command-ref/conf-file.html#conf-builders", + "conf-builders-use-substitutes": "command-ref/conf-file.html#conf-builders-use-substitutes", + "conf-compress-build-log": "command-ref/conf-file.html#conf-compress-build-log", + "conf-connect-timeout": "command-ref/conf-file.html#conf-connect-timeout", + "conf-cores": "command-ref/conf-file.html#conf-cores", + "conf-diff-hook": "command-ref/conf-file.html#conf-diff-hook", + "conf-env-keep-derivations": "command-ref/conf-file.html#conf-env-keep-derivations", + "conf-extra-binary-caches": "command-ref/conf-file.html#conf-extra-binary-caches", + "conf-extra-platforms": "command-ref/conf-file.html#conf-extra-platforms", + "conf-extra-sandbox-paths": "command-ref/conf-file.html#conf-extra-sandbox-paths", + "conf-extra-substituters": "command-ref/conf-file.html#conf-extra-substituters", + "conf-fallback": "command-ref/conf-file.html#conf-fallback", + "conf-fsync-metadata": "command-ref/conf-file.html#conf-fsync-metadata", + "conf-gc-keep-derivations": "command-ref/conf-file.html#conf-gc-keep-derivations", + "conf-gc-keep-outputs": "command-ref/conf-file.html#conf-gc-keep-outputs", + "conf-hashed-mirrors": "command-ref/conf-file.html#conf-hashed-mirrors", + "conf-http-connections": "command-ref/conf-file.html#conf-http-connections", + "conf-keep-build-log": "command-ref/conf-file.html#conf-keep-build-log", + "conf-keep-derivations": "command-ref/conf-file.html#conf-keep-derivations", + "conf-keep-env-derivations": "command-ref/conf-file.html#conf-keep-env-derivations", + "conf-keep-outputs": "command-ref/conf-file.html#conf-keep-outputs", + "conf-max-build-log-size": "command-ref/conf-file.html#conf-max-build-log-size", + "conf-max-free": "command-ref/conf-file.html#conf-max-free", + "conf-max-jobs": "command-ref/conf-file.html#conf-max-jobs", + "conf-max-silent-time": "command-ref/conf-file.html#conf-max-silent-time", + "conf-min-free": "command-ref/conf-file.html#conf-min-free", + "conf-narinfo-cache-negative-ttl": "command-ref/conf-file.html#conf-narinfo-cache-negative-ttl", + "conf-narinfo-cache-positive-ttl": "command-ref/conf-file.html#conf-narinfo-cache-positive-ttl", + "conf-netrc-file": "command-ref/conf-file.html#conf-netrc-file", + "conf-plugin-files": "command-ref/conf-file.html#conf-plugin-files", + "conf-post-build-hook": "command-ref/conf-file.html#conf-post-build-hook", + "conf-pre-build-hook": "command-ref/conf-file.html#conf-pre-build-hook", + "conf-require-sigs": "command-ref/conf-file.html#conf-require-sigs", + "conf-restrict-eval": "command-ref/conf-file.html#conf-restrict-eval", + "conf-run-diff-hook": "command-ref/conf-file.html#conf-run-diff-hook", + "conf-sandbox": "command-ref/conf-file.html#conf-sandbox", + "conf-sandbox-dev-shm-size": "command-ref/conf-file.html#conf-sandbox-dev-shm-size", + "conf-sandbox-paths": "command-ref/conf-file.html#conf-sandbox-paths", + "conf-secret-key-files": "command-ref/conf-file.html#conf-secret-key-files", + "conf-show-trace": "command-ref/conf-file.html#conf-show-trace", + "conf-stalled-download-timeout": "command-ref/conf-file.html#conf-stalled-download-timeout", + "conf-substitute": "command-ref/conf-file.html#conf-substitute", + "conf-substituters": "command-ref/conf-file.html#conf-substituters", + "conf-system": "command-ref/conf-file.html#conf-system", + "conf-system-features": "command-ref/conf-file.html#conf-system-features", + "conf-tarball-ttl": "command-ref/conf-file.html#conf-tarball-ttl", + "conf-timeout": "command-ref/conf-file.html#conf-timeout", + "conf-trace-function-calls": "command-ref/conf-file.html#conf-trace-function-calls", + "conf-trusted-binary-caches": "command-ref/conf-file.html#conf-trusted-binary-caches", + "conf-trusted-public-keys": "command-ref/conf-file.html#conf-trusted-public-keys", + "conf-trusted-substituters": "command-ref/conf-file.html#conf-trusted-substituters", + "conf-trusted-users": "command-ref/conf-file.html#conf-trusted-users", + "extra-sandbox-paths": "command-ref/conf-file.html#extra-sandbox-paths", + "sec-conf-file": "command-ref/conf-file.html", + "env-NIX_PATH": "command-ref/env-common.html#env-NIX_PATH", + "env-common": "command-ref/env-common.html", + "envar-remote": "command-ref/env-common.html#env-NIX_REMOTE", + "sec-common-env": "command-ref/env-common.html", + "ch-files": "command-ref/files.html", + "ch-main-commands": "command-ref/main-commands.html", + "opt-out-link": "command-ref/nix-build.html#opt-out-link", + "sec-nix-build": "command-ref/nix-build.html", + "sec-nix-channel": "command-ref/nix-channel.html", + "sec-nix-collect-garbage": "command-ref/nix-collect-garbage.html", + "sec-nix-copy-closure": "command-ref/nix-copy-closure.html", + "sec-nix-daemon": "command-ref/nix-daemon.html", + "refsec-nix-env-install-examples": "command-ref/nix-env.html#examples", + "rsec-nix-env-install": "command-ref/nix-env.html#operation---install", + "rsec-nix-env-set": "command-ref/nix-env.html#operation---set", + "rsec-nix-env-set-flag": "command-ref/nix-env.html#operation---set-flag", + "rsec-nix-env-upgrade": "command-ref/nix-env.html#operation---upgrade", + "sec-nix-env": "command-ref/nix-env.html", + "ssec-version-comparisons": "command-ref/nix-env.html#versions", + "sec-nix-hash": "command-ref/nix-hash.html", + "sec-nix-instantiate": "command-ref/nix-instantiate.html", + "sec-nix-prefetch-url": "command-ref/nix-prefetch-url.html", + "sec-nix-shell": "command-ref/nix-shell.html", + "ssec-nix-shell-shebang": "command-ref/nix-shell.html#use-as-a--interpreter", + "nixref-queries": "command-ref/nix-store.html#queries", + "opt-add-root": "command-ref/nix-store.html#opt-add-root", + "refsec-nix-store-dump": "command-ref/nix-store.html#operation---dump", + "refsec-nix-store-export": "command-ref/nix-store.html#operation---export", + "refsec-nix-store-import": "command-ref/nix-store.html#operation---import", + "refsec-nix-store-query": "command-ref/nix-store.html#operation---query", + "refsec-nix-store-verify": "command-ref/nix-store.html#operation---verify", + "rsec-nix-store-gc": "command-ref/nix-store.html#operation---gc", + "rsec-nix-store-generate-binary-cache-key": "command-ref/nix-store.html#operation---generate-binary-cache-key", + "rsec-nix-store-realise": "command-ref/nix-store.html#operation---realise", + "rsec-nix-store-serve": "command-ref/nix-store.html#operation---serve", + "sec-nix-store": "command-ref/nix-store.html", + "opt-I": "command-ref/opt-common.html#opt-I", + "opt-attr": "command-ref/opt-common.html#opt-attr", + "opt-common": "command-ref/opt-common.html", + "opt-cores": "command-ref/opt-common.html#opt-cores", + "opt-log-format": "command-ref/opt-common.html#opt-log-format", + "opt-max-jobs": "command-ref/opt-common.html#opt-max-jobs", + "opt-max-silent-time": "command-ref/opt-common.html#opt-max-silent-time", + "opt-timeout": "command-ref/opt-common.html#opt-timeout", + "sec-common-options": "command-ref/opt-common.html", + "ch-utilities": "command-ref/utilities.html", + "chap-hacking": "contributing/hacking.html", + "adv-attr-allowSubstitutes": "language/advanced-attributes.html#adv-attr-allowSubstitutes", + "adv-attr-allowedReferences": "language/advanced-attributes.html#adv-attr-allowedReferences", + "adv-attr-allowedRequisites": "language/advanced-attributes.html#adv-attr-allowedRequisites", + "adv-attr-disallowedReferences": "language/advanced-attributes.html#adv-attr-disallowedReferences", + "adv-attr-disallowedRequisites": "language/advanced-attributes.html#adv-attr-disallowedRequisites", + "adv-attr-exportReferencesGraph": "language/advanced-attributes.html#adv-attr-exportReferencesGraph", + "adv-attr-impureEnvVars": "language/advanced-attributes.html#adv-attr-impureEnvVars", + "adv-attr-outputHash": "language/advanced-attributes.html#adv-attr-outputHash", + "adv-attr-outputHashAlgo": "language/advanced-attributes.html#adv-attr-outputHashAlgo", + "adv-attr-outputHashMode": "language/advanced-attributes.html#adv-attr-outputHashMode", + "adv-attr-passAsFile": "language/advanced-attributes.html#adv-attr-passAsFile", + "adv-attr-preferLocalBuild": "language/advanced-attributes.html#adv-attr-preferLocalBuild", + "fixed-output-drvs": "language/advanced-attributes.html#adv-attr-outputHash", + "sec-advanced-attributes": "language/advanced-attributes.html", + "builtin-abort": "language/builtins.html#builtins-abort", + "builtin-add": "language/builtins.html#builtins-add", + "builtin-all": "language/builtins.html#builtins-all", + "builtin-any": "language/builtins.html#builtins-any", + "builtin-attrNames": "language/builtins.html#builtins-attrNames", + "builtin-attrValues": "language/builtins.html#builtins-attrValues", + "builtin-baseNameOf": "language/builtins.html#builtins-baseNameOf", + "builtin-bitAnd": "language/builtins.html#builtins-bitAnd", + "builtin-bitOr": "language/builtins.html#builtins-bitOr", + "builtin-bitXor": "language/builtins.html#builtins-bitXor", + "builtin-builtins": "language/builtins.html#builtins-builtins", + "builtin-compareVersions": "language/builtins.html#builtins-compareVersions", + "builtin-concatLists": "language/builtins.html#builtins-concatLists", + "builtin-concatStringsSep": "language/builtins.html#builtins-concatStringsSep", + "builtin-currentSystem": "language/builtins.html#builtins-currentSystem", + "builtin-deepSeq": "language/builtins.html#builtins-deepSeq", + "builtin-derivation": "language/builtins.html#builtins-derivation", + "builtin-dirOf": "language/builtins.html#builtins-dirOf", + "builtin-div": "language/builtins.html#builtins-div", + "builtin-elem": "language/builtins.html#builtins-elem", + "builtin-elemAt": "language/builtins.html#builtins-elemAt", + "builtin-fetchGit": "language/builtins.html#builtins-fetchGit", + "builtin-fetchTarball": "language/builtins.html#builtins-fetchTarball", + "builtin-fetchurl": "language/builtins.html#builtins-fetchurl", + "builtin-filterSource": "language/builtins.html#builtins-filterSource", + "builtin-foldl-prime": "language/builtins.html#builtins-foldl-prime", + "builtin-fromJSON": "language/builtins.html#builtins-fromJSON", + "builtin-functionArgs": "language/builtins.html#builtins-functionArgs", + "builtin-genList": "language/builtins.html#builtins-genList", + "builtin-getAttr": "language/builtins.html#builtins-getAttr", + "builtin-getEnv": "language/builtins.html#builtins-getEnv", + "builtin-hasAttr": "language/builtins.html#builtins-hasAttr", + "builtin-hashFile": "language/builtins.html#builtins-hashFile", + "builtin-hashString": "language/builtins.html#builtins-hashString", + "builtin-head": "language/builtins.html#builtins-head", + "builtin-import": "language/builtins.html#builtins-import", + "builtin-intersectAttrs": "language/builtins.html#builtins-intersectAttrs", + "builtin-isAttrs": "language/builtins.html#builtins-isAttrs", + "builtin-isBool": "language/builtins.html#builtins-isBool", + "builtin-isFloat": "language/builtins.html#builtins-isFloat", + "builtin-isFunction": "language/builtins.html#builtins-isFunction", + "builtin-isInt": "language/builtins.html#builtins-isInt", + "builtin-isList": "language/builtins.html#builtins-isList", + "builtin-isNull": "language/builtins.html#builtins-isNull", + "builtin-isString": "language/builtins.html#builtins-isString", + "builtin-length": "language/builtins.html#builtins-length", + "builtin-lessThan": "language/builtins.html#builtins-lessThan", + "builtin-listToAttrs": "language/builtins.html#builtins-listToAttrs", + "builtin-map": "language/builtins.html#builtins-map", + "builtin-match": "language/builtins.html#builtins-match", + "builtin-mul": "language/builtins.html#builtins-mul", + "builtin-parseDrvName": "language/builtins.html#builtins-parseDrvName", + "builtin-path": "language/builtins.html#builtins-path", + "builtin-pathExists": "language/builtins.html#builtins-pathExists", + "builtin-placeholder": "language/builtins.html#builtins-placeholder", + "builtin-readDir": "language/builtins.html#builtins-readDir", + "builtin-readFile": "language/builtins.html#builtins-readFile", + "builtin-removeAttrs": "language/builtins.html#builtins-removeAttrs", + "builtin-replaceStrings": "language/builtins.html#builtins-replaceStrings", + "builtin-seq": "language/builtins.html#builtins-seq", + "builtin-sort": "language/builtins.html#builtins-sort", + "builtin-split": "language/builtins.html#builtins-split", + "builtin-splitVersion": "language/builtins.html#builtins-splitVersion", + "builtin-stringLength": "language/builtins.html#builtins-stringLength", + "builtin-sub": "language/builtins.html#builtins-sub", + "builtin-substring": "language/builtins.html#builtins-substring", + "builtin-tail": "language/builtins.html#builtins-tail", + "builtin-throw": "language/builtins.html#builtins-throw", + "builtin-toFile": "language/builtins.html#builtins-toFile", + "builtin-toJSON": "language/builtins.html#builtins-toJSON", + "builtin-toPath": "language/builtins.html#builtins-toPath", + "builtin-toString": "language/builtins.html#builtins-toString", + "builtin-toXML": "language/builtins.html#builtins-toXML", + "builtin-trace": "language/builtins.html#builtins-trace", + "builtin-tryEval": "language/builtins.html#builtins-tryEval", + "builtin-typeOf": "language/builtins.html#builtins-typeOf", + "ssec-builtins": "language/builtins.html", + "attr-system": "language/derivations.html#attr-system", + "ssec-derivation": "language/derivations.html", + "ch-expression-language": "language/index.html", + "sec-constructs": "language/constructs.html", + "sect-let-language": "language/constructs.html#let-language", + "ss-functions": "language/constructs.html#functions", + "sec-language-operators": "language/operators.html", + "table-operators": "language/operators.html", + "ssec-values": "language/values.html", + "gloss-closure": "glossary.html#gloss-closure", + "gloss-derivation": "glossary.html#gloss-derivation", + "gloss-deriver": "glossary.html#gloss-deriver", + "gloss-nar": "glossary.html#gloss-nar", + "gloss-output-path": "glossary.html#gloss-output-path", + "gloss-profile": "glossary.html#gloss-profile", + "gloss-reachable": "glossary.html#gloss-reachable", + "gloss-reference": "glossary.html#gloss-reference", + "gloss-substitute": "glossary.html#gloss-substitute", + "gloss-user-env": "glossary.html#gloss-user-env", + "gloss-validity": "glossary.html#gloss-validity", + "part-glossary": "glossary.html", + "sec-building-source": "installation/building-source.html", + "ch-env-variables": "installation/env-variables.html", + "sec-installer-proxy-settings": "installation/env-variables.html#proxy-environment-variables", + "sec-nix-ssl-cert-file": "installation/env-variables.html#nix_ssl_cert_file", + "sec-nix-ssl-cert-file-with-nix-daemon-and-macos": "installation/env-variables.html#nix_ssl_cert_file-with-macos-and-the-nix-daemon", + "chap-installation": "installation/installation.html", + "ch-installing-binary": "installation/installing-binary.html", + "sect-macos-installation": "installation/installing-binary.html#macos-installation", + "sect-macos-installation-change-store-prefix": "installation/installing-binary.html#macos-installation", + "sect-macos-installation-encrypted-volume": "installation/installing-binary.html#macos-installation", + "sect-macos-installation-recommended-notes": "installation/installing-binary.html#macos-installation", + "sect-macos-installation-symlink": "installation/installing-binary.html#macos-installation", + "sect-multi-user-installation": "installation/installing-binary.html#multi-user-installation", + "sect-nix-install-binary-tarball": "installation/installing-binary.html#installing-from-a-binary-tarball", + "sect-nix-install-pinned-version-url": "installation/installing-binary.html#installing-a-pinned-nix-version-from-a-url", + "sect-single-user-installation": "installation/installing-binary.html#single-user-installation", + "ch-installing-source": "installation/installing-source.html", + "ssec-multi-user": "installation/multi-user.html", + "ch-nix-security": "installation/nix-security.html", + "sec-obtaining-source": "installation/obtaining-source.html", + "sec-prerequisites-source": "installation/prerequisites-source.html", + "sec-single-user": "installation/single-user.html", + "ch-supported-platforms": "installation/supported-platforms.html", + "ch-upgrading-nix": "installation/upgrading.html", + "ch-about-nix": "introduction.html", + "chap-introduction": "introduction.html", + "ch-basic-package-mgmt": "package-management/basic-package-mgmt.html", + "ssec-binary-cache-substituter": "package-management/binary-cache-substituter.html", + "sec-channels": "package-management/channels.html", + "ssec-copy-closure": "package-management/copy-closure.html", + "sec-garbage-collection": "package-management/garbage-collection.html", + "ssec-gc-roots": "package-management/garbage-collector-roots.html", + "chap-package-management": "package-management/package-management.html", + "sec-profiles": "package-management/profiles.html", + "ssec-s3-substituter": "package-management/s3-substituter.html", + "ssec-s3-substituter-anonymous-reads": "package-management/s3-substituter.html#anonymous-reads-to-your-s3-compatible-binary-cache", + "ssec-s3-substituter-authenticated-reads": "package-management/s3-substituter.html#authenticated-reads-to-your-s3-binary-cache", + "ssec-s3-substituter-authenticated-writes": "package-management/s3-substituter.html#authenticated-writes-to-your-s3-compatible-binary-cache", + "sec-sharing-packages": "package-management/sharing-packages.html", + "ssec-ssh-substituter": "package-management/ssh-substituter.html", + "chap-quick-start": "quick-start.html", + "sec-relnotes": "release-notes/release-notes.html", + "ch-relnotes-0.10.1": "release-notes/rl-0.10.1.html", + "ch-relnotes-0.10": "release-notes/rl-0.10.html", + "ssec-relnotes-0.11": "release-notes/rl-0.11.html", + "ssec-relnotes-0.12": "release-notes/rl-0.12.html", + "ssec-relnotes-0.13": "release-notes/rl-0.13.html", + "ssec-relnotes-0.14": "release-notes/rl-0.14.html", + "ssec-relnotes-0.15": "release-notes/rl-0.15.html", + "ssec-relnotes-0.16": "release-notes/rl-0.16.html", + "ch-relnotes-0.5": "release-notes/rl-0.5.html", + "ch-relnotes-0.6": "release-notes/rl-0.6.html", + "ch-relnotes-0.7": "release-notes/rl-0.7.html", + "ch-relnotes-0.8.1": "release-notes/rl-0.8.1.html", + "ch-relnotes-0.8": "release-notes/rl-0.8.html", + "ch-relnotes-0.9.1": "release-notes/rl-0.9.1.html", + "ch-relnotes-0.9.2": "release-notes/rl-0.9.2.html", + "ch-relnotes-0.9": "release-notes/rl-0.9.html", + "ssec-relnotes-1.0": "release-notes/rl-1.0.html", + "ssec-relnotes-1.1": "release-notes/rl-1.1.html", + "ssec-relnotes-1.10": "release-notes/rl-1.10.html", + "ssec-relnotes-1.11.10": "release-notes/rl-1.11.10.html", + "ssec-relnotes-1.11": "release-notes/rl-1.11.html", + "ssec-relnotes-1.2": "release-notes/rl-1.2.html", + "ssec-relnotes-1.3": "release-notes/rl-1.3.html", + "ssec-relnotes-1.4": "release-notes/rl-1.4.html", + "ssec-relnotes-1.5.1": "release-notes/rl-1.5.1.html", + "ssec-relnotes-1.5.2": "release-notes/rl-1.5.2.html", + "ssec-relnotes-1.5": "release-notes/rl-1.5.html", + "ssec-relnotes-1.6.1": "release-notes/rl-1.6.1.html", + "ssec-relnotes-1.6.0": "release-notes/rl-1.6.html", + "ssec-relnotes-1.7": "release-notes/rl-1.7.html", + "ssec-relnotes-1.8": "release-notes/rl-1.8.html", + "ssec-relnotes-1.9": "release-notes/rl-1.9.html", + "ssec-relnotes-2.0": "release-notes/rl-2.0.html", + "ssec-relnotes-2.1": "release-notes/rl-2.1.html", + "ssec-relnotes-2.2": "release-notes/rl-2.2.html", + "ssec-relnotes-2.3": "release-notes/rl-2.3.html" + }, + "language/values.html": { + "simple-values": "#primitives", + "lists": "#list", + "strings": "#string", + "lists": "#list", + "attribute-sets": "#attribute-set" + } }; -var isRoot = (document.location.pathname.endsWith('/') || document.location.pathname.endsWith('/index.html')) && path_to_root === ''; -if (isRoot && redirects[document.location.hash]) { - document.location.href = path_to_root + redirects[document.location.hash]; +// the following code matches the current page's URL against the set of redirects. +// +// it is written to minimize the latency between page load and redirect. +// therefore we avoid function calls, copying data, and unnecessary loops. +// IMPORTANT: we use stateful array operations and their order matters! +// +// matching URLs is more involved than it should be: +// +// 1. `document.location.pathname` can have an arbitrary prefix. +// +// 2. `path_to_root` is set by mdBook. it consists only of `../`s and +// determines the depth of `<path>` relative to the prefix: +// +// `document.location.pathname` +// |------------------------------| +// /<prefix>/<path>/[<file>[.html]][#<anchor>] +// |----| +// `path_to_root` has same number of path segments +// +// source: https://phaiax.github.io/mdBook/format/theme/index-hbs.html#data +// +// 3. the following paths are equivalent: +// +// /foo/bar/ +// /foo/bar/index.html +// /foo/bar/index +// +// 4. the following paths are also equivalent: +// +// /foo/bar/baz +// /foo/bar/baz.html +// + +let segments = document.location.pathname.split('/'); + +let file = segments.pop(); + +// normalize file name +if (file === '') { file = "index.html"; } +else if (!file.endsWith('.html')) { file = file + '.html'; } + +segments.push(file); + +// use `path_to_root` to discern prefix from path. +const depth = path_to_root.split('/').length; + +// remove segments containing prefix. the following works because +// 1. the original `document.location.pathname` is absolute, +// hence first element of `segments` is always empty. +// 2. last element of splitting `path_to_root` is also always empty. +// 3. last element of `segments` is the file name. +// +// visual example: +// +// '/foo/bar/baz.html'.split('/') -> [ '', 'foo', 'bar', 'baz.html' ] +// '../'.split('/') -> [ '..', '' ] +// +// the following operations will then result in +// +// path = 'bar/baz.html' +// +segments.splice(0, segments.length - depth); +const path = segments.join('/'); + +// anchor starts with the hash character (`#`), +// but our redirect declarations don't, so we strip it. +// example: +// document.location.hash -> '#foo' +// document.location.hash.substring(1) -> 'foo' +const anchor = document.location.hash.substring(1); + +const redirect = redirects[path]; +if (redirect) { + const target = redirect[anchor]; + if (target) { + document.location.href = target; + } } diff --git a/doc/manual/src/SUMMARY.md.in b/doc/manual/src/SUMMARY.md.in index c8cb72fc0..f3e2c9ae6 100644 --- a/doc/manual/src/SUMMARY.md.in +++ b/doc/manual/src/SUMMARY.md.in @@ -26,21 +26,15 @@ - [Copying Closures via SSH](package-management/copy-closure.md) - [Serving a Nix store via SSH](package-management/ssh-substituter.md) - [Serving a Nix store via S3](package-management/s3-substituter.md) -- [Writing Nix Expressions](expressions/writing-nix-expressions.md) - - [A Simple Nix Expression](expressions/simple-expression.md) - - [Expression Syntax](expressions/expression-syntax.md) - - [Build Script](expressions/build-script.md) - - [Arguments and Variables](expressions/arguments-variables.md) - - [Building and Testing](expressions/simple-building-testing.md) - - [Generic Builder Syntax](expressions/generic-builder.md) - - [Nix Expression Language](expressions/expression-language.md) - - [Data Types](expressions/language-values.md) - - [Language Constructs](expressions/language-constructs.md) - - [Operators](expressions/language-operators.md) - - [Derivations](expressions/derivations.md) - - [Advanced Attributes](expressions/advanced-attributes.md) - - [Built-in Constants](expressions/builtin-constants.md) - - [Built-in Functions](expressions/builtins.md) +- [Nix Language](language/index.md) + - [Data Types](language/values.md) + - [Language Constructs](language/constructs.md) + - [String interpolation](language/string-interpolation.md) + - [Operators](language/operators.md) + - [Derivations](language/derivations.md) + - [Advanced Attributes](language/advanced-attributes.md) + - [Built-in Constants](language/builtin-constants.md) + - [Built-in Functions](language/builtins.md) - [Advanced Topics](advanced-topics/advanced-topics.md) - [Remote Builds](advanced-topics/distributed-builds.md) - [Tuning Cores and Jobs](advanced-topics/cores-vs-jobs.md) @@ -50,10 +44,10 @@ - [Common Options](command-ref/opt-common.md) - [Common Environment Variables](command-ref/env-common.md) - [Main Commands](command-ref/main-commands.md) - - [nix-env](command-ref/nix-env.md) - [nix-build](command-ref/nix-build.md) - [nix-shell](command-ref/nix-shell.md) - [nix-store](command-ref/nix-store.md) + - [nix-env](command-ref/nix-env.md) - [Utilities](command-ref/utilities.md) - [nix-channel](command-ref/nix-channel.md) - [nix-collect-garbage](command-ref/nix-collect-garbage.md) @@ -63,15 +57,20 @@ - [nix-instantiate](command-ref/nix-instantiate.md) - [nix-prefetch-url](command-ref/nix-prefetch-url.md) - [Experimental Commands](command-ref/experimental-commands.md) -@manpages@ +{{#include ./command-ref/new-cli/SUMMARY.md}} - [Files](command-ref/files.md) - [nix.conf](command-ref/conf-file.md) +- [Architecture](architecture/architecture.md) - [Glossary](glossary.md) - [Contributing](contributing/contributing.md) - [Hacking](contributing/hacking.md) - [CLI guideline](contributing/cli-guideline.md) - [Release Notes](release-notes/release-notes.md) - [Release X.Y (202?-??-??)](release-notes/rl-next.md) + - [Release 2.14 (2023-02-28)](release-notes/rl-2.14.md) + - [Release 2.13 (2023-01-17)](release-notes/rl-2.13.md) + - [Release 2.12 (2022-12-06)](release-notes/rl-2.12.md) + - [Release 2.11 (2022-08-25)](release-notes/rl-2.11.md) - [Release 2.10 (2022-07-11)](release-notes/rl-2.10.md) - [Release 2.9 (2022-05-30)](release-notes/rl-2.9.md) - [Release 2.8 (2022-04-19)](release-notes/rl-2.8.md) diff --git a/doc/manual/src/advanced-topics/diff-hook.md b/doc/manual/src/advanced-topics/diff-hook.md index 161e64b2a..4a742c160 100644 --- a/doc/manual/src/advanced-topics/diff-hook.md +++ b/doc/manual/src/advanced-topics/diff-hook.md @@ -121,37 +121,3 @@ error: are not valid, so checking is not possible Run the build without `--check`, and then try with `--check` again. - -# Automatic and Optionally Enforced Determinism Verification - -Automatically verify every build at build time by executing the build -multiple times. - -Setting `repeat` and `enforce-determinism` in your `nix.conf` permits -the automated verification of every build Nix performs. - -The following configuration will run each build three times, and will -require the build to be deterministic: - - enforce-determinism = true - repeat = 2 - -Setting `enforce-determinism` to false as in the following -configuration will run the build multiple times, execute the build -hook, but will allow the build to succeed even if it does not build -reproducibly: - - enforce-determinism = false - repeat = 1 - -An example output of this configuration: - -```console -$ nix-build ./test.nix -A unstable -this derivation will be built: - /nix/store/ch6llwpr2h8c3jmnf3f2ghkhx59aa97f-unstable.drv -building '/nix/store/ch6llwpr2h8c3jmnf3f2ghkhx59aa97f-unstable.drv' (round 1/2)... -building '/nix/store/ch6llwpr2h8c3jmnf3f2ghkhx59aa97f-unstable.drv' (round 2/2)... -output '/nix/store/6xg356v9gl03hpbbg8gws77n19qanh02-unstable' of '/nix/store/ch6llwpr2h8c3jmnf3f2ghkhx59aa97f-unstable.drv' differs from '/nix/store/6xg356v9gl03hpbbg8gws77n19qanh02-unstable.check' from previous round -/nix/store/6xg356v9gl03hpbbg8gws77n19qanh02-unstable -``` diff --git a/doc/manual/src/advanced-topics/post-build-hook.md b/doc/manual/src/advanced-topics/post-build-hook.md index fcb52d878..1479cc3a4 100644 --- a/doc/manual/src/advanced-topics/post-build-hook.md +++ b/doc/manual/src/advanced-topics/post-build-hook.md @@ -33,12 +33,17 @@ distribute the public key for verifying the authenticity of the paths. example-nix-cache-1:1/cKDz3QCCOmwcztD2eV6Coggp6rqc9DGjWv7C0G+rM= ``` -Then, add the public key and the cache URL to your `nix.conf`'s -`trusted-public-keys` and `substituters` options: +Then update [`nix.conf`](../command-ref/conf-file.md) on any machine that will access the cache. +Add the cache URL to [`substituters`](../command-ref/conf-file.md#conf-substituters) and the public key to [`trusted-public-keys`](../command-ref/conf-file.md#conf-trusted-public-keys): substituters = https://cache.nixos.org/ s3://example-nix-cache trusted-public-keys = cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY= example-nix-cache-1:1/cKDz3QCCOmwcztD2eV6Coggp6rqc9DGjWv7C0G+rM= +Machines that build for the cache must sign derivations using the private key. +On those machines, add the path to the key file to the [`secret-key-files`](../command-ref/conf-file.md#conf-secret-key-files) field in their [`nix.conf`](../command-ref/conf-file.md): + + secret-key-files = /etc/nix/key.private + We will restart the Nix daemon in a later step. # Implementing the build hook @@ -52,14 +57,12 @@ set -eu set -f # disable globbing export IFS=' ' -echo "Signing paths" $OUT_PATHS -nix store sign --key-file /etc/nix/key.private $OUT_PATHS echo "Uploading paths" $OUT_PATHS -exec nix copy --to 's3://example-nix-cache' $OUT_PATHS +exec nix copy --to "s3://example-nix-cache" $OUT_PATHS ``` > **Note** -> +> > The `$OUT_PATHS` variable is a space-separated list of Nix store > paths. In this case, we expect and want the shell to perform word > splitting to make each output path its own argument to `nix diff --git a/doc/manual/src/architecture/architecture.md b/doc/manual/src/architecture/architecture.md new file mode 100644 index 000000000..e51958052 --- /dev/null +++ b/doc/manual/src/architecture/architecture.md @@ -0,0 +1,115 @@ +# Architecture + +This chapter describes how Nix works. +It should help users understand why Nix behaves as it does, and it should help developers understand how to modify Nix and how to write similar tools. + +## Overview + +Nix consists of [hierarchical layers]. + +[hierarchical layers]: https://en.m.wikipedia.org/wiki/Multitier_architecture#Layers + +The following [concept map] shows its main components (rectangles), the objects they operate on (rounded rectangles), and their interactions (connecting phrases): + +[concept map]: https://en.m.wikipedia.org/wiki/Concept_map + +``` + + .----------------. + | Nix expression |----------. + '----------------' | + | passed to + | | ++----------|-------------------|--------------------------------+ +| Nix | V | +| | +-------------------------+ | +| | | commmand line interface |------. | +| | +-------------------------+ | | +| | | | | +| evaluated by calls manages | +| | | | | +| | V | | +| | +--------------------+ | | +| '-------->| language evaluator | | | +| +--------------------+ | | +| | | | +| produces | | +| | V | +| +----------------------------|------------------------------+ | +| | store | | | +| | referenced by V builds | | +| | .-------------. .------------. .--------------. | | +| | | build input |----->| build plan |----->| build result | | | +| | '-------------' '------------' '--------------' | | +| +-------------------------------------------------|---------+ | ++---------------------------------------------------|-----------+ + | + represented as + | + V + .---------------. + | file | + '---------------' +``` + +At the top is the [command line interface](../command-ref/command-ref.md) that drives the underlying layers. + +The [Nix language](../language/index.md) evaluator transforms Nix expressions into self-contained *build plans*, which are used to derive *build results* from referenced *build inputs*. + +The command line interface and Nix expressions are what users deal with most. + +> **Note** +> The Nix language itself does not have a notion of *packages* or *configurations*. +> As far as we are concerned here, the inputs and results of a build plan are just data. + +Underlying the command line interface and the Nix language evaluator is the [Nix store](../glossary.md#gloss-store), a mechanism to keep track of build plans, data, and references between them. +It can also execute build plans to produce new data, which are made available to the operating system as files. + +A build plan itself is a series of *build tasks*, together with their build inputs. + +> **Important** +> A build task in Nix is called [derivation](../glossary.md#gloss-derivation). + +Each build task has a special build input executed as *build instructions* in order to perform the build. +The result of a build task can be input to another build task. + +The following [data flow diagram] shows a build plan for illustration. +Build inputs used as instructions to a build task are marked accordingly: + +[data flow diagram]: https://en.m.wikipedia.org/wiki/Data-flow_diagram + +``` ++--------------------------------------------------------------------+ +| build plan | +| | +| .-------------. | +| | build input |---------. | +| '-------------' | | +| instructions | +| | | +| v | +| .-------------. .----------. | +| | build input |-->( build task )-------. | +| '-------------' '----------' | | +| instructions | +| | | +| v | +| .-------------. .----------. .--------------. | +| | build input |---------. ( build task )--->| build result | | +| '-------------' | '----------' '--------------' | +| instructions ^ | +| | | | +| v | | +| .-------------. .----------. | | +| | build input |-->( build task )-------' | +| '-------------' '----------' | +| ^ | +| | | +| | | +| .-------------. | | +| | build input |---------' | +| '-------------' | +| | ++--------------------------------------------------------------------+ +``` + diff --git a/doc/manual/src/command-ref/env-common.md b/doc/manual/src/command-ref/env-common.md index 983079794..2be55be52 100644 --- a/doc/manual/src/command-ref/env-common.md +++ b/doc/manual/src/command-ref/env-common.md @@ -7,42 +7,11 @@ Most Nix commands interpret the following environment variables: `nix-shell`. It can have the values `pure` or `impure`. - [`NIX_PATH`]{#env-NIX_PATH}\ - A colon-separated list of directories used to look up Nix - expressions enclosed in angle brackets (i.e., `<path>`). For - instance, the value - - /home/eelco/Dev:/etc/nixos - - will cause Nix to look for paths relative to `/home/eelco/Dev` and - `/etc/nixos`, in this order. It is also possible to match paths - against a prefix. For example, the value - - nixpkgs=/home/eelco/Dev/nixpkgs-branch:/etc/nixos - - will cause Nix to search for `<nixpkgs/path>` in - `/home/eelco/Dev/nixpkgs-branch/path` and `/etc/nixos/nixpkgs/path`. - - If a path in the Nix search path starts with `http://` or - `https://`, it is interpreted as the URL of a tarball that will be - downloaded and unpacked to a temporary location. The tarball must - consist of a single top-level directory. For example, setting - `NIX_PATH` to - - nixpkgs=https://github.com/NixOS/nixpkgs/archive/master.tar.gz - - tells Nix to download and use the current contents of the - `master` branch in the `nixpkgs` repository. - - The URLs of the tarballs from the official nixos.org channels (see - [the manual for `nix-channel`](nix-channel.md)) can be abbreviated - as `channel:<channel-name>`. For instance, the following two - values of `NIX_PATH` are equivalent: - - nixpkgs=channel:nixos-21.05 - nixpkgs=https://nixos.org/channels/nixos-21.05/nixexprs.tar.xz - - The Nix search path can also be extended using the `-I` option to - many Nix commands, which takes precedence over `NIX_PATH`. + A colon-separated list of directories used to look up the location of Nix + expressions using [paths](../language/values.md#type-path) + enclosed in angle brackets (i.e., `<path>`), + e.g. `/home/eelco/Dev:/etc/nixos`. It can be extended using the + [`-I` option](./opt-common.md#opt-I). If `NIX_PATH` is not set at all, Nix will fall back to the following list in [impure](./conf-file.md#conf-pure-eval) and [unrestricted](./conf-file.md#conf-restrict-eval) evaluation mode: @@ -133,3 +102,16 @@ Most Nix commands interpret the following environment variables: variable sets the initial size of the heap in bytes. It defaults to 384 MiB. Setting it to a low value reduces memory consumption, but will increase runtime due to the overhead of garbage collection. + +## XDG Base Directory + +New Nix commands conform to the [XDG Base Directory Specification], and use the following environment variables to determine locations of various state and configuration files: + +- [`XDG_CONFIG_HOME`]{#env-XDG_CONFIG_HOME} (default `~/.config`) +- [`XDG_STATE_HOME`]{#env-XDG_STATE_HOME} (default `~/.local/state`) +- [`XDG_CACHE_HOME`]{#env-XDG_CACHE_HOME} (default `~/.cache`) + +Classic Nix commands can also be made to follow this standard using the [`use-xdg-base-directories`] configuration option. + +[XDG Base Directory Specification]: https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html +[`use-xdg-base-directories`]: ../command-ref/conf-file.md#conf-use-xdg-base-directories
\ No newline at end of file diff --git a/doc/manual/src/command-ref/nix-build.md b/doc/manual/src/command-ref/nix-build.md index 49c6f3f55..937b046b8 100644 --- a/doc/manual/src/command-ref/nix-build.md +++ b/doc/manual/src/command-ref/nix-build.md @@ -37,10 +37,12 @@ directory containing at least a file named `default.nix`. `nix-build` is essentially a wrapper around [`nix-instantiate`](nix-instantiate.md) (to translate a high-level Nix -expression to a low-level store derivation) and [`nix-store +expression to a low-level [store derivation]) and [`nix-store --realise`](nix-store.md#operation---realise) (to build the store derivation). +[store derivation]: ../glossary.md#gloss-store-derivation + > **Warning** > > The result of the build is automatically registered as a root of the @@ -53,16 +55,18 @@ All options not listed here are passed to `nix-store --realise`, except for `--arg` and `--attr` / `-A` which are passed to `nix-instantiate`. - - [`--no-out-link`]{#opt-no-out-link}\ + - <span id="opt-no-out-link">[`--no-out-link`](#opt-no-out-link)<span> + Do not create a symlink to the output path. Note that as a result the output does not become a root of the garbage collector, and so - might be deleted by `nix-store - --gc`. + might be deleted by `nix-store --gc`. + + - <span id="opt-dry-run">[`--dry-run`](#opt-dry-run)</span> - - [`--dry-run`]{#opt-dry-run}\ Show what store paths would be built or downloaded. - - [`--out-link`]{#opt-out-link} / `-o` *outlink*\ + - <span id="opt-out-link">[`--out-link`](#opt-out-link)</span> / `-o` *outlink* + Change the name of the symlink to the output path created from `result` to *outlink*. diff --git a/doc/manual/src/command-ref/nix-copy-closure.md b/doc/manual/src/command-ref/nix-copy-closure.md index 7047d3012..cd8e351bb 100644 --- a/doc/manual/src/command-ref/nix-copy-closure.md +++ b/doc/manual/src/command-ref/nix-copy-closure.md @@ -30,8 +30,8 @@ Since `nix-copy-closure` calls `ssh`, you may be asked to type in the appropriate password or passphrase. In fact, you may be asked _twice_ because `nix-copy-closure` currently connects twice to the remote machine, first to get the set of paths missing on the target machine, -and second to send the dump of those paths. If this bothers you, use -`ssh-agent`. +and second to send the dump of those paths. When using public key +authentication, you can avoid typing the passphrase with `ssh-agent`. # Options @@ -47,7 +47,9 @@ and second to send the dump of those paths. If this bothers you, use Enable compression of the SSH connection. - `--include-outputs`\ - Also copy the outputs of store derivations included in the closure. + Also copy the outputs of [store derivation]s included in the closure. + + [store derivation]: ../glossary.md#gloss-store-derivation - `--use-substitutes` / `-s`\ Attempt to download missing paths on the target machine using Nix’s diff --git a/doc/manual/src/command-ref/nix-daemon.md b/doc/manual/src/command-ref/nix-daemon.md index e91cb01dd..04b483be3 100644 --- a/doc/manual/src/command-ref/nix-daemon.md +++ b/doc/manual/src/command-ref/nix-daemon.md @@ -8,6 +8,6 @@ # Description -The Nix daemon is necessary in multi-user Nix installations. It performs -build actions and other operations on the Nix store on behalf of +The Nix daemon is necessary in multi-user Nix installations. It runs +build tasks and other operations on the Nix store on behalf of unprivileged users. diff --git a/doc/manual/src/command-ref/nix-env.md b/doc/manual/src/command-ref/nix-env.md index a372c5eae..f4fa5b50c 100644 --- a/doc/manual/src/command-ref/nix-env.md +++ b/doc/manual/src/command-ref/nix-env.md @@ -198,17 +198,19 @@ a number of possible ways: another. - If `--from-expression` is given, *args* are Nix - [functions](../expressions/language-constructs.md#functions) + [functions](../language/constructs.md#functions) that are called with the active Nix expression as their single argument. The derivations returned by those function calls are installed. This allows derivations to be specified in an unambiguous way, which is necessary if there are multiple derivations with the same name. - - If *args* are store derivations, then these are + - If *args* are [store derivation]s, then these are [realised](nix-store.md#operation---realise), and the resulting output paths are installed. + [store derivation]: ../glossary.md#gloss-store-derivation + - If *args* are store paths that are not store derivations, then these are [realised](nix-store.md#operation---realise) and installed. @@ -280,7 +282,7 @@ To copy the store path with symbolic name `gcc` from another profile: $ nix-env -i --from-profile /nix/var/nix/profiles/foo gcc ``` -To install a specific store derivation (typically created by +To install a specific [store derivation] (typically created by `nix-instantiate`): ```console @@ -665,7 +667,7 @@ derivation is shown unless `--no-name` is specified. Print the `system` attribute of the derivation. - `--drv-path`\ - Print the path of the store derivation. + Print the path of the [store derivation]. - `--out-path`\ Print the output path of the derivation. diff --git a/doc/manual/src/command-ref/nix-hash.md b/doc/manual/src/command-ref/nix-hash.md index 45f67f1c5..90fedfb14 100644 --- a/doc/manual/src/command-ref/nix-hash.md +++ b/doc/manual/src/command-ref/nix-hash.md @@ -6,9 +6,7 @@ `nix-hash` [`--flat`] [`--base32`] [`--truncate`] [`--type` *hashAlgo*] *path…* -`nix-hash` `--to-base16` *hash…* - -`nix-hash` `--to-base32` *hash…* +`nix-hash` [`--to-base16`|`--to-base32`|`--to-base64`|`--to-sri`] [`--type` *hashAlgo*] *hash…* # Description @@ -35,11 +33,23 @@ md5sum`. The result is identical to that produced by the GNU commands `md5sum` and `sha1sum`. + - `--base16`\ + Print the hash in a hexadecimal representation (default). + - `--base32`\ Print the hash in a base-32 representation rather than hexadecimal. This base-32 representation is more compact and can be used in Nix expressions (such as in calls to `fetchurl`). + - `--base64`\ + Similar to --base32, but print the hash in a base-64 representation, + which is more compact than the base-32 one. + + - `--sri`\ + Print the hash in SRI format with base-64 encoding. + The type of hash algorithm will be prepended to the hash string, + followed by a hyphen (-) and the base-64 hash body. + - `--truncate`\ Truncate hashes longer than 160 bits (such as SHA-256) to 160 bits. @@ -55,6 +65,14 @@ md5sum`. Don’t hash anything, but convert the hexadecimal hash representation *hash* to base-32. + - `--to-base64`\ + Don’t hash anything, but convert the hexadecimal hash representation + *hash* to base-64. + + - `--to-sri`\ + Don’t hash anything, but convert the hexadecimal hash representation + *hash* to SRI. + # Examples Computing the same hash as `nix-prefetch-url`: @@ -81,9 +99,18 @@ $ nix-store --dump test/ | md5sum (for comparison) $ nix-hash --type sha1 test/ e4fd8ba5f7bbeaea5ace89fe10255536cd60dab6 +$ nix-hash --type sha1 --base16 test/ +e4fd8ba5f7bbeaea5ace89fe10255536cd60dab6 + $ nix-hash --type sha1 --base32 test/ nvd61k9nalji1zl9rrdfmsmvyyjqpzg4 +$ nix-hash --type sha1 --base64 test/ +5P2Lpfe76upazon+ECVVNs1g2rY= + +$ nix-hash --type sha1 --sri test/ +sha1-5P2Lpfe76upazon+ECVVNs1g2rY= + $ nix-hash --type sha256 --flat test/ error: reading file `test/': Is a directory @@ -91,7 +118,7 @@ $ nix-hash --type sha256 --flat test/world 5891b5b522d5df086d0ff0b110fbd9d21bb4fc7163af34d08286a2e846f6be03 ``` -Converting between hexadecimal and base-32: +Converting between hexadecimal, base-32, base-64, and SRI: ```console $ nix-hash --type sha1 --to-base32 e4fd8ba5f7bbeaea5ace89fe10255536cd60dab6 @@ -99,4 +126,13 @@ nvd61k9nalji1zl9rrdfmsmvyyjqpzg4 $ nix-hash --type sha1 --to-base16 nvd61k9nalji1zl9rrdfmsmvyyjqpzg4 e4fd8ba5f7bbeaea5ace89fe10255536cd60dab6 + +$ nix-hash --type sha1 --to-base64 e4fd8ba5f7bbeaea5ace89fe10255536cd60dab6 +5P2Lpfe76upazon+ECVVNs1g2rY= + +$ nix-hash --type sha1 --to-sri nvd61k9nalji1zl9rrdfmsmvyyjqpzg4 +sha1-5P2Lpfe76upazon+ECVVNs1g2rY= + +$ nix-hash --to-base16 sha1-5P2Lpfe76upazon+ECVVNs1g2rY= +e4fd8ba5f7bbeaea5ace89fe10255536cd60dab6 ``` diff --git a/doc/manual/src/command-ref/nix-instantiate.md b/doc/manual/src/command-ref/nix-instantiate.md index 2e198daed..432fb2608 100644 --- a/doc/manual/src/command-ref/nix-instantiate.md +++ b/doc/manual/src/command-ref/nix-instantiate.md @@ -17,13 +17,14 @@ # Description -The command `nix-instantiate` generates [store -derivations](../glossary.md) from (high-level) Nix expressions. It -evaluates the Nix expressions in each of *files* (which defaults to +The command `nix-instantiate` produces [store derivation]s from (high-level) Nix expressions. +It evaluates the Nix expressions in each of *files* (which defaults to *./default.nix*). Each top-level expression should evaluate to a derivation, a list of derivations, or a set of derivations. The paths of the resulting store derivations are printed on standard output. +[store derivation]: ../glossary.md#gloss-store-derivation + If *files* is the character `-`, then a Nix expression will be read from standard input. @@ -51,7 +52,7 @@ standard input. - `--strict`\ When used with `--eval`, recursively evaluate list elements and attributes. Normally, such sub-expressions are left unevaluated - (since the Nix expression language is lazy). + (since the Nix language is lazy). > **Warning** > @@ -66,7 +67,7 @@ standard input. When used with `--eval`, print the resulting value as an XML representation of the abstract syntax tree rather than as an ATerm. The schema is the same as that used by the [`toXML` - built-in](../expressions/builtins.md). + built-in](../language/builtins.md). - `--read-write-mode`\ When used with `--eval`, perform evaluation in read/write mode so @@ -79,8 +80,7 @@ standard input. # Examples -Instantiating store derivations from a Nix expression, and building them -using `nix-store`: +Instantiate [store derivation]s from a Nix expression, and build them using `nix-store`: ```console $ nix-instantiate test.nix (instantiate) diff --git a/doc/manual/src/command-ref/nix-store.md b/doc/manual/src/command-ref/nix-store.md index dc8faba68..17cbd0461 100644 --- a/doc/manual/src/command-ref/nix-store.md +++ b/doc/manual/src/command-ref/nix-store.md @@ -22,7 +22,8 @@ This section lists the options that are common to all operations. These options are allowed for every subcommand, though they may not always have an effect. - - [`--add-root`]{#opt-add-root} *path*\ + - <span id="opt-add-root">[`--add-root`](#opt-add-root)</span> *path* + Causes the result of a realisation (`--realise` and `--force-realise`) to be registered as a root of the garbage collector. *path* will be created as a symlink to the resulting @@ -53,6 +54,11 @@ have an effect. created by sequentially numbering symlinks beyond the first one (e.g., `foo`, `foo-2`, `foo-3`, and so on). + - <span id="opt-stdin">[`--stdin`](#opt-stdin)</span> + + Read *paths…* from the standard input. + Useful for chaining nix-store commands. + # Operation `--realise` ## Synopsis @@ -65,13 +71,13 @@ The operation `--realise` essentially “builds” the specified store paths. Realisation is a somewhat overloaded term: - If the store path is a *derivation*, realisation ensures that the - output paths of the derivation are [valid](../glossary.md) (i.e., + output paths of the derivation are [valid] (i.e., the output path and its closure exist in the file system). This can be done in several ways. First, it is possible that the outputs are already valid, in which case we are done - immediately. Otherwise, there may be [substitutes](../glossary.md) + immediately. Otherwise, there may be [substitutes] that produce the outputs (e.g., by downloading them). Finally, the - outputs can be produced by performing the build action described + outputs can be produced by running the build task described by the derivation. - If the store path is not a derivation, realisation ensures that the @@ -81,6 +87,9 @@ paths. Realisation is a somewhat overloaded term: produced through substitutes. If there are no (successful) substitutes, realisation fails. +[valid]: ../glossary.md#gloss-validity +[substitutes]: ../glossary.md#gloss-substitute + The output path of each derivation is printed on standard output. (For non-derivations argument, the argument itself is printed.) @@ -104,10 +113,6 @@ The following flags are available: previous build, the new output path is left in `/nix/store/name.check.` - See also the `build-repeat` configuration option, which repeats a - derivation a number of times and prevents its outputs from being - registered as “valid” in the Nix store unless they are identical. - Special exit codes: - `100`\ @@ -121,7 +126,7 @@ Special exit codes: - `102`\ Hash mismatch, the build output was rejected because it does not match the [`outputHash` attribute of the - derivation](../expressions/advanced-attributes.md). + derivation](../language/advanced-attributes.md). - `104`\ Not deterministic, the build succeeded in check mode but the @@ -140,8 +145,10 @@ or. ## Examples -This operation is typically used to build store derivations produced by -[`nix-instantiate`](nix-instantiate.md): +This operation is typically used to build [store derivation]s produced by +[`nix-instantiate`](./nix-instantiate.md): + +[store derivation]: ../glossary.md#gloss-store-derivation ```console $ nix-store -r $(nix-instantiate ./test.nix) @@ -156,6 +163,12 @@ To test whether a previously-built derivation is deterministic: $ nix-build '<nixpkgs>' -A hello --check -K ``` +Use [`--read-log`](#operation---read-log) to show the stderr and stdout of a build: + +```console +$ nix-store --read-log $(nix-instantiate ./test.nix) +``` + # Operation `--serve` ## Synopsis @@ -290,8 +303,8 @@ error: cannot delete path `/nix/store/zq0h41l75vlb4z45kzgjjmsjxvcv1qk7-mesa-6.4' ## Description -The operation `--query` displays various bits of information about the -store paths . The queries are described below. At most one query can be +The operation `--query` displays information about [store path]s. +The queries are described below. At most one query can be specified. The default query is `--outputs`. The paths *paths* may also be symlinks from outside of the Nix store, to @@ -301,7 +314,7 @@ symlink. ## Common query options - `--use-output`; `-u`\ - For each argument to the query that is a store derivation, apply the + For each argument to the query that is a [store derivation], apply the query to the output path of the derivation instead. - `--force-realise`; `-f`\ @@ -311,17 +324,17 @@ symlink. ## Queries - `--outputs`\ - Prints out the [output paths](../glossary.md) of the store + Prints out the [output path]s of the store derivations *paths*. These are the paths that will be produced when the derivation is built. - `--requisites`; `-R`\ - Prints out the [closure](../glossary.md) of the store path *paths*. + Prints out the [closure] of the given *paths*. This query has one option: - `--include-outputs` - Also include the existing output paths of store derivations, + Also include the existing output paths of [store derivation]s, and their closures. This query can be used to implement various kinds of deployment. A @@ -333,10 +346,12 @@ symlink. derivation and specifying the option `--include-outputs`. - `--references`\ - Prints the set of [references](../glossary.md) of the store paths + Prints the set of [references]s of the store paths *paths*, that is, their immediate dependencies. (For *all* dependencies, use `--requisites`.) + [reference]: ../glossary.md#gloss-reference + - `--referrers`\ Prints the set of *referrers* of the store paths *paths*, that is, the store paths currently existing in the Nix store that refer to @@ -351,11 +366,13 @@ symlink. in the Nix store that are dependent on *paths*. - `--deriver`; `-d`\ - Prints the [deriver](../glossary.md) of the store paths *paths*. If + Prints the [deriver] of the store paths *paths*. If the path has no deriver (e.g., if it is a source file), or if the deriver is not known (e.g., in the case of a binary-only deployment), the string `unknown-deriver` is printed. + [deriver]: ../glossary.md#gloss-deriver + - `--graph`\ Prints the references graph of the store paths *paths* in the format of the `dot` tool of AT\&T's [Graphviz @@ -375,12 +392,12 @@ symlink. Prints the references graph of the store paths *paths* in the [GraphML](http://graphml.graphdrawing.org/) file format. This can be used to visualise dependency graphs. To obtain a build-time - dependency graph, apply this to a store derivation. To obtain a + dependency graph, apply this to a [store derivation]. To obtain a runtime dependency graph, apply it to an output path. - `--binding` *name*; `-b` *name*\ Prints the value of the attribute *name* (i.e., environment - variable) of the store derivations *paths*. It is an error for a + variable) of the [store derivation]s *paths*. It is an error for a derivation to not have the specified attribute. - `--hash`\ @@ -621,7 +638,7 @@ written to standard output. A NAR archive is like a TAR or Zip archive, but it contains only the information that Nix considers important. For instance, timestamps are -elided because all files in the Nix store have their timestamp set to 0 +elided because all files in the Nix store have their timestamp set to 1 anyway. Likewise, all permissions are left out except for the execute bit, because all files in the Nix store have 444 or 555 permission. diff --git a/doc/manual/src/command-ref/opt-common.md b/doc/manual/src/command-ref/opt-common.md index 51d7de18a..e612c416f 100644 --- a/doc/manual/src/command-ref/opt-common.md +++ b/doc/manual/src/command-ref/opt-common.md @@ -145,7 +145,7 @@ Most Nix commands accept the following command-line options: expression evaluator will automatically try to call functions that it encounters. It can automatically call functions for which every argument has a [default - value](../expressions/language-constructs.md#functions) (e.g., + value](../language/constructs.md#functions) (e.g., `{ argName ? defaultValue }: ...`). With `--arg`, you can also call functions that have arguments without a default value (or override a default value). That is, if the evaluator encounters a @@ -164,7 +164,7 @@ Most Nix commands accept the following command-line options: So if you call this Nix expression (e.g., when you do `nix-env -iA pkgname`), the function will be called automatically using the - value [`builtins.currentSystem`](../expressions/builtins.md) for + value [`builtins.currentSystem`](../language/builtins.md) for the `system` argument. You can override this using `--arg`, e.g., `nix-env -iA pkgname --arg system \"i686-freebsd\"`. (Note that since the argument is a Nix string literal, you have to escape the diff --git a/doc/manual/src/contributing/cli-guideline.md b/doc/manual/src/contributing/cli-guideline.md index 01a1b1e73..e53d2d178 100644 --- a/doc/manual/src/contributing/cli-guideline.md +++ b/doc/manual/src/contributing/cli-guideline.md @@ -389,6 +389,88 @@ colors, no emojis and using ASCII instead of Unicode symbols). The same should happen when TTY is not detected on STDERR. We should not display progress / status section, but only print warnings and errors. +## Returning future proof JSON + +The schema of JSON output should allow for backwards compatible extension. This section explains how to achieve this. + +Two definitions are helpful here, because while JSON only defines one "key-value" +object type, we use it to cover two use cases: + + - **dictionary**: a map from names to value that all have the same type. In + C++ this would be a `std::map` with string keys. + - **record**: a fixed set of attributes each with their own type. In C++, this + would be represented by a `struct`. + +It is best not to mix these use cases, as that may lead to incompatibilities when the schema changes. For example, adding a record field to a dictionary breaks consumers that assume all JSON object fields to have the same meaning and type. + +This leads to the following guidelines: + + - The top-level (root) value must be a record. + + Otherwise, one can not change the structure of a command's output. + + - The value of a dictionary item must be a record. + + Otherwise, the item type can not be extended. + + - List items should be records. + + Otherwise, one can not change the structure of the list items. + + If the order of the items does not matter, and each item has a unique key that is a string, consider representing the list as a dictionary instead. If the order of the items needs to be preserved, return a list of records. + + - Streaming JSON should return records. + + An example of a streaming JSON format is [JSON lines](https://jsonlines.org/), where each line represents a JSON value. These JSON values can be considered top-level values or list items, and they must be records. + +### Examples + + +This is bad, because all keys must be assumed to be store implementations: + +```json +{ + "local": { ... }, + "remote": { ... }, + "http": { ... } +} +``` + +This is good, because the it is extensible at the root, and is somewhat self-documenting: + +```json +{ + "storeTypes": { "local": { ... }, ... }, + "pluginSupport": true +} +``` + +While the dictionary of store types seems like a very complete response at first, a use case may arise that warrants returning additional information. +For example, the presence of plugin support may be crucial information for a client to proceed when their desired store type is missing. + + + +The following representation is bad because it is not extensible: + +```json +{ "outputs": [ "out" "bin" ] } +``` + +However, simply converting everything to records is not enough, because the order of outputs must be preserved: + +```json +{ "outputs": { "bin": {}, "out": {} } } +``` + +The first item is the default output. Deriving this information from the outputs ordering is not great, but this is how Nix currently happens to work. +While it is possible for a JSON parser to preserve the order of fields, we can not rely on this capability to be present in all JSON libraries. + +This representation is extensible and preserves the ordering: + +```json +{ "outputs": [ { "outputName": "out" }, { "outputName": "bin" } ] } +``` + ## Dialog with the user CLIs don't always make it clear when an action has taken place. For every diff --git a/doc/manual/src/contributing/hacking.md b/doc/manual/src/contributing/hacking.md index 59ce5cac7..ca69f076a 100644 --- a/doc/manual/src/contributing/hacking.md +++ b/doc/manual/src/contributing/hacking.md @@ -8,52 +8,83 @@ $ git clone https://github.com/NixOS/nix.git $ cd nix ``` -To build Nix for the current operating system/architecture use +The following instructions assume you already have some version of Nix installed locally, so that you can use it to set up the development environment. If you don't have it installed, follow the [installation instructions]. + +[installation instructions]: ../installation/installation.md + +## Nix with flakes + +This section assumes you are using Nix with [flakes] enabled. See the [next section](#classic-nix) for equivalent instructions which don't require flakes. + +[flakes]: ../command-ref/new-cli/nix3-flake.md#description + +To build all dependencies and start a shell in which all environment +variables are set up so that those dependencies can be found: ```console -$ nix-build +$ nix develop ``` -or if you have a flake-enabled nix: +This shell also adds `./outputs/bin/nix` to your `$PATH` so you can run `nix` immediately after building it. + +To get a shell with one of the other [supported compilation environments](#compilation-environments): ```console -$ nix build +$ nix develop .#native-clang11StdenvPackages ``` -This will build `defaultPackage` attribute defined in the `flake.nix` -file. To build for other platforms add one of the following suffixes to -it: aarch64-linux, i686-linux, x86\_64-darwin, x86\_64-linux. i.e. +> **Note** +> +> Use `ccacheStdenv` to drastically improve rebuild time. +> By default, [ccache](https://ccache.dev) keeps artifacts in `~/.cache/ccache/`. + +To build Nix itself in this shell: ```console -$ nix-build -A defaultPackage.x86_64-linux +[nix-shell]$ ./bootstrap.sh +[nix-shell]$ ./configure $configureFlags --prefix=$(pwd)/outputs/out +[nix-shell]$ make -j $NIX_BUILD_CORES ``` -To build all dependencies and start a shell in which all environment -variables are set up so that those dependencies can be found: +To install it in `$(pwd)/outputs` and test it: ```console -$ nix-shell +[nix-shell]$ make install +[nix-shell]$ make installcheck -j $NIX_BUILD_CORES +[nix-shell]$ nix --version +nix (Nix) 2.12 ``` -or if you have a flake-enabled nix: +To build a release version of Nix: ```console -$ nix develop +$ nix build ``` -To get a shell with a different compilation environment (e.g. stdenv, -gccStdenv, clangStdenv, clang11Stdenv): +You can also build Nix for one of the [supported target platforms](#target-platforms). + +## Classic Nix + +This section is for Nix without [flakes]. + +To build all dependencies and start a shell in which all environment +variables are set up so that those dependencies can be found: ```console -$ nix-shell -A devShells.x86_64-linux.clang11StdenvPackages +$ nix-shell ``` -or if you have a flake-enabled nix: +To get a shell with one of the other [supported compilation environments](#compilation-environments): ```console -$ nix develop .#clang11StdenvPackages +$ nix-shell -A devShells.x86_64-linux.native-clang11StdenvPackages ``` +> **Note** +> +> You can use `native-ccacheStdenvPackages` to drastically improve rebuild time. +> By default, [ccache](https://ccache.dev) keeps artifacts in `~/.cache/ccache/`. + To build Nix itself in this shell: ```console @@ -68,38 +99,186 @@ To install it in `$(pwd)/outputs` and test it: [nix-shell]$ make install [nix-shell]$ make installcheck -j $NIX_BUILD_CORES [nix-shell]$ ./outputs/out/bin/nix --version -nix (Nix) 3.0 +nix (Nix) 2.12 ``` -If you have a flakes-enabled Nix you can replace: +To build Nix for the current operating system and CPU architecture use ```console -$ nix-shell +$ nix-build ``` -by: +You can also build Nix for one of the [supported target platforms](#target-platforms). + +## Platforms + +As specified in [`flake.nix`], Nix can be built for various platforms: + +- `aarch64-linux` +- `i686-linux` +- `x86_64-darwin` +- `x86_64-linux` + +[`flake.nix`]: https://github.com/nixos/nix/blob/master/flake.nix + +In order to build Nix for a different platform than the one you're currently +on, you need to have some way for your system Nix to build code for that +platform. Common solutions include [remote builders] and [binfmt emulation] +(only supported on NixOS). + +[remote builders]: ../advanced-topics/distributed-builds.md +[binfmt emulation]: https://nixos.org/manual/nixos/stable/options.html#opt-boot.binfmt.emulatedSystems + +These solutions let Nix perform builds as if you're on the native platform, so +executing the build is as simple as ```console -$ nix develop +$ nix build .#packages.aarch64-linux.default +``` + +for flake-enabled Nix, or + +```console +$ nix-build -A packages.aarch64-linux.default ``` -## Testing +for classic Nix. + +You can use any of the other supported platforms in place of `aarch64-linux`. + +Cross-compiled builds are available for ARMv6 and ARMv7, and Nix on unsupported platforms can be bootstrapped by adding more `crossSystems` in `flake.nix`. + +## Compilation environments + +Nix can be compiled using multiple environments: + +- `stdenv`: default; +- `gccStdenv`: force the use of `gcc` compiler; +- `clangStdenv`: force the use of `clang` compiler; +- `ccacheStdenv`: enable [ccache], a compiler cache to speed up compilation. + +To build with one of those environments, you can use + +```console +$ nix build .#nix-ccacheStdenv +``` -Nix comes with three different flavors of tests: unit, functional and integration. +for flake-enabled Nix, or + +```console +$ nix-build -A nix-ccacheStdenv +``` + +for classic Nix. + +You can use any of the other supported environments in place of `nix-ccacheStdenv`. + +## Editor integration + +The `clangd` LSP server is installed by default on the `clang`-based `devShell`s. +See [supported compilation environments](#compilation-environments) and instructions how to set up a shell [with flakes](#nix-with-flakes) or in [classic Nix](#classic-nix). + +To use the LSP with your editor, you first need to [set up `clangd`](https://clangd.llvm.org/installation#project-setup) by running: + +```console +make clean && bear -- make -j$NIX_BUILD_CORES install +``` + +Configure your editor to use the `clangd` from the shell, either by running it inside the development shell, or by using [nix-direnv](https://github.com/nix-community/nix-direnv) and [the appropriate editor plugin](https://github.com/direnv/direnv/wiki#editor-integration). + +> **Note** +> +> For some editors (e.g. Visual Studio Code), you may need to install a [special extension](https://open-vsx.org/extension/llvm-vs-code-extensions/vscode-clangd) for the editor to interact with `clangd`. +> Some other editors (e.g. Emacs, Vim) need a plugin to support LSP servers in general (e.g. [lsp-mode](https://github.com/emacs-lsp/lsp-mode) for Emacs and [vim-lsp](https://github.com/prabirshrestha/vim-lsp) for vim). +> Editor-specific setup is typically opinionated, so we will not cover it here in more detail. + +## Running tests ### Unit-tests The unit-tests for each Nix library (`libexpr`, `libstore`, etc..) are defined under `src/{library_name}/tests` using the -[googletest](https://google.github.io/googletest/) framework. +[googletest](https://google.github.io/googletest/) and +[rapidcheck](https://github.com/emil-e/rapidcheck) frameworks. You can run the whole testsuite with `make check`, or the tests for a specific component with `make libfoo-tests_RUN`. Finer-grained filtering is also possible using the [--gtest_filter](https://google.github.io/googletest/advanced.html#running-a-subset-of-the-tests) command-line option. ### Functional tests The functional tests reside under the `tests` directory and are listed in `tests/local.mk`. -The whole testsuite can be run with `make install && make installcheck`. -Individual tests can be run with `make tests/{testName}.sh.test`. +Each test is a bash script. + +The whole test suite can be run with: + +```shell-session +$ make install && make installcheck +ran test tests/foo.sh... [PASS] +ran test tests/bar.sh... [PASS] +... +``` + +Individual tests can be run with `make`: + +```shell-session +$ make tests/${testName}.sh.test +ran test tests/${testName}.sh... [PASS] +``` + +or without `make`: + +```shell-session +$ ./mk/run-test.sh tests/${testName}.sh +ran test tests/${testName}.sh... [PASS] +``` + +To see the complete output, one can also run: + +```shell-session +$ ./mk/debug-test.sh tests/${testName}.sh ++ foo +output from foo ++ bar +output from bar +... +``` + +The test script will then be traced with `set -x` and the output displayed as it happens, regardless of whether the test succeeds or fails. + +#### Debugging failing functional tests + +When a functional test fails, it usually does so somewhere in the middle of the script. + +To figure out what's wrong, it is convenient to run the test regularly up to the failing `nix` command, and then run that command with a debugger like GDB. + +For example, if the script looks like: + +```bash +foo +nix blah blub +bar +``` +edit it like so: + +```diff + foo +-nix blah blub ++gdb --args nix blah blub + bar +``` + +Then, running the test with `./mk/debug-test.sh` will drop you into GDB once the script reaches that point: + +```shell-session +$ ./mk/debug-test.sh tests/${testName}.sh +... ++ gdb blash blub +GNU gdb (GDB) 12.1 +... +(gdb) +``` + +One can debug the Nix invocation in all the usual ways. +For example, enter `run` to start the Nix invocation. ### Integration tests @@ -108,3 +287,137 @@ These tests include everything that needs to interact with external services or Because these tests are expensive and require more than what the standard github-actions setup provides, they only run on the master branch (on <https://hydra.nixos.org/jobset/nix/master>). You can run them manually with `nix build .#hydraJobs.tests.{testName}` or `nix-build -A hydraJobs.tests.{testName}` + +### Installer tests + +After a one-time setup, the Nix repository's GitHub Actions continuous integration (CI) workflow can test the installer each time you push to a branch. + +Creating a Cachix cache for your installer tests and adding its authorization token to GitHub enables [two installer-specific jobs in the CI workflow](https://github.com/NixOS/nix/blob/88a45d6149c0e304f6eb2efcc2d7a4d0d569f8af/.github/workflows/ci.yml#L50-L91): + +- The `installer` job generates installers for the platforms below and uploads them to your Cachix cache: + - `x86_64-linux` + - `armv6l-linux` + - `armv7l-linux` + - `x86_64-darwin` + +- The `installer_test` job (which runs on `ubuntu-latest` and `macos-latest`) will try to install Nix with the cached installer and run a trivial Nix command. + +#### One-time setup + +1. Have a GitHub account with a fork of the [Nix repository](https://github.com/NixOS/nix). +2. At cachix.org: + - Create or log in to an account. + - Create a Cachix cache using the format `<github-username>-nix-install-tests`. + - Navigate to the new cache > Settings > Auth Tokens. + - Generate a new Cachix auth token and copy the generated value. +3. At github.com: + - Navigate to your Nix fork > Settings > Secrets > Actions > New repository secret. + - Name the secret `CACHIX_AUTH_TOKEN`. + - Paste the copied value of the Cachix cache auth token. + +#### Using the CI-generated installer for manual testing + +After the CI run completes, you can check the output to extract the installer URL: +1. Click into the detailed view of the CI run. +2. Click into any `installer_test` run (the URL you're here to extract will be the same in all of them). +3. Click into the `Run cachix/install-nix-action@v...` step and click the detail triangle next to the first log line (it will also be `Run cachix/install-nix-action@v...`) +4. Copy the value of `install_url` +5. To generate an install command, plug this `install_url` and your GitHub username into this template: + + ```console + curl -L <install_url> | sh -s -- --tarball-url-prefix https://<github-username>-nix-install-tests.cachix.org/serve + ``` + +<!-- #### Manually generating test installers + +There's obviously a manual way to do this, and it's still the only way for +platforms that lack GA runners. + +I did do this back in Fall 2020 (before the GA approach encouraged here). I'll +sketch what I recall in case it encourages someone to fill in detail, but: I +didn't know what I was doing at the time and had to fumble/ask around a lot-- +so I don't want to uphold any of it as "right". It may have been dumb or +the _hard_ way from the getgo. Fundamentals may have changed since. + +Here's the build command I used to do this on and for x86_64-darwin: +nix build --out-link /tmp/foo ".#checks.x86_64-darwin.binaryTarball" + +I used the stable out-link to make it easier to script the next steps: +link=$(readlink /tmp/foo) +cp $link/*-darwin.tar.xz ~/somewheres + +I've lost the last steps and am just going from memory: + +From here, I think I had to extract and modify the `install` script to point +it at this tarball (which I scped to my own site, but it might make more sense +to just share them locally). I extracted this script once and then just +search/replaced in it for each new build. + +The installer now supports a `--tarball-url-prefix` flag which _may_ have +solved this need? +--> + +### Checking links in the manual + +The build checks for broken internal links. +This happens late in the process, so `nix build` is not suitable for iterating. +To build the manual incrementally, run: + +```console +make html -j $NIX_BUILD_CORES +``` + +In order to reflect changes to the [Makefile], clear all generated files before re-building: + +[Makefile]: https://github.com/NixOS/nix/blob/master/doc/manual/local.mk + +```console +rm $(git ls-files doc/manual/ -o | grep -F '.md') && rmdir doc/manual/src/command-ref/new-cli && make html -j $NIX_BUILD_CORES +``` + +[`mdbook-linkcheck`] does not implement checking [URI fragments] yet. + +[`mdbook-linkcheck`]: https://github.com/Michael-F-Bryan/mdbook-linkcheck +[URI fragments]: https://en.m.wikipedia.org/wiki/URI_fragment + +#### `@docroot@` variable + +`@docroot@` provides a base path for links that occur in reusable snippets or other documentation that doesn't have a base path of its own. + +If a broken link occurs in a snippet that was inserted into multiple generated files in different directories, use `@docroot@` to reference the `doc/manual/src` directory. + +If the `@docroot@` literal appears in an error message from the `mdbook-linkcheck` tool, the `@docroot@` replacement needs to be applied to the generated source file that mentions it. +See existing `@docroot@` logic in the [Makefile]. +Regular markdown files used for the manual have a base path of their own and they can use relative paths instead of `@docroot@`. + +## API documentation + +Doxygen API documentation is [available +online](https://hydra.nixos.org/job/nix/master/internal-api-docs/latest/download-by-type/doc/internal-api-docs). You +can also build and view it yourself: + +```console +# nix build .#hydraJobs.internal-api-docs +# xdg-open ./result/share/doc/nix/internal-api/html/index.html +``` + +or inside a `nix develop` shell by running: + +``` +# make internal-api-html +# xdg-open ./outputs/doc/share/doc/nix/internal-api/html/index.html +``` + +## Coverage analysis + +A coverage analysis report is [available +online](https://hydra.nixos.org/job/nix/master/coverage/latest/download-by-type/report/coverage). You +can build it yourself: + +``` +# nix build .#hydraJobs.coverage +# xdg-open ./result/coverage/index.html +``` + +Metrics about the change in line/function coverage over time are also +[available](https://hydra.nixos.org/job/nix/master/coverage#tabs-charts). diff --git a/doc/manual/src/expressions/arguments-variables.md b/doc/manual/src/expressions/arguments-variables.md deleted file mode 100644 index 12198c879..000000000 --- a/doc/manual/src/expressions/arguments-variables.md +++ /dev/null @@ -1,80 +0,0 @@ -# Arguments and Variables - -The [Nix expression for GNU Hello](expression-syntax.md) is a -function; it is missing some arguments that have to be filled in -somewhere. In the Nix Packages collection this is done in the file -`pkgs/top-level/all-packages.nix`, where all Nix expressions for -packages are imported and called with the appropriate arguments. Here -are some fragments of `all-packages.nix`, with annotations of what -they mean: - -```nix -... - -rec { ① - - hello = import ../applications/misc/hello/ex-1 ② { ③ - inherit fetchurl stdenv perl; - }; - - perl = import ../development/interpreters/perl { ④ - inherit fetchurl stdenv; - }; - - fetchurl = import ../build-support/fetchurl { - inherit stdenv; ... - }; - - stdenv = ...; - -} -``` - -1. This file defines a set of attributes, all of which are concrete - derivations (i.e., not functions). In fact, we define a *mutually - recursive* set of attributes. That is, the attributes can refer to - each other. This is precisely what we want since we want to “plug” - the various packages into each other. - -2. Here we *import* the Nix expression for GNU Hello. The import - operation just loads and returns the specified Nix expression. In - fact, we could just have put the contents of the Nix expression - for GNU Hello in `all-packages.nix` at this point. That would be - completely equivalent, but it would make `all-packages.nix` rather - bulky. - - Note that we refer to `../applications/misc/hello/ex-1`, not - `../applications/misc/hello/ex-1/default.nix`. When you try to - import a directory, Nix automatically appends `/default.nix` to the - file name. - -3. This is where the actual composition takes place. Here we *call* the - function imported from `../applications/misc/hello/ex-1` with a set - containing the things that the function expects, namely `fetchurl`, - `stdenv`, and `perl`. We use inherit again to use the attributes - defined in the surrounding scope (we could also have written - `fetchurl = fetchurl;`, etc.). - - The result of this function call is an actual derivation that can be - built by Nix (since when we fill in the arguments of the function, - what we get is its body, which is the call to `stdenv.mkDerivation` - in the [Nix expression for GNU Hello](expression-syntax.md)). - - > **Note** - > - > Nixpkgs has a convenience function `callPackage` that imports and - > calls a function, filling in any missing arguments by passing the - > corresponding attribute from the Nixpkgs set, like this: - > - > ```nix - > hello = callPackage ../applications/misc/hello/ex-1 { }; - > ``` - > - > If necessary, you can set or override arguments: - > - > ```nix - > hello = callPackage ../applications/misc/hello/ex-1 { stdenv = myStdenv; }; - > ``` - -4. Likewise, we have to instantiate Perl, `fetchurl`, and the standard - environment. diff --git a/doc/manual/src/expressions/build-script.md b/doc/manual/src/expressions/build-script.md deleted file mode 100644 index b1eacae88..000000000 --- a/doc/manual/src/expressions/build-script.md +++ /dev/null @@ -1,70 +0,0 @@ -# Build Script - -Here is the builder referenced from Hello's Nix expression (stored in -`pkgs/applications/misc/hello/ex-1/builder.sh`): - -```bash -source $stdenv/setup ① - -PATH=$perl/bin:$PATH ② - -tar xvfz $src ③ -cd hello-* -./configure --prefix=$out ④ -make ⑤ -make install -``` - -The builder can actually be made a lot shorter by using the *generic -builder* functions provided by `stdenv`, but here we write out the build -steps to elucidate what a builder does. It performs the following steps: - -1. When Nix runs a builder, it initially completely clears the - environment (except for the attributes declared in the derivation). - This is done to prevent undeclared inputs from being used in the - build process. If for example the `PATH` contained `/usr/bin`, then - you might accidentally use `/usr/bin/gcc`. - - So the first step is to set up the environment. This is done by - calling the `setup` script of the standard environment. The - environment variable `stdenv` points to the location of the - standard environment being used. (It wasn't specified explicitly - as an attribute in Hello's Nix expression, but `mkDerivation` adds - it automatically.) - -2. Since Hello needs Perl, we have to make sure that Perl is in the - `PATH`. The `perl` environment variable points to the location of - the Perl package (since it was passed in as an attribute to the - derivation), so `$perl/bin` is the directory containing the Perl - interpreter. - -3. Now we have to unpack the sources. The `src` attribute was bound to - the result of fetching the Hello source tarball from the network, so - the `src` environment variable points to the location in the Nix - store to which the tarball was downloaded. After unpacking, we `cd` - to the resulting source directory. - - The whole build is performed in a temporary directory created in - `/tmp`, by the way. This directory is removed after the builder - finishes, so there is no need to clean up the sources afterwards. - Also, the temporary directory is always newly created, so you don't - have to worry about files from previous builds interfering with the - current build. - -4. GNU Hello is a typical Autoconf-based package, so we first have to - run its `configure` script. In Nix every package is stored in a - separate location in the Nix store, for instance - `/nix/store/9a54ba97fb71b65fda531012d0443ce2-hello-2.1.1`. Nix - computes this path by cryptographically hashing all attributes of - the derivation. The path is passed to the builder through the `out` - environment variable. So here we give `configure` the parameter - `--prefix=$out` to cause Hello to be installed in the expected - location. - -5. Finally we build Hello (`make`) and install it into the location - specified by `out` (`make install`). - -If you are wondering about the absence of error checking on the result -of various commands called in the builder: this is because the shell -script is evaluated with Bash's `-e` option, which causes the script to -be aborted if any command fails without an error check. diff --git a/doc/manual/src/expressions/expression-language.md b/doc/manual/src/expressions/expression-language.md deleted file mode 100644 index 267fcb983..000000000 --- a/doc/manual/src/expressions/expression-language.md +++ /dev/null @@ -1,12 +0,0 @@ -# Nix Expression Language - -The Nix expression language is a pure, lazy, functional language. Purity -means that operations in the language don't have side-effects (for -instance, there is no variable assignment). Laziness means that -arguments to functions are evaluated only when they are needed. -Functional means that functions are “normal” values that can be passed -around and manipulated in interesting ways. The language is not a -full-featured, general purpose language. Its main job is to describe -packages, compositions of packages, and the variability within packages. - -This section presents the various features of the language. diff --git a/doc/manual/src/expressions/expression-syntax.md b/doc/manual/src/expressions/expression-syntax.md deleted file mode 100644 index 6b93e692c..000000000 --- a/doc/manual/src/expressions/expression-syntax.md +++ /dev/null @@ -1,93 +0,0 @@ -# Expression Syntax - -Here is a Nix expression for GNU Hello: - -```nix -{ stdenv, fetchurl, perl }: ① - -stdenv.mkDerivation { ② - name = "hello-2.1.1"; ③ - builder = ./builder.sh; ④ - src = fetchurl { ⑤ - url = "ftp://ftp.nluug.nl/pub/gnu/hello/hello-2.1.1.tar.gz"; - sha256 = "1md7jsfd8pa45z73bz1kszpp01yw6x5ljkjk2hx7wl800any6465"; - }; - inherit perl; ⑥ -} -``` - -This file is actually already in the Nix Packages collection in -`pkgs/applications/misc/hello/ex-1/default.nix`. It is customary to -place each package in a separate directory and call the single Nix -expression in that directory `default.nix`. The file has the following -elements (referenced from the figure by number): - -1. This states that the expression is a *function* that expects to be - called with three arguments: `stdenv`, `fetchurl`, and `perl`. They - are needed to build Hello, but we don't know how to build them here; - that's why they are function arguments. `stdenv` is a package that - is used by almost all Nix Packages; it provides a - “standard” environment consisting of the things you would expect - in a basic Unix environment: a C/C++ compiler (GCC, to be precise), - the Bash shell, fundamental Unix tools such as `cp`, `grep`, `tar`, - etc. `fetchurl` is a function that downloads files. `perl` is the - Perl interpreter. - - Nix functions generally have the form `{ x, y, ..., z }: e` where - `x`, `y`, etc. are the names of the expected arguments, and where - *e* is the body of the function. So here, the entire remainder of - the file is the body of the function; when given the required - arguments, the body should describe how to build an instance of - the Hello package. - -2. So we have to build a package. Building something from other stuff - is called a *derivation* in Nix (as opposed to sources, which are - built by humans instead of computers). We perform a derivation by - calling `stdenv.mkDerivation`. `mkDerivation` is a function - provided by `stdenv` that builds a package from a set of - *attributes*. A set is just a list of key/value pairs where each - key is a string and each value is an arbitrary Nix - expression. They take the general form `{ name1 = expr1; ... - nameN = exprN; }`. - -3. The attribute `name` specifies the symbolic name and version of - the package. Nix doesn't really care about these things, but they - are used by for instance `nix-env -q` to show a “human-readable” - name for packages. This attribute is required by `mkDerivation`. - -4. The attribute `builder` specifies the builder. This attribute can - sometimes be omitted, in which case `mkDerivation` will fill in a - default builder (which does a `configure; make; make install`, in - essence). Hello is sufficiently simple that the default builder - would suffice, but in this case, we will show an actual builder - for educational purposes. The value `./builder.sh` refers to the - shell script shown in the [next section](build-script.md), - discussed below. - -5. The builder has to know what the sources of the package are. Here, - the attribute `src` is bound to the result of a call to the - `fetchurl` function. Given a URL and a SHA-256 hash of the expected - contents of the file at that URL, this function builds a derivation - that downloads the file and checks its hash. So the sources are a - dependency that like all other dependencies is built before Hello - itself is built. - - Instead of `src` any other name could have been used, and in fact - there can be any number of sources (bound to different attributes). - However, `src` is customary, and it's also expected by the default - builder (which we don't use in this example). - -6. Since the derivation requires Perl, we have to pass the value of the - `perl` function argument to the builder. All attributes in the set - are actually passed as environment variables to the builder, so - declaring an attribute - - ```nix - perl = perl; - ``` - - will do the trick: it binds an attribute `perl` to the function - argument which also happens to be called `perl`. However, it looks a - bit silly, so there is a shorter syntax. The `inherit` keyword - causes the specified attributes to be bound to whatever variables - with the same name happen to be in scope. diff --git a/doc/manual/src/expressions/generic-builder.md b/doc/manual/src/expressions/generic-builder.md deleted file mode 100644 index cf26b5f82..000000000 --- a/doc/manual/src/expressions/generic-builder.md +++ /dev/null @@ -1,66 +0,0 @@ -# Generic Builder Syntax - -Recall that the [build script for GNU Hello](build-script.md) looked -something like this: - -```bash -PATH=$perl/bin:$PATH -tar xvfz $src -cd hello-* -./configure --prefix=$out -make -make install -``` - -The builders for almost all Unix packages look like this — set up some -environment variables, unpack the sources, configure, build, and -install. For this reason the standard environment provides some Bash -functions that automate the build process. Here is what a builder using -the generic build facilities looks like: - -```bash -buildInputs="$perl" ① - -source $stdenv/setup ② - -genericBuild ③ -``` - -Here is what each line means: - -1. The `buildInputs` variable tells `setup` to use the indicated - packages as “inputs”. This means that if a package provides a `bin` - subdirectory, it's added to `PATH`; if it has a `include` - subdirectory, it's added to GCC's header search path; and so on. - (This is implemented in a modular way: `setup` tries to source the - file `pkg/nix-support/setup-hook` of all dependencies. These “setup - hooks” can then set up whatever environment variables they want; for - instance, the setup hook for Perl sets the `PERL5LIB` environment - variable to contain the `lib/site_perl` directories of all inputs.) - -2. The function `genericBuild` is defined in the file `$stdenv/setup`. - -3. The final step calls the shell function `genericBuild`, which - performs the steps that were done explicitly in the previous build - script. The generic builder is smart enough to figure out whether - to unpack the sources using `gzip`, `bzip2`, etc. It can be - customised in many ways; see the Nixpkgs manual for details. - -Discerning readers will note that the `buildInputs` could just as well -have been set in the Nix expression, like this: - -```nix - buildInputs = [ perl ]; -``` - -The `perl` attribute can then be removed, and the builder becomes even -shorter: - -```bash -source $stdenv/setup -genericBuild -``` - -In fact, `mkDerivation` provides a default builder that looks exactly -like that, so it is actually possible to omit the builder for Hello -entirely. diff --git a/doc/manual/src/expressions/language-operators.md b/doc/manual/src/expressions/language-operators.md deleted file mode 100644 index 268b44f4c..000000000 --- a/doc/manual/src/expressions/language-operators.md +++ /dev/null @@ -1,28 +0,0 @@ -# Operators - -The table below lists the operators in the Nix expression language, in -order of precedence (from strongest to weakest binding). - -| Name | Syntax | Associativity | Description | Precedence | -| ------------------------ | ----------------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------- | -| Select | *e* `.` *attrpath* \[ `or` *def* \] | none | Select attribute denoted by the attribute path *attrpath* from set *e*. (An attribute path is a dot-separated list of attribute names.) If the attribute doesn’t exist, return *def* if provided, otherwise abort evaluation. | 1 | -| Application | *e1* *e2* | left | Call function *e1* with argument *e2*. | 2 | -| Arithmetic Negation | `-` *e* | none | Arithmetic negation. | 3 | -| Has Attribute | *e* `?` *attrpath* | none | Test whether set *e* contains the attribute denoted by *attrpath*; return `true` or `false`. | 4 | -| List Concatenation | *e1* `++` *e2* | right | List concatenation. | 5 | -| Multiplication | *e1* `*` *e2*, | left | Arithmetic multiplication. | 6 | -| Division | *e1* `/` *e2* | left | Arithmetic division. | 6 | -| Addition | *e1* `+` *e2* | left | Arithmetic addition. | 7 | -| Subtraction | *e1* `-` *e2* | left | Arithmetic subtraction. | 7 | -| String Concatenation | *string1* `+` *string2* | left | String concatenation. | 7 | -| Not | `!` *e* | none | Boolean negation. | 8 | -| Update | *e1* `//` *e2* | right | Return a set consisting of the attributes in *e1* and *e2* (with the latter taking precedence over the former in case of equally named attributes). | 9 | -| Less Than | *e1* `<` *e2*, | none | Arithmetic/lexicographic comparison. | 10 | -| Less Than or Equal To | *e1* `<=` *e2* | none | Arithmetic/lexicographic comparison. | 10 | -| Greater Than | *e1* `>` *e2* | none | Arithmetic/lexicographic comparison. | 10 | -| Greater Than or Equal To | *e1* `>=` *e2* | none | Arithmetic/lexicographic comparison. | 10 | -| Equality | *e1* `==` *e2* | none | Equality. | 11 | -| Inequality | *e1* `!=` *e2* | none | Inequality. | 11 | -| Logical AND | *e1* `&&` *e2* | left | Logical AND. | 12 | -| Logical OR | *e1* <code>||</code> *e2* | left | Logical OR. | 13 | -| Logical Implication | *e1* `->` *e2* | none | Logical implication (equivalent to <code>!e1 || e2</code>). | 14 | diff --git a/doc/manual/src/expressions/simple-building-testing.md b/doc/manual/src/expressions/simple-building-testing.md deleted file mode 100644 index 7f0d8f841..000000000 --- a/doc/manual/src/expressions/simple-building-testing.md +++ /dev/null @@ -1,61 +0,0 @@ -# Building and Testing - -You can now try to build Hello. Of course, you could do `nix-env -f . -iA -hello`, but you may not want to install a possibly broken package just -yet. The best way to test the package is by using the command -`nix-build`, which builds a Nix expression and creates a symlink named -`result` in the current directory: - -```console -$ nix-build -A hello -building path `/nix/store/632d2b22514d...-hello-2.1.1' -hello-2.1.1/ -hello-2.1.1/intl/ -hello-2.1.1/intl/ChangeLog -... - -$ ls -l result -lrwxrwxrwx ... 2006-09-29 10:43 result -> /nix/store/632d2b22514d...-hello-2.1.1 - -$ ./result/bin/hello -Hello, world! -``` - -The `-A` option selects the `hello` attribute. This is faster than -using the symbolic package name specified by the `name` attribute -(which also happens to be `hello`) and is unambiguous (there can be -multiple packages with the symbolic name `hello`, but there can be -only one attribute in a set named `hello`). - -`nix-build` registers the `./result` symlink as a garbage collection -root, so unless and until you delete the `./result` symlink, the output -of the build will be safely kept on your system. You can use -`nix-build`’s `-o` switch to give the symlink another name. - -Nix has transactional semantics. Once a build finishes successfully, Nix -makes a note of this in its database: it registers that the path denoted -by `out` is now “valid”. If you try to build the derivation again, Nix -will see that the path is already valid and finish immediately. If a -build fails, either because it returns a non-zero exit code, because Nix -or the builder are killed, or because the machine crashes, then the -output paths will not be registered as valid. If you try to build the -derivation again, Nix will remove the output paths if they exist (e.g., -because the builder died half-way through `make -install`) and try again. Note that there is no “negative caching”: Nix -doesn't remember that a build failed, and so a failed build can always -be repeated. This is because Nix cannot distinguish between permanent -failures (e.g., a compiler error due to a syntax error in the source) -and transient failures (e.g., a disk full condition). - -Nix also performs locking. If you run multiple Nix builds -simultaneously, and they try to build the same derivation, the first Nix -instance that gets there will perform the build, while the others block -(or perform other derivations if available) until the build finishes: - -```console -$ nix-build -A hello -waiting for lock on `/nix/store/0h5b7hp8d4hqfrw8igvx97x1xawrjnac-hello-2.1.1x' -``` - -So it is always safe to run multiple instances of Nix in parallel (which -isn’t the case with, say, `make`). diff --git a/doc/manual/src/expressions/simple-expression.md b/doc/manual/src/expressions/simple-expression.md deleted file mode 100644 index 857f71b9b..000000000 --- a/doc/manual/src/expressions/simple-expression.md +++ /dev/null @@ -1,23 +0,0 @@ -# A Simple Nix Expression - -This section shows how to add and test the [GNU Hello -package](http://www.gnu.org/software/hello/hello.html) to the Nix -Packages collection. Hello is a program that prints out the text “Hello, -world\!”. - -To add a package to the Nix Packages collection, you generally need to -do three things: - -1. Write a Nix expression for the package. This is a file that - describes all the inputs involved in building the package, such as - dependencies, sources, and so on. - -2. Write a *builder*. This is a shell script that builds the package - from the inputs. (In fact, it can be written in any language, but - typically it's a `bash` shell script.) - -3. Add the package to the file `pkgs/top-level/all-packages.nix`. The - Nix expression written in the first step is a *function*; it - requires other packages in order to build it. In this step you put - it all together, i.e., you call the function with the right - arguments to build the actual package. diff --git a/doc/manual/src/expressions/writing-nix-expressions.md b/doc/manual/src/expressions/writing-nix-expressions.md deleted file mode 100644 index 5664108e7..000000000 --- a/doc/manual/src/expressions/writing-nix-expressions.md +++ /dev/null @@ -1,12 +0,0 @@ -This chapter shows you how to write Nix expressions, which instruct Nix -how to build packages. It starts with a simple example (a Nix expression -for GNU Hello), and then moves on to a more in-depth look at the Nix -expression language. - -> **Note** -> -> This chapter is mostly about the Nix expression language. For more -> extensive information on adding packages to the Nix Packages -> collection (such as functions in the standard environment and coding -> conventions), please consult [its -> manual](http://nixos.org/nixpkgs/manual/). diff --git a/doc/manual/src/glossary.md b/doc/manual/src/glossary.md index 3448b971b..c916af4bc 100644 --- a/doc/manual/src/glossary.md +++ b/doc/manual/src/glossary.md @@ -1,26 +1,111 @@ # Glossary - [derivation]{#gloss-derivation}\ - A description of a build action. The result of a derivation is a + A description of a build task. The result of a derivation is a store object. Derivations are typically specified in Nix expressions - using the [`derivation` primitive](expressions/derivations.md). These are + using the [`derivation` primitive](./language/derivations.md). These are translated into low-level *store derivations* (implicitly by `nix-env` and `nix-build`, or explicitly by `nix-instantiate`). + [derivation]: #gloss-derivation + + - [store derivation]{#gloss-store-derivation}\ + A [derivation] represented as a `.drv` file in the [store]. + It has a [store path], like any [store object]. + + Example: `/nix/store/g946hcz4c8mdvq2g8vxx42z51qb71rvp-git-2.38.1.drv` + + See [`nix show-derivation`](./command-ref/new-cli/nix3-show-derivation.md) (experimental) for displaying the contents of store derivations. + + [store derivation]: #gloss-store-derivation + + - [instantiate]{#gloss-instantiate}, instantiation\ + Translate a [derivation] into a [store derivation]. + + See [`nix-instantiate`](./command-ref/nix-instantiate.md). + + [instantiate]: #gloss-instantiate + + - [realise]{#gloss-realise}, realisation\ + Ensure a [store path] is [valid][validity]. + + This means either running the `builder` executable as specified in the corresponding [derivation] or fetching a pre-built [store object] from a [substituter]. + + See [`nix-build`](./command-ref/nix-build.md) and [`nix-store --realise`](./command-ref/nix-store.md#operation---realise). + + See [`nix build`](./command-ref/new-cli/nix3-build.md) (experimental). + + [realise]: #gloss-realise + + - [content-addressed derivation]{#gloss-content-addressed-derivation}\ + A derivation which has the + [`__contentAddressed`](./language/advanced-attributes.md#adv-attr-__contentAddressed) + attribute set to `true`. + + - [fixed-output derivation]{#gloss-fixed-output-derivation}\ + A derivation which includes the + [`outputHash`](./language/advanced-attributes.md#adv-attr-outputHash) attribute. + - [store]{#gloss-store}\ The location in the file system where store objects live. Typically `/nix/store`. + From the perspective of the location where Nix is + invoked, the Nix store can be referred to + as a "_local_" or a "_remote_" one: + + + A *local store* exists on the filesystem of + the machine where Nix is invoked. You can use other + local stores by passing the `--store` flag to the + `nix` command. Local stores can be used for building derivations. + + + A *remote store* exists anywhere other than the + local filesystem. One example is the `/nix/store` + directory on another machine, accessed via `ssh` or + served by the `nix-serve` Perl script. + + [store]: #gloss-store + + - [chroot store]{#gloss-chroot-store}\ + A local store whose canonical path is anything other than `/nix/store`. + + - [binary cache]{#gloss-binary-cache}\ + A *binary cache* is a Nix store which uses a different format: its + metadata and signatures are kept in `.narinfo` files rather than in a + Nix database. This different format simplifies serving store objects + over the network, but cannot host builds. Examples of binary caches + include S3 buckets and the [NixOS binary + cache](https://cache.nixos.org). + - [store path]{#gloss-store-path}\ - The location in the file system of a store object, i.e., an + The location of a [store object] in the file system, i.e., an immediate child of the Nix store directory. + Example: `/nix/store/a040m110amc4h71lds2jmr8qrkj2jhxd-git-2.38.1` + + [store path]: #gloss-store-path + - [store object]{#gloss-store-object}\ A file that is an immediate child of the Nix store directory. These can be regular files, but also entire directory trees. Store objects can be sources (objects copied from outside of the store), - derivation outputs (objects produced by running a build action), or - derivations (files describing a build action). + derivation outputs (objects produced by running a build task), or + derivations (files describing a build task). + + [store object]: #gloss-store-object + + - [input-addressed store object]{#gloss-input-addressed-store-object}\ + A store object produced by building a + non-[content-addressed](#gloss-content-addressed-derivation), + non-[fixed-output](#gloss-fixed-output-derivation) + derivation. + + - [output-addressed store object]{#gloss-output-addressed-store-object}\ + A store object whose store path hashes its content. This + includes derivations, the outputs of + [content-addressed derivations](#gloss-content-addressed-derivation), + and the outputs of + [fixed-output derivations](#gloss-fixed-output-derivation). - [substitute]{#gloss-substitute}\ A substitute is a command invocation stored in the Nix database that @@ -29,6 +114,13 @@ store object by downloading a pre-built version of the store object from some server. + - [substituter]{#gloss-substituter}\ + A *substituter* is an additional store from which Nix will + copy store objects it doesn't have. For details, see the + [`substituters` option](./command-ref/conf-file.md#conf-substituters). + + [substituter]: #gloss-substituter + - [purity]{#gloss-purity}\ The assumption that equal Nix derivations when run always produce the same output. This cannot be guaranteed in general (e.g., a @@ -43,14 +135,13 @@ then be built. - [reference]{#gloss-reference}\ - A store path `P` is said to have a reference to a store path `Q` if - the store object at `P` contains the path `Q` somewhere. The - *references* of a store path are the set of store paths to which it - has a reference. + A [store object] `O` is said to have a *reference* to a store object `P` if a [store path] to `P` appears in the contents of `O`. - A derivation can reference other derivations and sources (but not - output paths), whereas an output path only references other output - paths. + Store objects can refer to both other store objects and themselves. + References from a store object to itself are called *self-references*. + References other than a self-reference must not form a cycle. + + [reference]: #gloss-reference - [reachable]{#gloss-reachable}\ A store path `Q` is reachable from another store path `P` if `Q` @@ -67,38 +158,62 @@ files could be missing. The command `nix-store -qR` prints out closures of store paths. - As an example, if the store object at path `P` contains a reference - to path `Q`, then `Q` is in the closure of `P`. Further, if `Q` + As an example, if the [store object] at path `P` contains a [reference] + to a store object at path `Q`, then `Q` is in the closure of `P`. Further, if `Q` references `R` then `R` is also in the closure of `P`. + [closure]: #gloss-closure + - [output path]{#gloss-output-path}\ - A store path produced by a derivation. + A [store path] produced by a [derivation]. + + [output path]: #gloss-output-path - [deriver]{#gloss-deriver}\ - The deriver of an *output path* is the store - derivation that built it. + The [store derivation] that produced an [output path]. - [validity]{#gloss-validity}\ - A store path is considered *valid* if it exists in the file system, - is listed in the Nix database as being valid, and if all paths in - its closure are also valid. + A store path is valid if all [store object]s in its [closure] can be read from the [store]. + + For a local store, this means: + - The store path leads to an existing [store object] in that [store]. + - The store path is listed in the Nix database as being valid. + - All paths in the store path's [closure] are valid. + + [validity]: #gloss-validity - [user environment]{#gloss-user-env}\ An automatically generated store object that consists of a set of symlinks to “active” applications, i.e., other store paths. These are generated automatically by - [`nix-env`](command-ref/nix-env.md). See *profiles*. + [`nix-env`](./command-ref/nix-env.md). See *profiles*. - [profile]{#gloss-profile}\ A symlink to the current *user environment* of a user, e.g., `/nix/var/nix/profiles/default`. + - [installable]{#gloss-installable}\ + Something that can be realised in the Nix store. + + See [installables](./command-ref/new-cli/nix.md#installables) for [`nix` commands](./command-ref/new-cli/nix.md) (experimental) for details. + - [NAR]{#gloss-nar}\ A *N*ix *AR*chive. This is a serialisation of a path in the Nix store. It can contain regular files, directories and symbolic links. NARs are generated and unpacked using `nix-store --dump` and `nix-store --restore`. + - [`∅`]{#gloss-emtpy-set}\ The empty set symbol. In the context of profile history, this denotes a package is not present in a particular version of the profile. + - [`ε`]{#gloss-epsilon}\ The epsilon symbol. In the context of a package, this means the version is empty. More precisely, the derivation does not have a version attribute. + + - [string interpolation]{#gloss-string-interpolation}\ + Expanding expressions enclosed in `${ }` within a [string], [path], or [attribute name]. + + See [String interpolation](./language/string-interpolation.md) for details. + + [string]: ./language/values.md#type-string + [path]: ./language/values.md#type-path + [attribute name]: ./language/values.md#attribute-set diff --git a/doc/manual/src/installation/env-variables.md b/doc/manual/src/installation/env-variables.md index bb35c0e9f..fb8155a80 100644 --- a/doc/manual/src/installation/env-variables.md +++ b/doc/manual/src/installation/env-variables.md @@ -27,7 +27,7 @@ Set the environment variable and install Nix ```console $ export NIX_SSL_CERT_FILE=/etc/ssl/my-certificate-bundle.crt -$ sh <(curl -L https://nixos.org/nix/install) +$ curl -L https://nixos.org/nix/install | sh ``` In the shell profile and rc files (for example, `/etc/bashrc`, @@ -38,7 +38,7 @@ export NIX_SSL_CERT_FILE=/etc/ssl/my-certificate-bundle.crt ``` > **Note** -> +> > You must not add the export and then do the install, as the Nix > installer will detect the presence of Nix configuration, and abort. diff --git a/doc/manual/src/installation/installation.md b/doc/manual/src/installation/installation.md index b40c5b95f..dafdeb667 100644 --- a/doc/manual/src/installation/installation.md +++ b/doc/manual/src/installation/installation.md @@ -1,2 +1,38 @@ -This section describes how to install and configure Nix for first-time -use. +# Installation + +This section describes how to install and configure Nix for first-time use. + +The current recommended option on Linux and MacOS is [multi-user](#multi-user). + +## Multi-user + +This installation offers better sharing, improved isolation, and more security +over a single user installation. + +This option requires either: + +* Linux running systemd, with SELinux disabled +* MacOS + +```console +$ bash <(curl -L https://nixos.org/nix/install) --daemon +``` + +## Single-user + +> Single-user is not supported on Mac. + +This installation has less requirements than the multi-user install, however it +cannot offer equivalent sharing, isolation, or security. + +This option is suitable for systems without systemd. + +```console +$ bash <(curl -L https://nixos.org/nix/install) --no-daemon +``` + +## Distributions + +The Nix community maintains installers for several distributions. + +They can be found in the [`nix-community/nix-installers`](https://github.com/nix-community/nix-installers) repository. diff --git a/doc/manual/src/installation/installing-binary.md b/doc/manual/src/installation/installing-binary.md index 9fb9c80c3..e3fd962bd 100644 --- a/doc/manual/src/installation/installing-binary.md +++ b/doc/manual/src/installation/installing-binary.md @@ -3,7 +3,7 @@ The easiest way to install Nix is to run the following command: ```console -$ sh <(curl -L https://nixos.org/nix/install) +$ curl -L https://nixos.org/nix/install | sh ``` This will run the installer interactively (causing it to explain what @@ -13,7 +13,7 @@ for your platform: - multi-user on macOS > **Notes on read-only filesystem root in macOS 10.15 Catalina +** - > + > > - It took some time to support this cleanly. You may see posts, > examples, and tutorials using obsolete workarounds. > - Supporting it cleanly made macOS installs too complex to qualify @@ -27,12 +27,12 @@ you can authenticate with `sudo`. To explicitly select a single-user installation on your system: ```console -$ sh <(curl -L https://nixos.org/nix/install) --no-daemon +$ curl -L https://nixos.org/nix/install | sh -s -- --no-daemon ``` This will perform a single-user installation of Nix, meaning that `/nix` -is owned by the invoking user. You should run this under your usual user -account, *not* as root. The script will invoke `sudo` to create `/nix` +is owned by the invoking user. You can run this under your usual user +account or root. The script will invoke `sudo` to create `/nix` if it doesn’t already exist. If you don’t have `sudo`, you should manually create `/nix` first as root, e.g.: @@ -66,16 +66,16 @@ You can instruct the installer to perform a multi-user installation on your system: ```console -$ sh <(curl -L https://nixos.org/nix/install) --daemon +$ curl -L https://nixos.org/nix/install | sh -s -- --daemon ``` The multi-user installation of Nix will create build users between the user IDs 30001 and 30032, and a group with the group ID 30000. You -should run this under your usual user account, *not* as root. The script +can run this under your usual user account or root. The script will invoke `sudo` as needed. > **Note** -> +> > If you need Nix to use a different group ID or user ID set, you will > have to download the tarball manually and [edit the install > script](#installing-from-a-binary-tarball). @@ -88,19 +88,51 @@ extension. The installer will also create `/etc/profile.d/nix.sh`. ### Linux +If you are on Linux with systemd: + +1. Remove the Nix daemon service: + + ```console + sudo systemctl stop nix-daemon.service + sudo systemctl disable nix-daemon.socket nix-daemon.service + sudo systemctl daemon-reload + ``` + +1. Remove systemd service files: + + ```console + sudo rm /etc/systemd/system/nix-daemon.service /etc/systemd/system/nix-daemon.socket + ``` + +1. The installer script uses systemd-tmpfiles to create the socket directory. + You may also want to remove the configuration for that: + + ```console + sudo rm /etc/tmpfiles.d/nix-daemon.conf + ``` + +Remove files created by Nix: + ```console -sudo rm -rf /etc/profile/nix.sh /etc/nix /nix ~root/.nix-profile ~root/.nix-defexpr ~root/.nix-channels ~/.nix-profile ~/.nix-defexpr ~/.nix-channels - -# If you are on Linux with systemd, you will need to run: -sudo systemctl stop nix-daemon.socket -sudo systemctl stop nix-daemon.service -sudo systemctl disable nix-daemon.socket -sudo systemctl disable nix-daemon.service -sudo systemctl daemon-reload +sudo rm -rf /nix /etc/nix /etc/profile/nix.sh ~root/.nix-profile ~root/.nix-defexpr ~root/.nix-channels ~/.nix-profile ~/.nix-defexpr ~/.nix-channels ``` -There may also be references to Nix in `/etc/profile`, `/etc/bashrc`, -and `/etc/zshrc` which you may remove. +Remove build users and their group: + +```console +for i in $(seq 1 32); do + sudo userdel nixbld$i +done +sudo groupdel nixbld +``` + +There may also be references to Nix in + +- `/etc/profile` +- `/etc/bashrc` +- `/etc/zshrc` + +which you may remove. ### macOS @@ -148,7 +180,8 @@ and `/etc/zshrc` which you may remove. This will remove all the build users that no longer serve a purpose. 4. Edit fstab using `sudo vifs` to remove the line mounting the Nix Store - volume on `/nix`, which looks like this, + volume on `/nix`, which looks like + `UUID=<uuid> /nix apfs rw,noauto,nobrowse,suid,owners` or `LABEL=Nix\040Store /nix apfs rw,nobrowse`. This will prevent automatic mounting of the Nix Store volume. @@ -167,7 +200,7 @@ and `/etc/zshrc` which you may remove. removed next. 7. Remove the Nix Store volume: - + ```console sudo diskutil apfs deleteVolume /nix ``` @@ -175,8 +208,20 @@ and `/etc/zshrc` which you may remove. This will remove the Nix Store volume and everything that was added to the store. + If the output indicates that the command couldn't remove the volume, you should + make sure you don't have an _unmounted_ Nix Store volume. Look for a + "Nix Store" volume in the output of the following command: + + ```console + diskutil list + ``` + + If you _do_ see a "Nix Store" volume, delete it by re-running the diskutil + deleteVolume command, but replace `/nix` with the store volume's `diskXsY` + identifier. + > **Note** -> +> > After you complete the steps here, you will still have an empty `/nix` > directory. This is an expected sign of a successful uninstall. The empty > `/nix` directory will disappear the next time you reboot. @@ -191,8 +236,7 @@ and `/etc/zshrc` which you may remove. <!-- Note: anchors above to catch permalinks to old explanations --> We believe we have ironed out how to cleanly support the read-only root -on modern macOS. New installs will do this automatically, and you can -also re-run a new installer to convert your existing setup. +on modern macOS. New installs will do this automatically. This section previously detailed the situation, options, and trade-offs, but it now only outlines what the installer does. You don't need to know @@ -243,7 +287,7 @@ These install scripts can be used the same as the main NixOS.org installation script: ```console -$ sh <(curl -L https://nixos.org/nix/install) +$ curl -L https://nixos.org/nix/install | sh ``` In the same directory of the install script are sha256 sums, and gpg diff --git a/doc/manual/src/installation/prerequisites-source.md b/doc/manual/src/installation/prerequisites-source.md index 6f4eb3008..5a708f11b 100644 --- a/doc/manual/src/installation/prerequisites-source.md +++ b/doc/manual/src/installation/prerequisites-source.md @@ -71,3 +71,8 @@ <http://libcpuid.sourceforge.net>. This is an optional dependency and can be disabled by providing a `--disable-cpuid` to the `configure` script. + + - Unless `./configure --disable-tests` is specified, GoogleTest (GTest) and + RapidCheck are required, which are available at + <https://google.github.io/googletest/> and + <https://github.com/emil-e/rapidcheck> respectively. diff --git a/doc/manual/src/introduction.md b/doc/manual/src/introduction.md index d87487a07..b54346db8 100644 --- a/doc/manual/src/introduction.md +++ b/doc/manual/src/introduction.md @@ -104,7 +104,7 @@ a currently running program. Packages are built from _Nix expressions_, which is a simple functional language. A Nix expression describes everything that goes -into a package build action (a “derivation”): other packages, sources, +into a package build task (a “derivation”): other packages, sources, the build script, environment variables for the build script, etc. Nix tries very hard to ensure that Nix expressions are _deterministic_: building a Nix expression twice should yield the same diff --git a/doc/manual/src/expressions/advanced-attributes.md b/doc/manual/src/language/advanced-attributes.md index 2e7e80ed0..5a63236e5 100644 --- a/doc/manual/src/expressions/advanced-attributes.md +++ b/doc/manual/src/language/advanced-attributes.md @@ -207,13 +207,13 @@ Derivations can declare some infrequently used optional attributes. the hash in either hexadecimal or base-32 notation. (See the [`nix-hash` command](../command-ref/nix-hash.md) for information about converting to and from base-32 notation.) - + - [`__contentAddressed`]{#adv-attr-__contentAddressed} If this **experimental** attribute is set to true, then the derivation outputs will be stored in a content-addressed location rather than the traditional input-addressed one. - This only has an effect if the `ca-derivation` experimental feature is enabled. - + This only has an effect if the `ca-derivations` experimental feature is enabled. + Setting this attribute also requires setting `outputHashMode` and `outputHashAlgo` like for *fixed-output derivations* (see above). - [`passAsFile`]{#adv-attr-passAsFile}\ @@ -255,3 +255,78 @@ Derivations can declare some infrequently used optional attributes. > substituted. Thus it is usually a good idea to align `system` with > `builtins.currentSystem` when setting `allowSubstitutes` to > `false`. For most trivial derivations this should be the case. + + - [`__structuredAttrs`]{#adv-attr-structuredAttrs}\ + If the special attribute `__structuredAttrs` is set to `true`, the other derivation + attributes are serialised in JSON format and made available to the + builder via the file `.attrs.json` in the builder’s temporary + directory. This obviates the need for [`passAsFile`](#adv-attr-passAsFile) since JSON files + have no size restrictions, unlike process environments. + + It also makes it possible to tweak derivation settings in a structured way; see + [`outputChecks`](#adv-attr-outputChecks) for example. + + As a convenience to Bash builders, + Nix writes a script named `.attrs.sh` to the builder’s directory + that initialises shell variables corresponding to all attributes + that are representable in Bash. This includes non-nested + (associative) arrays. For example, the attribute `hardening.format = true` + ends up as the Bash associative array element `${hardening[format]}`. + + - [`outputChecks`]{#adv-attr-outputChecks}\ + When using [structured attributes](#adv-attr-structuredAttrs), the `outputChecks` + attribute allows defining checks per-output. + + In addition to + [`allowedReferences`](#adv-attr-allowedReferences), [`allowedRequisites`](#adv-attr-allowedRequisites), + [`disallowedReferences`](#adv-attr-disallowedReferences) and [`disallowedRequisites`](#adv-attr-disallowedRequisites), + the following attributes are available: + + - `maxSize` defines the maximum size of the resulting [store object](../glossary.md#gloss-store-object). + - `maxClosureSize` defines the maximum size of the output's closure. + - `ignoreSelfRefs` controls whether self-references should be considered when + checking for allowed references/requisites. + + Example: + + ```nix + __structuredAttrs = true; + + outputChecks.out = { + # The closure of 'out' must not be larger than 256 MiB. + maxClosureSize = 256 * 1024 * 1024; + + # It must not refer to the C compiler or to the 'dev' output. + disallowedRequisites = [ stdenv.cc "dev" ]; + }; + + outputChecks.dev = { + # The 'dev' output must not be larger than 128 KiB. + maxSize = 128 * 1024; + }; + ``` + + - [`unsafeDiscardReferences`]{#adv-attr-unsafeDiscardReferences}\ + > **Warning** + > This is an experimental feature. + > + > To enable it, add the following to [nix.conf](../command-ref/conf-file.md): + > + > ``` + > extra-experimental-features = discard-references + > ``` + + When using [structured attributes](#adv-attr-structuredAttrs), the + attribute `unsafeDiscardReferences` is an attribute set with a boolean value for each output name. + If set to `true`, it disables scanning the output for runtime dependencies. + + Example: + + ```nix + __structuredAttrs = true; + unsafeDiscardReferences.out = true; + ``` + + This is useful, for example, when generating self-contained filesystem images with + their own embedded Nix store: hashes found inside such an image refer + to the embedded store and not to the host's Nix store. diff --git a/doc/manual/src/expressions/builtin-constants.md b/doc/manual/src/language/builtin-constants.md index 78d066a82..78d066a82 100644 --- a/doc/manual/src/expressions/builtin-constants.md +++ b/doc/manual/src/language/builtin-constants.md diff --git a/doc/manual/src/expressions/builtins-prefix.md b/doc/manual/src/language/builtins-prefix.md index c631a8453..c631a8453 100644 --- a/doc/manual/src/expressions/builtins-prefix.md +++ b/doc/manual/src/language/builtins-prefix.md diff --git a/doc/manual/src/expressions/builtins-suffix.md b/doc/manual/src/language/builtins-suffix.md index a74db2857..a74db2857 100644 --- a/doc/manual/src/expressions/builtins-suffix.md +++ b/doc/manual/src/language/builtins-suffix.md diff --git a/doc/manual/src/expressions/language-constructs.md b/doc/manual/src/language/constructs.md index 1c01f2cc7..1c01f2cc7 100644 --- a/doc/manual/src/expressions/language-constructs.md +++ b/doc/manual/src/language/constructs.md diff --git a/doc/manual/src/expressions/derivations.md b/doc/manual/src/language/derivations.md index 3391ec0d8..043a38191 100644 --- a/doc/manual/src/expressions/derivations.md +++ b/doc/manual/src/language/derivations.md @@ -1,7 +1,7 @@ # Derivations The most important built-in function is `derivation`, which is used to -describe a single derivation (a build action). It takes as input a set, +describe a single derivation (a build task). It takes as input a set, the attributes of which specify the inputs of the build. - There must be an attribute named [`system`]{#attr-system} whose value must be a diff --git a/doc/manual/src/language/index.md b/doc/manual/src/language/index.md new file mode 100644 index 000000000..3eabe1a02 --- /dev/null +++ b/doc/manual/src/language/index.md @@ -0,0 +1,581 @@ +# Nix Language + +The Nix language is + +- *domain-specific* + + It only exists for the Nix package manager: + to describe packages and configurations as well as their variants and compositions. + It is not intended for general purpose use. + +- *declarative* + + There is no notion of executing sequential steps. + Dependencies between operations are established only through data. + +- *pure* + + Values cannot change during computation. + Functions always produce the same output if their input does not change. + +- *functional* + + Functions are like any other value. + Functions can be assigned to names, taken as arguments, or returned by functions. + +- *lazy* + + Expressions are only evaluated when their value is needed. + +- *dynamically typed* + + Type errors are only detected when expressions are evaluated. + +# Overview + +This is an incomplete overview of language features, by example. + +<table> + <tr> + <th> + Example + </th> + <th> + Description + </th> + </tr> + <tr> + <td> + + + *Basic values* + + + </td> + <td> + + + + </td> + </tr> + <tr> + <td> + + `"hello world"` + + </td> + <td> + + A string + + </td> + </tr> + <tr> + <td> + + ``` + '' + multi + line + string + '' + ``` + + </td> + <td> + + A multi-line string. Strips common prefixed whitespace. Evaluates to `"multi\n line\n string"`. + + </td> + </tr> + <tr> + <td> + + `"hello ${ { a = "world"; }.a }"` + + `"1 2 ${toString 3}"` + + `"${pkgs.bash}/bin/sh"` + + </td> + <td> + + String interpolation (expands to `"hello world"`, `"1 2 3"`, `"/nix/store/<hash>-bash-<version>/bin/sh"`) + + </td> + </tr> + <tr> + <td> + + `true`, `false` + + </td> + <td> + + Booleans + + </td> + </tr> + <tr> + <td> + + `null` + + </td> + <td> + + Null value + + </td> + </tr> + <tr> + <td> + + `123` + + </td> + <td> + + An integer + + </td> + </tr> + <tr> + <td> + + `3.141` + + </td> + <td> + + A floating point number + + </td> + </tr> + <tr> + <td> + + `/etc` + + </td> + <td> + + An absolute path + + </td> + </tr> + <tr> + <td> + + `./foo.png` + + </td> + <td> + + A path relative to the file containing this Nix expression + + </td> + </tr> + <tr> + <td> + + `~/.config` + + </td> + <td> + + A home path. Evaluates to the `"<user's home directory>/.config"`. + + </td> + </tr> + <tr> + <td> + + `<nixpkgs>` + + </td> + <td> + + Search path for Nix files. Value determined by [`$NIX_PATH` environment variable](../command-ref/env-common.md#env-NIX_PATH). + + </td> + </tr> + <tr> + <td> + + *Compound values* + + </td> + <td> + + + + </td> + </tr> + <tr> + <td> + + `{ x = 1; y = 2; }` + + </td> + <td> + + A set with attributes named `x` and `y` + + </td> + </tr> + <tr> + <td> + + `{ foo.bar = 1; }` + + </td> + <td> + + A nested set, equivalent to `{ foo = { bar = 1; }; }` + + </td> + </tr> + <tr> + <td> + + `rec { x = "foo"; y = x + "bar"; }` + + </td> + <td> + + A recursive set, equivalent to `{ x = "foo"; y = "foobar"; }` + + </td> + </tr> + <tr> + <td> + + `[ "foo" "bar" "baz" ]` + + `[ 1 2 3 ]` + + `[ (f 1) { a = 1; b = 2; } [ "c" ] ]` + + </td> + <td> + + Lists with three elements. + + </td> + </tr> + <tr> + <td> + + *Operators* + + </td> + <td> + + + + </td> + </tr> + <tr> + <td> + + `"foo" + "bar"` + + </td> + <td> + + String concatenation + + </td> + </tr> + <tr> + <td> + + `1 + 2` + + </td> + <td> + + Integer addition + + </td> + </tr> + <tr> + <td> + + `"foo" == "f" + "oo"` + + </td> + <td> + + Equality test (evaluates to `true`) + + </td> + </tr> + <tr> + <td> + + `"foo" != "bar"` + + </td> + <td> + + Inequality test (evaluates to `true`) + + </td> + </tr> + <tr> + <td> + + `!true` + + </td> + <td> + + Boolean negation + + </td> + </tr> + <tr> + <td> + + `{ x = 1; y = 2; }.x` + + </td> + <td> + + Attribute selection (evaluates to `1`) + + </td> + </tr> + <tr> + <td> + + `{ x = 1; y = 2; }.z or 3` + + </td> + <td> + + Attribute selection with default (evaluates to `3`) + + </td> + </tr> + <tr> + <td> + + `{ x = 1; y = 2; } // { z = 3; }` + + </td> + <td> + + Merge two sets (attributes in the right-hand set taking precedence) + + </td> + </tr> + <tr> + <td> + + *Control structures* + + </td> + <td> + + + + </td> + </tr> + <tr> + <td> + + `if 1 + 1 == 2 then "yes!" else "no!"` + + </td> + <td> + + Conditional expression + + </td> + </tr> + <tr> + <td> + + `assert 1 + 1 == 2; "yes!"` + + </td> + <td> + + Assertion check (evaluates to `"yes!"`). + + </td> + </tr> + <tr> + <td> + + `let x = "foo"; y = "bar"; in x + y` + + </td> + <td> + + Variable definition + + </td> + </tr> + <tr> + <td> + + `with builtins; head [ 1 2 3 ]` + + </td> + <td> + + Add all attributes from the given set to the scope (evaluates to `1`) + + </td> + </tr> + <tr> + <td> + + *Functions (lambdas)* + + </td> + <td> + + + + </td> + </tr> + <tr> + <td> + + `x: x + 1` + + </td> + <td> + + A function that expects an integer and returns it increased by 1 + + </td> + </tr> + <tr> + <td> + + `x: y: x + y` + + </td> + <td> + + Curried function, equivalent to `x: (y: x + y)`. Can be used like a function that takes two arguments and returns their sum. + + </td> + </tr> + <tr> + <td> + + `(x: x + 1) 100` + + </td> + <td> + + A function call (evaluates to 101) + + </td> + </tr> + <tr> + <td> + + `let inc = x: x + 1; in inc (inc (inc 100))` + + </td> + <td> + + A function bound to a variable and subsequently called by name (evaluates to 103) + + </td> + </tr> + <tr> + <td> + + `{ x, y }: x + y` + + </td> + <td> + + A function that expects a set with required attributes `x` and `y` and concatenates them + + </td> + </tr> + <tr> + <td> + + `{ x, y ? "bar" }: x + y` + + </td> + <td> + + A function that expects a set with required attribute `x` and optional `y`, using `"bar"` as default value for `y` + + </td> + </tr> + <tr> + <td> + + `{ x, y, ... }: x + y` + + </td> + <td> + + A function that expects a set with required attributes `x` and `y` and ignores any other attributes + + </td> + </tr> + <tr> + <td> + + `{ x, y } @ args: x + y` + + `args @ { x, y }: x + y` + + </td> + <td> + + A function that expects a set with required attributes `x` and `y`, and binds the whole set to `args` + + </td> + </tr> + <tr> + <td> + + *Built-in functions* + + </td> + <td> + + + + </td> + </tr> + <tr> + <td> + + `import ./foo.nix` + + </td> + <td> + + Load and return Nix expression in given file + + </td> + </tr> + <tr> + <td> + + `map (x: x + x) [ 1 2 3 ]` + + </td> + <td> + + Apply a function to every element of a list (evaluates to `[ 2 4 6 ]`) + + </td> + </tr> +</table> diff --git a/doc/manual/src/language/operators.md b/doc/manual/src/language/operators.md new file mode 100644 index 000000000..90b325597 --- /dev/null +++ b/doc/manual/src/language/operators.md @@ -0,0 +1,167 @@ +# Operators + +| Name | Syntax | Associativity | Precedence | +|----------------------------------------|--------------------------------------------|---------------|------------| +| [Attribute selection] | *attrset* `.` *attrpath* \[ `or` *expr* \] | none | 1 | +| Function application | *func* *expr* | left | 2 | +| [Arithmetic negation][arithmetic] | `-` *number* | none | 3 | +| [Has attribute] | *attrset* `?` *attrpath* | none | 4 | +| List concatenation | *list* `++` *list* | right | 5 | +| [Multiplication][arithmetic] | *number* `*` *number* | left | 6 | +| [Division][arithmetic] | *number* `/` *number* | left | 6 | +| [Subtraction][arithmetic] | *number* `-` *number* | left | 7 | +| [Addition][arithmetic] | *number* `+` *number* | left | 7 | +| [String concatenation] | *string* `+` *string* | left | 7 | +| [Path concatenation] | *path* `+` *path* | left | 7 | +| [Path and string concatenation] | *path* `+` *string* | left | 7 | +| [String and path concatenation] | *string* `+` *path* | left | 7 | +| Logical negation (`NOT`) | `!` *bool* | none | 8 | +| [Update] | *attrset* `//` *attrset* | right | 9 | +| [Less than][Comparison] | *expr* `<` *expr* | none | 10 | +| [Less than or equal to][Comparison] | *expr* `<=` *expr* | none | 10 | +| [Greater than][Comparison] | *expr* `>` *expr* | none | 10 | +| [Greater than or equal to][Comparison] | *expr* `>=` *expr* | none | 10 | +| [Equality] | *expr* `==` *expr* | none | 11 | +| Inequality | *expr* `!=` *expr* | none | 11 | +| Logical conjunction (`AND`) | *bool* `&&` *bool* | left | 12 | +| Logical disjunction (`OR`) | *bool* <code>\|\|</code> *bool* | left | 13 | +| [Logical implication] | *bool* `->` *bool* | none | 14 | + +[string]: ./values.md#type-string +[path]: ./values.md#type-path +[number]: ./values.md#type-number +[list]: ./values.md#list +[attribute set]: ./values.md#attribute-set + +## Attribute selection + +Select the attribute denoted by attribute path *attrpath* from [attribute set] *attrset*. +If the attribute doesn’t exist, return *value* if provided, otherwise abort evaluation. + +<!-- FIXME: the following should to into its own language syntax section, but that needs more work to fit in well --> + +An attribute path is a dot-separated list of attribute names. +An attribute name can be an identifier or a string. + +> *attrpath* = *name* [ `.` *name* ]... +> *name* = *identifier* | *string* +> *identifier* ~ `[a-zA-Z_][a-zA-Z0-9_'-]*` + +[Attribute selection]: #attribute-selection + +## Has attribute + +> *attrset* `?` *attrpath* + +Test whether [attribute set] *attrset* contains the attribute denoted by *attrpath*. +The result is a [Boolean] value. + +[Boolean]: ./values.md#type-boolean + +[Has attribute]: #has-attribute + +## Arithmetic + +Numbers are type-compatible: +Pure integer operations will always return integers, whereas any operation involving at least one floating point number return a floating point number. + +See also [Comparison] and [Equality]. + +The `+` operator is overloaded to also work on strings and paths. + +[arithmetic]: #arithmetic + +## String concatenation + +> *string* `+` *string* + +Concatenate two [string]s and merge their string contexts. + +[String concatenation]: #string-concatenation + +## Path concatenation + +> *path* `+` *path* + +Concatenate two [path]s. +The result is a path. + +[Path concatenation]: #path-concatenation + +## Path and string concatenation + +> *path* + *string* + +Concatenate *[path]* with *[string]*. +The result is a path. + +> **Note** +> +> The string must not have a string context that refers to a [store path]. + +[Path and string concatenation]: #path-and-string-concatenation + +## String and path concatenation + +> *string* + *path* + +Concatenate *[string]* with *[path]*. +The result is a string. + +> **Important** +> +> The file or directory at *path* must exist and is copied to the [store]. +> The path appears in the result as the corresponding [store path]. + +[store path]: ../glossary.md#gloss-store-path +[store]: ../glossary.md#gloss-store + +[String and path concatenation]: #string-and-path-concatenation + +## Update + +> *attrset1* // *attrset2* + +Update [attribute set] *attrset1* with names and values from *attrset2*. + +The returned attribute set will have of all the attributes in *attrset1* and *attrset2*. +If an attribute name is present in both, the attribute value from the latter is taken. + +[Update]: #update + +## Comparison + +Comparison is + +- [arithmetic] for [number]s +- lexicographic for [string]s and [path]s +- item-wise lexicographic for [list]s: + elements at the same index in both lists are compared according to their type and skipped if they are equal. + +All comparison operators are implemented in terms of `<`, and the following equivalencies hold: + +| comparison | implementation | +|--------------|-----------------------| +| *a* `<=` *b* | `! (` *b* `<` *a* `)` | +| *a* `>` *b* | *b* `<` *a* | +| *a* `>=` *b* | `! (` *a* `<` *b* `)` | + +[Comparison]: #comparison-operators + +## Equality + +- [Attribute sets][attribute set] and [list]s are compared recursively, and therefore are fully evaluated. +- Comparison of [function]s always returns `false`. +- Numbers are type-compatible, see [arithmetic] operators. +- Floating point numbers only differ up to a limited precision. + +[function]: ./constructs.md#functions + +[Equality]: #equality + +## Logical implication + +Equivalent to `!`*b1* `||` *b2*. + +[Logical implication]: #logical-implication + diff --git a/doc/manual/src/language/string-interpolation.md b/doc/manual/src/language/string-interpolation.md new file mode 100644 index 000000000..ddc6b8230 --- /dev/null +++ b/doc/manual/src/language/string-interpolation.md @@ -0,0 +1,82 @@ +# String interpolation + +String interpolation is a language feature where a [string], [path], or [attribute name] can contain expressions enclosed in `${ }` (dollar-sign with curly brackets). + +Such a string is an *interpolated string*, and an expression inside is an *interpolated expression*. + +Interpolated expressions must evaluate to one of the following: + +- a [string] +- a [path] +- a [derivation] + +[string]: ./values.md#type-string +[path]: ./values.md#type-path +[attribute name]: ./values.md#attribute-set +[derivation]: ../glossary.md#gloss-derivation + +## Examples + +### String + +Rather than writing + +```nix +"--with-freetype2-library=" + freetype + "/lib" +``` + +(where `freetype` is a [derivation]), you can instead write + +```nix +"--with-freetype2-library=${freetype}/lib" +``` + +The latter is automatically translated to the former. + +A more complicated example (from the Nix expression for [Qt](http://www.trolltech.com/products/qt)): + +```nix +configureFlags = " + -system-zlib -system-libpng -system-libjpeg + ${if openglSupport then "-dlopen-opengl + -L${mesa}/lib -I${mesa}/include + -L${libXmu}/lib -I${libXmu}/include" else ""} + ${if threadSupport then "-thread" else "-no-thread"} +"; +``` + +Note that Nix expressions and strings can be arbitrarily nested; +in this case the outer string contains various interpolated expressions that themselves contain strings (e.g., `"-thread"`), some of which in turn contain interpolated expressions (e.g., `${mesa}`). + +### Path + +Rather than writing + +```nix +./. + "/" + foo + "-" + bar + ".nix" +``` + +or + +```nix +./. + "/${foo}-${bar}.nix" +``` + +you can instead write + +```nix +./${foo}-${bar}.nix +``` + +### Attribute name + +Attribute names can be created dynamically with string interpolation: + +```nix +let name = "foo"; in +{ + ${name} = "bar"; +} +``` + + { foo = "bar"; } diff --git a/doc/manual/src/expressions/language-values.md b/doc/manual/src/language/values.md index f09400d02..c85124278 100644 --- a/doc/manual/src/expressions/language-values.md +++ b/doc/manual/src/language/values.md @@ -13,41 +13,9 @@ returns and tabs can be written as `\n`, `\r` and `\t`, respectively. - You can include the result of an expression into a string by - enclosing it in `${...}`, a feature known as *antiquotation*. The - enclosed expression must evaluate to something that can be coerced - into a string (meaning that it must be a string, a path, or a - derivation). For instance, rather than writing + You can include the results of other expressions into a string by enclosing them in `${ }`, a feature known as [string interpolation]. - ```nix - "--with-freetype2-library=" + freetype + "/lib" - ``` - - (where `freetype` is a derivation), you can instead write the more - natural - - ```nix - "--with-freetype2-library=${freetype}/lib" - ``` - - The latter is automatically translated to the former. A more - complicated example (from the Nix expression for - [Qt](http://www.trolltech.com/products/qt)): - - ```nix - configureFlags = " - -system-zlib -system-libpng -system-libjpeg - ${if openglSupport then "-dlopen-opengl - -L${mesa}/lib -I${mesa}/include - -L${libXmu}/lib -I${libXmu}/include" else ""} - ${if threadSupport then "-thread" else "-no-thread"} - "; - ``` - - Note that Nix expressions and strings can be arbitrarily nested; in - this case the outer string contains various antiquotations that - themselves contain strings (e.g., `"-thread"`), some of which in - turn contain expressions (e.g., `${mesa}`). + [string interpolation]: ./string-interpolation.md The second way to write string literals is as an *indented string*, which is enclosed between pairs of *double single-quotes*, like so: @@ -75,7 +43,7 @@ Note that the whitespace and newline following the opening `''` is ignored if there is no non-whitespace text on the initial line. - Antiquotation (`${expr}`) is supported in indented strings. + Indented strings support [string interpolation]. Since `${` and `''` have special meaning in indented strings, you need a way to quote them. `$` can be escaped by prefixing it with @@ -117,9 +85,10 @@ Numbers, which can be *integers* (like `123`) or *floating point* (like `123.43` or `.27e13`). - Numbers are type-compatible: pure integer operations will always - return integers, whereas any operation involving at least one - floating point number will have a floating point number as a result. + See [arithmetic] and [comparison] operators for semantics. + + [arithmetic]: ./operators.md#arithmetic + [comparison]: ./operators.md#comparison - <a id="type-path" href="#type-path">Path</a> @@ -143,12 +112,23 @@ environment variable `NIX_PATH` will be searched for the given file or directory name. - Antiquotation is supported in any paths except those in angle brackets. - `./${foo}-${bar}.nix` is a more convenient way of writing - `./. + "/" + foo + "-" + bar + ".nix"` or `./. + "/${foo}-${bar}.nix"`. At - least one slash must appear *before* any antiquotations for this to be - recognized as a path. `a.${foo}/b.${bar}` is a syntactically valid division - operation. `./a.${foo}/b.${bar}` is a path. + When an [interpolated string][string interpolation] evaluates to a path, the path is first copied into the Nix store and the resulting string is the [store path] of the newly created [store object]. + + [store path]: ../glossary.md#gloss-store-path + [store object]: ../glossary.md#gloss-store-object + + For instance, evaluating `"${./foo.txt}"` will cause `foo.txt` in the current directory to be copied into the Nix store and result in the string `"/nix/store/<hash>-foo.txt"`. + + Note that the Nix language assumes that all input files will remain _unchanged_ while evaluating a Nix expression. + For example, assume you used a file path in an interpolated string during a `nix repl` session. + Later in the same session, after having changed the file contents, evaluating the interpolated string with the file path again might not return a new store path, since Nix might not re-read the file contents. + + Paths themselves, except those in angle brackets (`< >`), support [string interpolation]. + + At least one slash (`/`) must appear *before* any interpolated expression for the result to be recognized as a path. + + `a.${foo}/b.${bar}` is a syntactically valid division operation. + `./a.${foo}/b.${bar}` is a path. - <a id="type-boolean" href="#type-boolean">Boolean</a> @@ -221,23 +201,33 @@ will evaluate to `"Xyzzy"` because there is no `c` attribute in the set. You can use arbitrary double-quoted strings as attribute names: ```nix -{ "foo ${bar}" = 123; "nix-1.0" = 456; }."foo ${bar}" +{ "$!@#?" = 123; }."$!@#?" +``` + +```nix +let bar = "bar"; in +{ "foo ${bar}" = 123; }."foo ${bar}" ``` -This will evaluate to `123` (Assuming `bar` is antiquotable). In the -case where an attribute name is just a single antiquotation, the quotes -can be dropped: +Both will evaluate to `123`. + +Attribute names support [string interpolation]: + +```nix +let bar = "foo"; in +{ foo = 123; }.${bar} +``` ```nix -{ foo = 123; }.${bar} or 456 +let bar = "foo"; in +{ ${bar} = 123; }.foo ``` -This will evaluate to `123` if `bar` evaluates to `"foo"` when coerced -to a string and `456` otherwise (again assuming `bar` is antiquotable). +Both will evaluate to `123`. In the special case where an attribute name inside of a set declaration -evaluates to `null` (which is normally an error, as `null` is not -antiquotable), that attribute is simply not added to the set: +evaluates to `null` (which is normally an error, as `null` cannot be coerced to +a string), that attribute is simply not added to the set: ```nix { ${if foo then "bar" else null} = true; } diff --git a/doc/manual/src/package-management/binary-cache-substituter.md b/doc/manual/src/package-management/binary-cache-substituter.md index ef738794b..5befad9f8 100644 --- a/doc/manual/src/package-management/binary-cache-substituter.md +++ b/doc/manual/src/package-management/binary-cache-substituter.md @@ -32,13 +32,13 @@ which should print something like: Priority: 30 On the client side, you can tell Nix to use your binary cache using -`--option extra-binary-caches`, e.g.: +`--substituters`, e.g.: ```console -$ nix-env -iA nixpkgs.firefox --option extra-binary-caches http://avalon:8080/ +$ nix-env -iA nixpkgs.firefox --substituters http://avalon:8080/ ``` -The option `extra-binary-caches` tells Nix to use this binary cache in +The option `substituters` tells Nix to use this binary cache in addition to your default caches, such as <https://cache.nixos.org>. Thus, for any path in the closure of Firefox, Nix will first check if the path is available on the server `avalon` or another binary caches. @@ -47,4 +47,4 @@ If not, it will fall back to building from source. You can also tell Nix to always use your binary cache by adding a line to the `nix.conf` configuration file like this: - binary-caches = http://avalon:8080/ https://cache.nixos.org/ + substituters = http://avalon:8080/ https://cache.nixos.org/ diff --git a/doc/manual/src/package-management/package-management.md b/doc/manual/src/package-management/package-management.md index bd26a09ab..d528112e2 100644 --- a/doc/manual/src/package-management/package-management.md +++ b/doc/manual/src/package-management/package-management.md @@ -1,5 +1,4 @@ This chapter discusses how to do package management with Nix, i.e., how to obtain, install, upgrade, and erase packages. This is the “user’s” perspective of the Nix system — people who want to *create* -packages should consult the [chapter on writing Nix -expressions](../expressions/writing-nix-expressions.md). +packages should consult the chapter on the [Nix language](../language/index.md). diff --git a/doc/manual/src/quick-start.md b/doc/manual/src/quick-start.md index b54e73500..651134c25 100644 --- a/doc/manual/src/quick-start.md +++ b/doc/manual/src/quick-start.md @@ -4,16 +4,16 @@ This chapter is for impatient people who don't like reading documentation. For more in-depth information you are kindly referred to subsequent chapters. -1. Install single-user Nix by running the following: +1. Install Nix by running the following: ```console - $ bash <(curl -L https://nixos.org/nix/install) + $ curl -L https://nixos.org/nix/install | sh ``` - This will install Nix in `/nix`. The install script will create - `/nix` using `sudo`, so make sure you have sufficient rights. (For - other installation methods, see - [here](installation/installation.md).) + The install script will use `sudo`, so make sure you have sufficient rights. + On Linux, `--daemon` can be omitted for a single-user install. + + For other installation methods, see [here](installation/installation.md). 1. See what installable packages are currently available in the channel: diff --git a/doc/manual/src/release-notes/rl-2.11.md b/doc/manual/src/release-notes/rl-2.11.md new file mode 100644 index 000000000..b322a4e5e --- /dev/null +++ b/doc/manual/src/release-notes/rl-2.11.md @@ -0,0 +1,5 @@ +# Release 2.11 (2022-08-24) + +* `nix copy` now copies the store paths in parallel as much as possible (again). + This doesn't apply for the `daemon` and `ssh-ng` stores which copy everything + in one batch to avoid latencies issues. diff --git a/doc/manual/src/release-notes/rl-2.12.md b/doc/manual/src/release-notes/rl-2.12.md new file mode 100644 index 000000000..e2045d7bf --- /dev/null +++ b/doc/manual/src/release-notes/rl-2.12.md @@ -0,0 +1,43 @@ +# Release 2.12 (2022-12-06) + +* On Linux, Nix can now run builds in a user namespace where they run + as root (UID 0) and have 65,536 UIDs available. + <!-- FIXME: move this to its own section about system features --> + This is primarily useful for running containers such as `systemd-nspawn` + inside a Nix build. For an example, see [`tests/systemd-nspawn/nix`][nspawn]. + + [nspawn]: https://github.com/NixOS/nix/blob/67bcb99700a0da1395fa063d7c6586740b304598/tests/systemd-nspawn.nix. + + A build can enable this by setting the derivation attribute: + + ``` + requiredSystemFeatures = [ "uid-range" ]; + ``` + + The `uid-range` [system feature] requires the [`auto-allocate-uids`] + setting to be enabled. + + [system feature]: ../command-ref/conf-file.md#conf-system-features + +* Nix can now automatically pick UIDs for builds, removing the need to + create `nixbld*` user accounts. See [`auto-allocate-uids`]. + + [`auto-allocate-uids`]: ../command-ref/conf-file.md#conf-auto-allocate-uids + +* On Linux, Nix has experimental support for running builds inside a + cgroup. See + [`use-cgroups`](../command-ref/conf-file.md#conf-use-cgroups). + +* `<nix/fetchurl.nix>` now accepts an additional argument `impure` which + defaults to `false`. If it is set to `true`, the `hash` and `sha256` + arguments will be ignored and the resulting derivation will have + `__impure` set to `true`, making it an impure derivation. + +* If `builtins.readFile` is called on a file with context, then only + the parts of the context that appear in the content of the file are + retained. This avoids a lot of spurious errors where strings end up + having a context just because they are read from a store path + ([#7260](https://github.com/NixOS/nix/pull/7260)). + +* `nix build --json` now prints some statistics about top-level + derivations, such as CPU statistics when cgroups are enabled. diff --git a/doc/manual/src/release-notes/rl-2.13.md b/doc/manual/src/release-notes/rl-2.13.md new file mode 100644 index 000000000..168708113 --- /dev/null +++ b/doc/manual/src/release-notes/rl-2.13.md @@ -0,0 +1,44 @@ +# Release 2.13 (2023-01-17) + +* The `repeat` and `enforce-determinism` options have been removed + since they had been broken under many circumstances for a long time. + +* You can now use [flake references] in the [old command line interface], e.g. + + [flake references]: ../command-ref/new-cli/nix3-flake.md#flake-references + [old command line interface]: ../command-ref/main-commands.md + + ```shell-session + # nix-build flake:nixpkgs -A hello + # nix-build -I nixpkgs=flake:github:NixOS/nixpkgs/nixos-22.05 \ + '<nixpkgs>' -A hello + # NIX_PATH=nixpkgs=flake:nixpkgs nix-build '<nixpkgs>' -A hello + ``` + +* Instead of "antiquotation", the more common term [string interpolation](../language/string-interpolation.md) is now used consistently. + Historical release notes were not changed. + +* Error traces have been reworked to provide detailed explanations and more + accurate error locations. A short excerpt of the trace is now shown by + default when an error occurs. + +* Allow explicitly selecting outputs in a store derivation installable, just like we can do with other sorts of installables. + For example, + ```shell-session + # nix build /nix/store/gzaflydcr6sb3567hap9q6srzx8ggdgg-glibc-2.33-78.drv^dev + ``` + now works just as + ```shell-session + # nix build nixpkgs#glibc^dev + ``` + does already. + +* On Linux, `nix develop` now sets the + [*personality*](https://man7.org/linux/man-pages/man2/personality.2.html) + for the development shell in the same way as the actual build of the + derivation. This makes shells for `i686-linux` derivations work + correctly on `x86_64-linux`. + +* You can now disable the global flake registry by setting the `flake-registry` + configuration option to an empty string. The same can be achieved at runtime with + `--flake-registry ""`. diff --git a/doc/manual/src/release-notes/rl-2.14.md b/doc/manual/src/release-notes/rl-2.14.md new file mode 100644 index 000000000..705c118bb --- /dev/null +++ b/doc/manual/src/release-notes/rl-2.14.md @@ -0,0 +1,22 @@ +# Release 2.14 (2023-02-28) + +* A new function `builtins.readFileType` is available. It is similar to + `builtins.readDir` but acts on a single file or directory. + +* In flakes, the `.outPath` attribute of a flake now always refers to + the directory containing the `flake.nix`. This was not the case for + when `flake.nix` was in a subdirectory of e.g. a Git repository. + The root of the source of a flake in a subdirectory is still + available in `.sourceInfo.outPath`. + +* In derivations that use structured attributes, you can now use `unsafeDiscardReferences` + to disable scanning a given output for runtime dependencies: + ```nix + __structuredAttrs = true; + unsafeDiscardReferences.out = true; + ``` + This is useful e.g. when generating self-contained filesystem images with + their own embedded Nix store: hashes found inside such an image refer + to the embedded store and not to the host's Nix store. + + This requires the `discard-references` experimental feature. diff --git a/doc/manual/src/release-notes/rl-next.md b/doc/manual/src/release-notes/rl-next.md index 78ae99f4b..b8d14e605 100644 --- a/doc/manual/src/release-notes/rl-next.md +++ b/doc/manual/src/release-notes/rl-next.md @@ -1,2 +1,29 @@ # Release X.Y (202?-??-??) +* Commands which take installables on the command line can now read them from the standard input if + passed the `--stdin` flag. This is primarily useful when you have a large amount of paths which + exceed the OS arg limit. + +* The `nix-hash` command now supports Base64 and SRI. Use the flags `--base64` + or `--sri` to specify the format of output hash as Base64 or SRI, and `--to-base64` + or `--to-sri` to convert a hash to Base64 or SRI format, respectively. + + As the choice of hash formats is no longer binary, the `--base16` flag is also added + to explicitly specify the Base16 format, which is still the default. + +* The special handling of an [installable](../command-ref/new-cli/nix.md#installables) with `.drv` suffix being interpreted as all of the given [store derivation](../glossary.md#gloss-store-derivation)'s output paths is removed, and instead taken as the literal store path that it represents. + + The new `^` syntax for store paths introduced in Nix 2.13 allows explicitly referencing output paths of a derivation. + Using this is better and more clear than relying on the now-removed `.drv` special handling. + + For example, + ```shell-session + $ nix path-info /nix/store/gzaflydcr6sb3567hap9q6srzx8ggdgg-glibc-2.33-78.drv + ``` + + now gives info about the derivation itself, while + + ```shell-session + $ nix path-info /nix/store/gzaflydcr6sb3567hap9q6srzx8ggdgg-glibc-2.33-78.drv^* + ``` + provides information about each of its outputs.
\ No newline at end of file diff --git a/doc/manual/utils.nix b/doc/manual/utils.nix index d4b18472f..d0643ef46 100644 --- a/doc/manual/utils.nix +++ b/doc/manual/utils.nix @@ -5,6 +5,32 @@ rec { concatStrings = concatStringsSep ""; + replaceStringsRec = from: to: string: + # recursively replace occurrences of `from` with `to` within `string` + # example: + # replaceStringRec "--" "-" "hello-----world" + # => "hello-world" + let + replaced = replaceStrings [ from ] [ to ] string; + in + if replaced == string then string else replaceStringsRec from to replaced; + + squash = replaceStringsRec "\n\n\n" "\n\n"; + + trim = string: + # trim trailing spaces and squash non-leading spaces + let + trimLine = line: + let + # separate leading spaces from the rest + parts = split "(^ *)" line; + spaces = head (elemAt parts 1); + rest = elemAt parts 2; + # drop trailing spaces + body = head (split " *$" rest); + in spaces + replaceStringsRec " " " " body; + in concatStringsSep "\n" (map trimLine (splitLines string)); + # FIXME: O(n^2) unique = foldl' (acc: e: if elem e acc then acc else acc ++ [ e ]) []; diff --git a/docker.nix b/docker.nix index 8e6aa227f..52199af66 100644 --- a/docker.nix +++ b/docker.nix @@ -2,11 +2,13 @@ , lib ? pkgs.lib , name ? "nix" , tag ? "latest" +, bundleNixpkgs ? true , channelName ? "nixpkgs" , channelURL ? "https://nixos.org/channels/nixpkgs-unstable" , extraPkgs ? [] , maxLayers ? 100 , nixConf ? {} +, flake-registry ? null }: let defaultPkgs = with pkgs; [ @@ -32,9 +34,20 @@ let root = { uid = 0; - shell = "/bin/bash"; + shell = "${pkgs.bashInteractive}/bin/bash"; home = "/root"; gid = 0; + groups = [ "root" ]; + description = "System administrator"; + }; + + nobody = { + uid = 65534; + shell = "${pkgs.shadow}/bin/nologin"; + home = "/var/empty"; + gid = 65534; + groups = [ "nobody" ]; + description = "Unprivileged account (don't use!)"; }; } // lib.listToAttrs ( @@ -56,6 +69,7 @@ let groups = { root.gid = 0; nixbld.gid = 30000; + nobody.gid = 65534; }; userToPasswd = ( @@ -139,10 +153,12 @@ let baseSystem = let nixpkgs = pkgs.path; - channel = pkgs.runCommand "channel-nixos" { } '' + channel = pkgs.runCommand "channel-nixos" { inherit bundleNixpkgs; } '' mkdir $out - ln -s ${nixpkgs} $out/nixpkgs - echo "[]" > $out/manifest.nix + if [ "$bundleNixpkgs" ]; then + ln -s ${nixpkgs} $out/nixpkgs + echo "[]" > $out/manifest.nix + fi ''; rootEnv = pkgs.buildPackages.buildEnv { name = "root-profile-env"; @@ -232,7 +248,16 @@ let mkdir -p $out/bin $out/usr/bin ln -s ${pkgs.coreutils}/bin/env $out/usr/bin/env ln -s ${pkgs.bashInteractive}/bin/bash $out/bin/sh - ''; + + '' + (lib.optionalString (flake-registry != null) '' + nixCacheDir="/root/.cache/nix" + mkdir -p $out$nixCacheDir + globalFlakeRegistryPath="$nixCacheDir/flake-registry.json" + ln -s ${flake-registry}/flake-registry.json $out$globalFlakeRegistryPath + mkdir -p $out/nix/var/nix/gcroots/auto + rootName=$(${pkgs.nix}/bin/nix --extra-experimental-features nix-command hash file --type sha1 --base32 <(echo -n $globalFlakeRegistryPath)) + ln -s $globalFlakeRegistryPath $out/nix/var/nix/gcroots/auto/$rootName + ''); in pkgs.dockerTools.buildLayeredImageWithNixDb { diff --git a/flake.lock b/flake.lock index a66c9cb1b..1d2aab5ed 100644 --- a/flake.lock +++ b/flake.lock @@ -1,5 +1,21 @@ { "nodes": { + "flake-compat": { + "flake": false, + "locked": { + "lastModified": 1673956053, + "narHash": "sha256-4gtG9iQuiKITOjNQQeQIpoIB6b16fm+504Ch3sNKLd8=", + "owner": "edolstra", + "repo": "flake-compat", + "rev": "35bb57c0c8d8b62bbfd284272c928ceb64ddbde9", + "type": "github" + }, + "original": { + "owner": "edolstra", + "repo": "flake-compat", + "type": "github" + } + }, "lowdown-src": { "flake": false, "locked": { @@ -18,16 +34,16 @@ }, "nixpkgs": { "locked": { - "lastModified": 1657693803, - "narHash": "sha256-G++2CJ9u0E7NNTAi9n5G8TdDmGJXcIjkJ3NF8cetQB8=", + "lastModified": 1670461440, + "narHash": "sha256-jy1LB8HOMKGJEGXgzFRLDU1CBGL0/LlkolgnqIsF0D8=", "owner": "NixOS", "repo": "nixpkgs", - "rev": "365e1b3a859281cf11b94f87231adeabbdd878a2", + "rev": "04a75b2eecc0acf6239acf9dd04485ff8d14f425", "type": "github" }, "original": { "owner": "NixOS", - "ref": "nixos-22.05-small", + "ref": "nixos-22.11-small", "repo": "nixpkgs", "type": "github" } @@ -50,6 +66,7 @@ }, "root": { "inputs": { + "flake-compat": "flake-compat", "lowdown-src": "lowdown-src", "nixpkgs": "nixpkgs", "nixpkgs-regression": "nixpkgs-regression" @@ -1,60 +1,68 @@ { description = "The purely functional package manager"; - inputs.nixpkgs.url = "github:NixOS/nixpkgs/nixos-22.05-small"; + inputs.nixpkgs.url = "github:NixOS/nixpkgs/nixos-22.11-small"; inputs.nixpkgs-regression.url = "github:NixOS/nixpkgs/215d4d0fd80ca5163643b03a33fde804a29cc1e2"; inputs.lowdown-src = { url = "github:kristapsdz/lowdown"; flake = false; }; + inputs.flake-compat = { url = "github:edolstra/flake-compat"; flake = false; }; - outputs = { self, nixpkgs, nixpkgs-regression, lowdown-src }: + outputs = { self, nixpkgs, nixpkgs-regression, lowdown-src, flake-compat }: let + inherit (nixpkgs) lib; - version = builtins.readFile ./.version + versionSuffix; + officialRelease = false; + + version = lib.fileContents ./.version + versionSuffix; versionSuffix = if officialRelease then "" else "pre${builtins.substring 0 8 (self.lastModifiedDate or self.lastModified or "19700101")}_${self.shortRev or "dirty"}"; - officialRelease = false; - linux64BitSystems = [ "x86_64-linux" "aarch64-linux" ]; linuxSystems = linux64BitSystems ++ [ "i686-linux" ]; systems = linuxSystems ++ [ "x86_64-darwin" "aarch64-darwin" ]; crossSystems = [ "armv6l-linux" "armv7l-linux" ]; - stdenvs = [ "gccStdenv" "clangStdenv" "clang11Stdenv" "stdenv" "libcxxStdenv" ]; + stdenvs = [ "gccStdenv" "clangStdenv" "clang11Stdenv" "stdenv" "libcxxStdenv" "ccacheStdenv" ]; - forAllSystems = f: nixpkgs.lib.genAttrs systems (system: f system); - forAllSystemsAndStdenvs = f: forAllSystems (system: - nixpkgs.lib.listToAttrs + forAllSystems = lib.genAttrs systems; + + forAllCrossSystems = lib.genAttrs crossSystems; + + forAllStdenvs = f: + lib.listToAttrs (map - (n: - nixpkgs.lib.nameValuePair "${n}Packages" ( - f system n - )) stdenvs - ) - ); + (stdenvName: { + name = "${stdenvName}Packages"; + value = f stdenvName; + }) + stdenvs); - forAllStdenvs = f: nixpkgs.lib.genAttrs stdenvs (stdenv: f stdenv); # Memoize nixpkgs for different platforms for efficiency. - nixpkgsFor = - let stdenvsPackages = forAllSystemsAndStdenvs - (system: stdenv: - import nixpkgs { - inherit system; - overlays = [ - (overlayFor (p: p.${stdenv})) - ]; - } - ); - in - # Add the `stdenvPackages` at toplevel, both because these are the ones - # we want most of the time and for backwards compatibility - forAllSystems (system: stdenvsPackages.${system} // stdenvsPackages.${system}.stdenvPackages); + nixpkgsFor = forAllSystems + (system: let + make-pkgs = crossSystem: stdenv: import nixpkgs { + inherit system crossSystem; + overlays = [ + (overlayFor (p: p.${stdenv})) + ]; + }; + stdenvs = forAllStdenvs (make-pkgs null); + native = stdenvs.stdenvPackages; + in { + inherit stdenvs native; + static = native.pkgsStatic; + cross = forAllCrossSystems (crossSystem: make-pkgs crossSystem "stdenv"); + }); - commonDeps = { pkgs, isStatic ? false }: with pkgs; rec { + commonDeps = + { pkgs + , isStatic ? pkgs.stdenv.hostPlatform.isStatic + }: + with pkgs; rec { # Use "busybox-sandbox-shell" if present, # if not (legacy) fallback and hope it's sufficient. sh = pkgs.busybox-sandbox-shell or (busybox.override { @@ -90,12 +98,21 @@ "LDFLAGS=-fuse-ld=gold" ]; + testConfigureFlags = [ + "RAPIDCHECK_HEADERS=${lib.getDev rapidcheck}/extras/gtest/include" + ]; + + internalApiDocsConfigureFlags = [ + "--enable-internal-api-docs" + ]; + nativeBuildDeps = [ buildPackages.bison buildPackages.flex (lib.getBin buildPackages.lowdown-nix) buildPackages.mdbook + buildPackages.mdbook-linkcheck buildPackages.autoconf-archive buildPackages.autoreconfHook buildPackages.pkg-config @@ -108,18 +125,26 @@ ++ lib.optionals stdenv.hostPlatform.isLinux [(buildPackages.util-linuxMinimal or buildPackages.utillinuxMinimal)]; buildDeps = - [ (curl.override { patchNetrcRegression = true; }) + [ curl bzip2 xz brotli editline openssl sqlite libarchive boost lowdown-nix - gtest ] ++ lib.optionals stdenv.isLinux [libseccomp] ++ lib.optional (stdenv.isLinux || stdenv.isDarwin) libsodium ++ lib.optional stdenv.hostPlatform.isx86_64 libcpuid; + checkDeps = [ + gtest + rapidcheck + ]; + + internalApiDocsDeps = [ + buildPackages.doxygen + ]; + awsDeps = lib.optional (stdenv.isLinux || stdenv.isDarwin) (aws-sdk-cpp.override { apis = ["s3" "transfer"]; @@ -133,13 +158,14 @@ patches = (o.patches or []) ++ [ ./boehmgc-coroutine-sp-fallback.diff ]; - })) + }) + ) nlohmann_json ]; }; installScriptFor = systems: - with nixpkgsFor.x86_64-linux; + with nixpkgsFor.x86_64-linux.native; runCommand "installer-script" { buildInputs = [ nix ]; } @@ -188,7 +214,7 @@ VERSION_SUFFIX = versionSuffix; nativeBuildInputs = nativeBuildDeps; - buildInputs = buildDeps ++ awsDeps; + buildInputs = buildDeps ++ awsDeps ++ checkDeps; propagatedBuildInputs = propagatedDeps; enableParallelBuilding = true; @@ -203,8 +229,9 @@ installCheckPhase = "make installcheck -j$NIX_BUILD_CORES -l$NIX_BUILD_CORES"; }; - binaryTarball = buildPackages: nix: pkgs: + binaryTarball = nix: pkgs: let + inherit (pkgs) buildPackages; inherit (pkgs) cacert; installerClosureInfo = buildPackages.closureInfo { rootPaths = [ nix cacert ]; }; in @@ -260,6 +287,7 @@ echo "file binary-dist $fn" >> $out/nix-support/hydra-build-products tar cvfJ $fn \ --owner=0 --group=0 --mode=u+rw,uga+r \ + --mtime='1970-01-01' \ --absolute-names \ --hard-dereference \ --transform "s,$TMPDIR/install,$dir/install," \ @@ -283,7 +311,15 @@ # Forward from the previous stage as we don’t want it to pick the lowdown override nixUnstable = prev.nixUnstable; - nix = with final; with commonDeps { inherit pkgs; }; currentStdenv.mkDerivation { + nix = + with final; + with commonDeps { + inherit pkgs; + inherit (currentStdenv.hostPlatform) isStatic; + }; + let + canRunInstalled = currentStdenv.buildPlatform.canExecute currentStdenv.hostPlatform; + in currentStdenv.mkDerivation (finalAttrs: { name = "nix-${version}"; inherit version; @@ -294,24 +330,27 @@ outputs = [ "out" "dev" "doc" ]; nativeBuildInputs = nativeBuildDeps; - buildInputs = buildDeps ++ awsDeps; + buildInputs = buildDeps + # There have been issues building these dependencies + ++ lib.optionals (currentStdenv.hostPlatform == currentStdenv.buildPlatform) awsDeps + ++ lib.optionals finalAttrs.doCheck checkDeps; propagatedBuildInputs = propagatedDeps; disallowedReferences = [ boost ]; - preConfigure = + preConfigure = lib.optionalString (! currentStdenv.hostPlatform.isStatic) '' # Copy libboost_context so we don't get all of Boost in our closure. # https://github.com/NixOS/nixpkgs/issues/45462 mkdir -p $out/lib cp -pd ${boost}/lib/{libboost_context*,libboost_thread*,libboost_system*} $out/lib rm -f $out/lib/*.a - ${lib.optionalString currentStdenv.isLinux '' + ${lib.optionalString currentStdenv.hostPlatform.isLinux '' chmod u+w $out/lib/*.so.* patchelf --set-rpath $out/lib:${currentStdenv.cc.cc.lib}/lib $out/lib/libboost_thread.so.* ''} - ${lib.optionalString currentStdenv.isDarwin '' + ${lib.optionalString currentStdenv.hostPlatform.isDarwin '' for LIB in $out/lib/*.dylib; do chmod u+w $LIB install_name_tool -id $LIB $LIB @@ -322,7 +361,11 @@ ''; configureFlags = configureFlags ++ - [ "--sysconfdir=/etc" ]; + [ "--sysconfdir=/etc" ] ++ + lib.optional stdenv.hostPlatform.isStatic "--enable-embedded-sandbox-shell" ++ + [ (lib.enableFeature finalAttrs.doCheck "tests") ] ++ + lib.optionals finalAttrs.doCheck testConfigureFlags ++ + lib.optional (!canRunInstalled) "--disable-doc-gen"; enableParallelBuilding = true; @@ -335,6 +378,10 @@ postInstall = '' mkdir -p $doc/nix-support echo "doc manual $doc/share/doc/nix/manual" >> $doc/nix-support/hydra-build-products + ${lib.optionalString currentStdenv.hostPlatform.isStatic '' + mkdir -p $out/nix-support + echo "file binary-dist $out/bin/nix" >> $out/nix-support/hydra-build-products + ''} ${lib.optionalString currentStdenv.isDarwin '' install_name_tool \ -change ${boost}/lib/libboost_context.dylib \ @@ -343,13 +390,16 @@ ''} ''; - doInstallCheck = true; + doInstallCheck = finalAttrs.doCheck; installCheckFlags = "sysconfdir=$(out)/etc"; + installCheckTarget = "installcheck"; # work around buggy detection in stdenv - separateDebugInfo = true; + separateDebugInfo = !currentStdenv.hostPlatform.isStatic; strictDeps = true; + hardeningDisable = lib.optional stdenv.hostPlatform.isStatic "pie"; + passthru.perl-bindings = with final; perl.pkgs.toPerlModule (currentStdenv.mkDerivation { name = "nix-perl-${version}"; @@ -363,7 +413,7 @@ buildInputs = [ nix - (curl.override { patchNetrcRegression = true; }) + curl bzip2 xz pkgs.perl @@ -382,8 +432,8 @@ postUnpack = "sourceRoot=$sourceRoot/perl"; }); - meta.platforms = systems; - }; + meta.platforms = lib.platforms.unix; + }); lowdown-nix = with final; currentStdenv.mkDerivation rec { name = "lowdown-0.9.0"; @@ -403,7 +453,20 @@ }; }; + nixos-lib = import (nixpkgs + "/nixos/lib") { }; + + # https://nixos.org/manual/nixos/unstable/index.html#sec-calling-nixos-tests + runNixOSTestFor = system: test: nixos-lib.runTest { + imports = [ test ]; + hostPkgs = nixpkgsFor.${system}.native; + defaults = { + nixpkgs.pkgs = nixpkgsFor.${system}.native; + }; + _module.args.nixpkgs = nixpkgs; + }; + in { + inherit nixpkgsFor; # A Nixpkgs overlay that overrides the 'nix' and # 'nix.perl-bindings' packages. @@ -412,30 +475,36 @@ hydraJobs = { # Binary package for various platforms. - build = nixpkgs.lib.genAttrs systems (system: self.packages.${system}.nix); + build = forAllSystems (system: self.packages.${system}.nix); + + buildStatic = lib.genAttrs linux64BitSystems (system: self.packages.${system}.nix-static); + + buildCross = forAllCrossSystems (crossSystem: + lib.genAttrs ["x86_64-linux"] (system: self.packages.${system}."nix-${crossSystem}")); - buildStatic = nixpkgs.lib.genAttrs linux64BitSystems (system: self.packages.${system}.nix-static); + buildNoGc = forAllSystems (system: self.packages.${system}.nix.overrideAttrs (a: { configureFlags = (a.configureFlags or []) ++ ["--enable-gc=no"];})); - buildCross = nixpkgs.lib.genAttrs crossSystems (crossSystem: - nixpkgs.lib.genAttrs ["x86_64-linux"] (system: self.packages.${system}."nix-${crossSystem}")); + buildNoTests = forAllSystems (system: + self.packages.${system}.nix.overrideAttrs (a: { + doCheck = + assert ! a?dontCheck; + false; + }) + ); # Perl bindings for various platforms. - perlBindings = nixpkgs.lib.genAttrs systems (system: self.packages.${system}.nix.perl-bindings); + perlBindings = forAllSystems (system: nixpkgsFor.${system}.native.nix.perl-bindings); # Binary tarball for various platforms, containing a Nix store # with the closure of 'nix' package, and the second half of # the installation script. - binaryTarball = nixpkgs.lib.genAttrs systems (system: binaryTarball nixpkgsFor.${system} nixpkgsFor.${system}.nix nixpkgsFor.${system}); - - binaryTarballCross = nixpkgs.lib.genAttrs ["x86_64-linux"] (system: builtins.listToAttrs (map (crossSystem: { - name = crossSystem; - value = let - nixpkgsCross = import nixpkgs { - inherit system crossSystem; - overlays = [ self.overlays.default ]; - }; - in binaryTarball nixpkgsFor.${system} self.packages.${system}."nix-${crossSystem}" nixpkgsCross; - }) crossSystems)); + binaryTarball = forAllSystems (system: binaryTarball nixpkgsFor.${system}.native.nix nixpkgsFor.${system}.native); + + binaryTarballCross = lib.genAttrs ["x86_64-linux"] (system: + forAllCrossSystems (crossSystem: + binaryTarball + self.packages.${system}."nix-${crossSystem}" + nixpkgsFor.${system}.cross.${crossSystem})); # The first half of the installation script. This is uploaded # to https://nixos.org/nix/install. It downloads the binary @@ -445,11 +514,11 @@ installerScriptForGHA = installScriptFor [ "x86_64-linux" "x86_64-darwin" "armv6l-linux" "armv7l-linux"]; # docker image with Nix inside - dockerImage = nixpkgs.lib.genAttrs linux64BitSystems (system: self.packages.${system}.dockerImage); + dockerImage = lib.genAttrs linux64BitSystems (system: self.packages.${system}.dockerImage); # Line coverage analysis. coverage = - with nixpkgsFor.x86_64-linux; + with nixpkgsFor.x86_64-linux.native; with commonDeps { inherit pkgs; }; releaseTools.coverageAnalysis { @@ -457,66 +526,74 @@ src = self; + configureFlags = testConfigureFlags; + enableParallelBuilding = true; nativeBuildInputs = nativeBuildDeps; - buildInputs = buildDeps ++ propagatedDeps ++ awsDeps; + buildInputs = buildDeps ++ propagatedDeps ++ awsDeps ++ checkDeps; dontInstall = false; doInstallCheck = true; + installCheckTarget = "installcheck"; # work around buggy detection in stdenv lcovFilter = [ "*/boost/*" "*-tab.*" ]; - # We call `dot', and even though we just use it to - # syntax-check generated dot files, it still requires some - # fonts. So provide those. - FONTCONFIG_FILE = texFunctions.fontsConf; + hardeningDisable = ["fortify"]; + }; + + # API docs for Nix's unstable internal C++ interfaces. + internal-api-docs = + with nixpkgsFor.x86_64-linux.native; + with commonDeps { inherit pkgs; }; + + stdenv.mkDerivation { + pname = "nix-internal-api-docs"; + inherit version; + + src = self; + + configureFlags = testConfigureFlags ++ internalApiDocsConfigureFlags; + + nativeBuildInputs = nativeBuildDeps; + buildInputs = buildDeps ++ propagatedDeps + ++ awsDeps ++ checkDeps ++ internalApiDocsDeps; + + dontBuild = true; + + installTargets = [ "internal-api-html" ]; + + postInstall = '' + mkdir -p $out/nix-support + echo "doc internal-api-docs $out/share/doc/nix/internal-api/html" >> $out/nix-support/hydra-build-products + ''; }; # System tests. - tests.remoteBuilds = import ./tests/remote-builds.nix { - system = "x86_64-linux"; - inherit nixpkgs; - overlay = self.overlays.default; - }; + tests.authorization = runNixOSTestFor "x86_64-linux" ./tests/nixos/authorization.nix; - tests.nix-copy-closure = import ./tests/nix-copy-closure.nix { - system = "x86_64-linux"; - inherit nixpkgs; - overlay = self.overlays.default; - }; + tests.remoteBuilds = runNixOSTestFor "x86_64-linux" ./tests/nixos/remote-builds.nix; - tests.nssPreload = (import ./tests/nss-preload.nix rec { - system = "x86_64-linux"; - inherit nixpkgs; - overlay = self.overlays.default; - }); + tests.nix-copy-closure = runNixOSTestFor "x86_64-linux" ./tests/nixos/nix-copy-closure.nix; - tests.githubFlakes = (import ./tests/github-flakes.nix rec { - system = "x86_64-linux"; - inherit nixpkgs; - overlay = self.overlays.default; - }); + tests.nssPreload = runNixOSTestFor "x86_64-linux" ./tests/nixos/nss-preload.nix; - tests.sourcehutFlakes = (import ./tests/sourcehut-flakes.nix rec { - system = "x86_64-linux"; - inherit nixpkgs; - overlay = self.overlays.default; - }); + tests.githubFlakes = runNixOSTestFor "x86_64-linux" ./tests/nixos/github-flakes.nix; + + tests.sourcehutFlakes = runNixOSTestFor "x86_64-linux" ./tests/nixos/sourcehut-flakes.nix; - tests.setuid = nixpkgs.lib.genAttrs + tests.containers = runNixOSTestFor "x86_64-linux" ./tests/nixos/containers/containers.nix; + + tests.setuid = lib.genAttrs ["i686-linux" "x86_64-linux"] - (system: - import ./tests/setuid.nix rec { - inherit nixpkgs system; - overlay = self.overlays.default; - }); + (system: runNixOSTestFor system ./tests/nixos/setuid.nix); + # Make sure that nix-env still produces the exact same result # on a particular version of Nixpkgs. tests.evalNixpkgs = - with nixpkgsFor.x86_64-linux; + with nixpkgsFor.x86_64-linux.native; runCommand "eval-nixos" { buildInputs = [ nix ]; } '' type -p nix-env @@ -526,13 +603,19 @@ mkdir $out ''; + tests.nixpkgsLibTests = + forAllSystems (system: + import (nixpkgs + "/lib/tests/release.nix") + { pkgs = nixpkgsFor.${system}.native; } + ); + metrics.nixpkgs = import "${nixpkgs-regression}/pkgs/top-level/metrics.nix" { - pkgs = nixpkgsFor.x86_64-linux; + pkgs = nixpkgsFor.x86_64-linux.native; nixpkgs = nixpkgs-regression; }; installTests = forAllSystems (system: - let pkgs = nixpkgsFor.${system}; in + let pkgs = nixpkgsFor.${system}.native; in pkgs.runCommand "install-tests" { againstSelf = testNixVersions pkgs pkgs.nix pkgs.pkgs.nix; againstCurrentUnstable = @@ -545,73 +628,30 @@ # againstLatestStable = testNixVersions pkgs pkgs.nix pkgs.nixStable; } "touch $out"); + installerTests = import ./tests/installer { + binaryTarballs = self.hydraJobs.binaryTarball; + inherit nixpkgsFor; + }; + }; checks = forAllSystems (system: { binaryTarball = self.hydraJobs.binaryTarball.${system}; perlBindings = self.hydraJobs.perlBindings.${system}; installTests = self.hydraJobs.installTests.${system}; - } // (nixpkgs.lib.optionalAttrs (builtins.elem system linux64BitSystems)) { + nixpkgsLibTests = self.hydraJobs.tests.nixpkgsLibTests.${system}; + } // (lib.optionalAttrs (builtins.elem system linux64BitSystems)) { dockerImage = self.hydraJobs.dockerImage.${system}; }); packages = forAllSystems (system: rec { - inherit (nixpkgsFor.${system}) nix; + inherit (nixpkgsFor.${system}.native) nix; default = nix; - } // (nixpkgs.lib.optionalAttrs (builtins.elem system linux64BitSystems) { - nix-static = let - nixpkgs = nixpkgsFor.${system}.pkgsStatic; - in with commonDeps { pkgs = nixpkgs; isStatic = true; }; nixpkgs.stdenv.mkDerivation { - name = "nix-${version}"; - - src = self; - - VERSION_SUFFIX = versionSuffix; - - outputs = [ "out" "dev" "doc" ]; - - nativeBuildInputs = nativeBuildDeps; - buildInputs = buildDeps ++ propagatedDeps; - - # Work around pkgsStatic disabling all tests. - # Remove in NixOS 22.11, see https://github.com/NixOS/nixpkgs/pull/140271. - preHook = - '' - doCheck=1 - doInstallCheck=1 - ''; - - configureFlags = - configureFlags ++ - [ "--sysconfdir=/etc" - "--enable-embedded-sandbox-shell" - ]; - - enableParallelBuilding = true; - - makeFlags = "profiledir=$(out)/etc/profile.d"; - - installFlags = "sysconfdir=$(out)/etc"; - - postInstall = '' - mkdir -p $doc/nix-support - echo "doc manual $doc/share/doc/nix/manual" >> $doc/nix-support/hydra-build-products - mkdir -p $out/nix-support - echo "file binary-dist $out/bin/nix" >> $out/nix-support/hydra-build-products - ''; - - installCheckFlags = "sysconfdir=$(out)/etc"; - - stripAllList = ["bin"]; - - strictDeps = true; - - hardeningDisable = [ "pie" ]; - }; - + } // (lib.optionalAttrs (builtins.elem system linux64BitSystems) { + nix-static = nixpkgsFor.${system}.static.nix; dockerImage = let - pkgs = nixpkgsFor.${system}; + pkgs = nixpkgsFor.${system}.native; image = import ./docker.nix { inherit pkgs; tag = version; }; in pkgs.runCommand @@ -623,68 +663,35 @@ ln -s ${image} $image echo "file binary-dist $image" >> $out/nix-support/hydra-build-products ''; - } - - // builtins.listToAttrs (map (crossSystem: { - name = "nix-${crossSystem}"; - value = let - nixpkgsCross = import nixpkgs { - inherit system crossSystem; - overlays = [ self.overlays.default ]; - }; - in with commonDeps { pkgs = nixpkgsCross; }; nixpkgsCross.stdenv.mkDerivation { - name = "nix-${version}"; - - src = self; - - VERSION_SUFFIX = versionSuffix; - - outputs = [ "out" "dev" "doc" ]; - - nativeBuildInputs = nativeBuildDeps; - buildInputs = buildDeps ++ propagatedDeps; - - configureFlags = [ "--sysconfdir=/etc" "--disable-doc-gen" ]; - - enableParallelBuilding = true; - - makeFlags = "profiledir=$(out)/etc/profile.d"; - - doCheck = true; - - installFlags = "sysconfdir=$(out)/etc"; - - postInstall = '' - mkdir -p $doc/nix-support - echo "doc manual $doc/share/doc/nix/manual" >> $doc/nix-support/hydra-build-products - mkdir -p $out/nix-support - echo "file binary-dist $out/bin/nix" >> $out/nix-support/hydra-build-products - ''; - - doInstallCheck = true; - installCheckFlags = "sysconfdir=$(out)/etc"; - }; - }) (if system == "x86_64-linux" then crossSystems else []))) - - // (builtins.listToAttrs (map (stdenvName: - nixpkgsFor.${system}.lib.nameValuePair - "nix-${stdenvName}" - nixpkgsFor.${system}."${stdenvName}Packages".nix - ) stdenvs))); - - devShells = forAllSystems (system: - forAllStdenvs (stdenv: - with nixpkgsFor.${system}; + } // builtins.listToAttrs (map + (crossSystem: { + name = "nix-${crossSystem}"; + value = nixpkgsFor.${system}.cross.${crossSystem}.nix; + }) + crossSystems) + // builtins.listToAttrs (map + (stdenvName: { + name = "nix-${stdenvName}"; + value = nixpkgsFor.${system}.stdenvs."${stdenvName}Packages".nix; + }) + stdenvs))); + + devShells = let + makeShell = pkgs: stdenv: with commonDeps { inherit pkgs; }; - nixpkgsFor.${system}.${stdenv}.mkDerivation { + stdenv.mkDerivation { name = "nix"; outputs = [ "out" "dev" "doc" ]; - nativeBuildInputs = nativeBuildDeps; - buildInputs = buildDeps ++ propagatedDeps ++ awsDeps; + nativeBuildInputs = nativeBuildDeps + ++ (lib.optionals stdenv.cc.isClang [ pkgs.bear pkgs.clang-tools ]); + + buildInputs = buildDeps ++ propagatedDeps + ++ awsDeps ++ checkDeps ++ internalApiDocsDeps; - inherit configureFlags; + configureFlags = configureFlags + ++ testConfigureFlags ++ internalApiDocsConfigureFlags; enableParallelBuilding = true; @@ -699,10 +706,21 @@ # Make bash completion work. XDG_DATA_DIRS+=:$out/share ''; - } - ) - // { default = self.devShells.${system}.stdenv; } - ); - + }; + in + forAllSystems (system: + let + makeShells = prefix: pkgs: + lib.mapAttrs' + (k: v: lib.nameValuePair "${prefix}-${k}" v) + (forAllStdenvs (stdenvName: makeShell pkgs pkgs.${stdenvName})); + in + (makeShells "native" nixpkgsFor.${system}.native) // + (makeShells "static" nixpkgsFor.${system}.static) // + (forAllCrossSystems (crossSystem: let pkgs = nixpkgsFor.${system}.cross.${crossSystem}; in makeShell pkgs pkgs.stdenv)) // + { + default = self.devShells.${system}.native-stdenvPackages; + } + ); }; } diff --git a/maintainers/README.md b/maintainers/README.md new file mode 100644 index 000000000..476a5f51e --- /dev/null +++ b/maintainers/README.md @@ -0,0 +1,131 @@ +# Nix maintainers team + +## Motivation + +The team's main responsibility is to set a direction for the development of Nix and ensure that the code is in good shape. + +We aim to achieve this by improving the contributor experience and attracting more maintainers – that is, by helping other people contributing to Nix and eventually taking responsibility – in order to scale the development process to match users' needs. + +### Objectives + +- It is obvious what is worthwhile to work on. +- It is easy to find the right place in the code to make a change. +- It is clear what is expected of a pull request. +- It is predictable how to get a change merged and released. + +### Tasks + +- Establish, communicate, and maintain a technical roadmap +- Improve documentation targeted at contributors + - Record architecture and design decisions + - Elaborate contribution guides and abide to them + - Define and assert quality criteria for contributions +- Maintain the issue tracker and triage pull requests +- Help contributors succeed with pull requests that address roadmap milestones +- Manage the release lifecycle +- Regularly publish reports on work done +- Engage with third parties in the interest of the project +- Ensure the required maintainer capacity for all of the above + +## Members + +- Eelco Dolstra (@edolstra) – Team lead +- Théophane Hufschmitt (@thufschmitt) +- Valentin Gagarin (@fricklerhandwerk) +- Thomas Bereknyei (@tomberek) +- Robert Hensing (@roberth) +- John Ericson (@Ericson2314) + +## Meeting protocol + +The team meets twice a week: + +- Discussion meeting: [Fridays 13:00-14:00 CET](https://calendar.google.com/calendar/event?eid=MHNtOGVuNWtrZXNpZHR2bW1sM3QyN2ZjaGNfMjAyMjExMjVUMTIwMDAwWiBiOW81MmZvYnFqYWs4b3E4bGZraGczdDBxZ0Bn) + + 1. Triage issues and pull requests from the _No Status_ column (30 min) + 2. Discuss issues and pull requests from the _To discuss_ column (30 min) + +- Work meeting: [Mondays 13:00-15:00 CET](https://calendar.google.com/calendar/event?eid=NTM1MG1wNGJnOGpmOTZhYms3bTB1bnY5cWxfMjAyMjExMjFUMTIwMDAwWiBiOW81MmZvYnFqYWs4b3E4bGZraGczdDBxZ0Bn) + + 1. Code review on pull requests from _In review_. + 2. Other chores and tasks. + +Meeting notes are collected on a [collaborative scratchpad](https://pad.lassul.us/Cv7FpYx-Ri-4VjUykQOLAw), and published on Discourse under the [Nix category](https://discourse.nixos.org/c/dev/nix/50). + +## Project board protocol + +The team uses a [GitHub project board](https://github.com/orgs/NixOS/projects/19/views/1) for tracking its work. + +Issues on the board progress through the following states: + +- No Status + + During the discussion meeting, the team triages new items. + To be considered, issues and pull requests must have a high-level description to provide the whole team with the necessary context at a glance. + + On every meeting, at least one item from each of the following categories is inspected: + + 1. [critical](https://github.com/NixOS/nix/labels/critical) + 2. [security](https://github.com/NixOS/nix/labels/security) + 3. [regression](https://github.com/NixOS/nix/labels/regression) + 4. [bug](https://github.com/NixOS/nix/issues?q=is%3Aopen+label%3Abug+sort%3Areactions-%2B1-desc) + + - [oldest pull requests](https://github.com/NixOS/nix/pulls?q=is%3Apr+is%3Aopen+sort%3Acreated-asc) + - [most popular pull requests](https://github.com/NixOS/nix/pulls?q=is%3Apr+is%3Aopen+sort%3Areactions-%2B1-desc) + - [oldest issues](https://github.com/NixOS/nix/issues?q=is%3Aissue+is%3Aopen+sort%3Acreated-asc) + - [most popular issues](https://github.com/NixOS/nix/issues?q=is%3Aissue+is%3Aopen+sort%3Areactions-%2B1-desc) + + Team members can also add pull requests or issues they would like the whole team to consider. + + If there is disagreement on the general idea behind an issue or pull request, it is moved to _To discuss_, otherwise to _In review_. + +- To discuss + + Pull requests and issues that are deemed important and controversial are discussed by the team during discussion meetings. + + This may be where the merit of the change itself or the implementation strategy is contested by a team member. + + As a general guideline, the order of items is determined as follows: + + - Prioritise pull requests over issues + + Contributors who took the time to implement concrete change proposals should not wait indefinitely. + + - Prioritise fixing bugs over documentation, improvements or new features + + The team values stability and accessibility higher than raw functionality. + + - Interleave issues and PRs + + This way issues without attempts at a solution get a chance to get addressed. + +- In review + + Pull requests in this column are reviewed together during work meetings. + This is both for spreading implementation knowledge and for establishing common values in code reviews. + + When the overall direction is agreed upon, even when further changes are required, the pull request is assigned to one team member. + +- Assigned for merging + + One team member is assigned to each of these pull requests. + They will communicate with the authors, and make the final approval once all remaining issues are addressed. + + If more substantive issues arise, the assignee can move the pull request back to _To discuss_ to involve the team again. + +The process is illustrated in the following diagram: + +```mermaid +flowchart TD + discuss[To discuss] + + review[To review] + + New --> |Disagreement on idea| discuss + New & discuss --> |Consensus on idea| review + + review --> |Consensus on implementation| Assigned + + Assigned --> |Implementation issues arise| review + Assigned --> |Remaining issues fixed| Merged +``` diff --git a/maintainers/backporting.md b/maintainers/backporting.md new file mode 100644 index 000000000..2424050c8 --- /dev/null +++ b/maintainers/backporting.md @@ -0,0 +1,12 @@ + +# Backporting + +To [automatically backport a pull request](https://github.com/NixOS/nix/blob/master/.github/workflows/backport.yml) to a release branch once it's merged, assign it a label of the form [`backport <branch>`](https://github.com/NixOS/nix/labels?q=backport). + +Since [GitHub Actions workflows will not trigger other workflows](https://docs.github.com/en/actions/using-workflows/triggering-a-workflow#triggering-a-workflow-from-a-workflow), checks on the automatic backport need to be triggered by another actor. +This is achieved by closing and reopening the backport pull request. + +This specifically affects the [`installer_test`] check. +Note that it only runs after the other tests, so it may take a while to appear. + +[`installer_test`]: https://github.com/NixOS/nix/blob/895dfc656a21f6252ddf48df0d1f215effa04ecb/.github/workflows/ci.yml#L70-L91 diff --git a/maintainers/release-process.md b/maintainers/release-process.md new file mode 100644 index 000000000..ec9e96489 --- /dev/null +++ b/maintainers/release-process.md @@ -0,0 +1,182 @@ +# Nix release process + +## Release artifacts + +The release process is intended to create the following for each +release: + +* A Git tag + +* Binary tarballs in https://releases.nixos.org/?prefix=nix/ + +* Docker images + +* Closures in https://cache.nixos.org + +* (Optionally) Updated `fallback-paths.nix` in Nixpkgs + +* An updated manual on https://nixos.org/manual/nix/stable/ + +## Creating a new release from the `master` branch + +* Make sure that the [Hydra `master` jobset](https://hydra.nixos.org/jobset/nix/master) succeeds. + +* In a checkout of the Nix repo, make sure you're on `master` and run + `git pull`. + +* Move the contents of `doc/manual/src/release-notes/rl-next.md` + (except the first line) to + `doc/manual/src/release-notes/rl-$VERSION.md` (where `$VERSION` is + the contents of `.version` *without* the patch level, e.g. `2.12` + rather than `2.12.0`). + +* Add a header to `doc/manual/src/release-notes/rl-$VERSION.md` like + + ``` + # Release 2.12 (2022-12-06) + ``` + +* Proof-read / edit / rearrange the release notes. Breaking changes + and highlights should go to the top. + +* Add a link to the release notes to `doc/manual/src/SUMMARY.md.in` + (*not* `SUMMARY.md`), e.g. + + ``` + - [Release 2.12 (2022-12-06)](release-notes/rl-2.12.md) + ``` + +* Run + + ```console + $ git checkout -b release-notes + $ git add doc/manual/src/release-notes/rl-$VERSION.md + $ git commit -a -m 'Release notes' + $ git push --set-upstream $REMOTE release-notes + ``` + +* Create a PR for `release-notes`. + +* Wait for the PR to be merged. + +* Create a branch for the release: + + ```console + $ git checkout master + $ git pull + $ git checkout -b $VERSION-maintenance + ``` + +* Mark the release as stable: + + ```console + $ git cherry-pick f673551e71942a52b6d7ae66af8b67140904a76a + ``` + + This removes the link to `rl-next.md` from the manual and sets + `officialRelease = true` in `flake.nix`. + +* Push the release branch: + + ```console + $ git push --set-upstream origin $VERSION-maintenance + ``` + +* Create a jobset for the release branch on Hydra as follows: + + * Go to the jobset of the previous release + (e.g. https://hydra.nixos.org/jobset/nix/maintenance-2.11). + + * Select `Actions -> Clone this jobset`. + + * Set identifier to `maintenance-$VERSION`. + + * Set description to `$VERSION release branch`. + + * Set flake URL to `github:NixOS/nix/$VERSION-maintenance`. + + * Hit `Create jobset`. + +* Wait for the new jobset to evaluate and build. If impatient, go to + the evaluation and select `Actions -> Bump builds to front of + queue`. + +* When the jobset evaluation has succeeded building, take note of the + evaluation ID (e.g. `1780832` in + `https://hydra.nixos.org/eval/1780832`). + +* Tag the release and upload the release artifacts to + [`releases.nixos.org`](https://releases.nixos.org/) and [Docker Hub](https://hub.docker.com/): + + ```console + $ IS_LATEST=1 ./maintainers/upload-release.pl <EVAL-ID> + ``` + + Note: `IS_LATEST=1` causes the `latest-release` branch to be + force-updated. This is used by the `nixos.org` website to get the + [latest Nix manual](https://nixos.org/manual/nixpkgs/unstable/). + + TODO: This script requires the right AWS credentials. Document. + + TODO: This script currently requires a + `/home/eelco/Dev/nix-pristine` and + `/home/eelco/Dev/nixpkgs-pristine`. + + TODO: trigger nixos.org netlify: https://docs.netlify.com/configure-builds/build-hooks/ + +* Prepare for the next point release by editing `.version` to + e.g. + + ```console + $ echo 2.12.1 > .version + $ git commit -a -m 'Bump version' + $ git push + ``` + + Commit and push this to the maintenance branch. + +* Bump the version of `master`: + + ```console + $ git checkout master + $ git pull + $ NEW_VERSION=2.13.0 + $ echo -n $NEW_VERSION > .version + $ git checkout -b bump-$NEW_VERSION + $ git commit -a -m 'Bump version' + $ git push --set-upstream origin bump-$NEW_VERSION + ``` + + Make a pull request and auto-merge it. + +* Create a milestone for the next release, move all unresolved issues + from the previous milestone, and close the previous milestone. Set + the date for the next milestone 6 weeks from now. + +* Create a backport label. + +* Post an [announcement on Discourse](https://discourse.nixos.org/c/announcements/8), including the contents of + `rl-$VERSION.md`. + +## Creating a point release + +* Wait for the desired evaluation of the maintenance jobset to finish + building. + +* Run + + ```console + $ IS_LATEST=1 ./maintainers/upload-release.pl <EVAL-ID> + ``` + + Omit `IS_LATEST=1` when creating a point release that is not on the + most recent stable branch. This prevents `nixos.org` to going back + to an older release. + +* Bump the version number of the release branch as above (e.g. to + `2.12.2`). + +## Recovering from mistakes + +`upload-release.pl` should be idempotent. For instance a wrong `IS_LATEST` value can be fixed that way, by running the script on the actual latest release. + diff --git a/maintainers/upload-release.pl b/maintainers/upload-release.pl index d3ef63db8..77469148a 100755 --- a/maintainers/upload-release.pl +++ b/maintainers/upload-release.pl @@ -115,10 +115,6 @@ sub downloadFile { write_file("$tmpFile.sha256", $sha256_actual); - if (! -e "$tmpFile.asc") { - system("gpg2 --detach-sign --armor $tmpFile") == 0 or die "unable to sign $tmpFile\n"; - } - return $sha256_expected; } @@ -194,7 +190,7 @@ for my $fn (glob "$tmpDir/*") { my $configuration = (); $configuration->{content_type} = "application/octet-stream"; - if ($fn =~ /.sha256|.asc|install/) { + if ($fn =~ /.sha256|install/) { # Text files $configuration->{content_type} = "text/plain"; } diff --git a/misc/launchd/org.nixos.nix-daemon.plist.in b/misc/launchd/org.nixos.nix-daemon.plist.in index da1970f69..5fa489b20 100644 --- a/misc/launchd/org.nixos.nix-daemon.plist.in +++ b/misc/launchd/org.nixos.nix-daemon.plist.in @@ -28,7 +28,7 @@ <key>SoftResourceLimits</key> <dict> <key>NumberOfFiles</key> - <integer>4096</integer> + <integer>1048576</integer> </dict> </dict> </plist> diff --git a/misc/systemd/nix-daemon.service.in b/misc/systemd/nix-daemon.service.in index e3ac42beb..f46413630 100644 --- a/misc/systemd/nix-daemon.service.in +++ b/misc/systemd/nix-daemon.service.in @@ -9,7 +9,7 @@ ConditionPathIsReadWrite=@localstatedir@/nix/daemon-socket [Service] ExecStart=@@bindir@/nix-daemon nix-daemon --daemon KillMode=process -LimitNOFILE=4096 +LimitNOFILE=1048576 [Install] WantedBy=multi-user.target diff --git a/misc/zsh/completion.zsh b/misc/zsh/completion.zsh index e702c721e..f9b3dca74 100644 --- a/misc/zsh/completion.zsh +++ b/misc/zsh/completion.zsh @@ -10,14 +10,15 @@ function _nix() { local -a suggestions declare -a suggestions for suggestion in ${res:1}; do - # FIXME: This doesn't work properly if the suggestion word contains a `:` - # itself - suggestions+="${suggestion/ /:}" + suggestions+=("${suggestion%% *}") done + local -a args if [[ "$tpe" == filenames ]]; then - compadd -f + args+=('-f') + elif [[ "$tpe" == attrs ]]; then + args+=('-S' '') fi - _describe 'nix' suggestions + compadd -J nix "${args[@]}" -a suggestions } _nix "$@" diff --git a/mk/common-test.sh b/mk/common-test.sh new file mode 100644 index 000000000..0a2e4c1c2 --- /dev/null +++ b/mk/common-test.sh @@ -0,0 +1,11 @@ +TESTS_ENVIRONMENT=("TEST_NAME=${test%.*}" 'NIX_REMOTE=') + +: ${BASH:=/usr/bin/env bash} + +init_test () { + cd tests && env "${TESTS_ENVIRONMENT[@]}" $BASH -e init.sh 2>/dev/null > /dev/null +} + +run_test_proper () { + cd $(dirname $test) && env "${TESTS_ENVIRONMENT[@]}" $BASH -e $(basename $test) +} diff --git a/mk/debug-test.sh b/mk/debug-test.sh new file mode 100755 index 000000000..b5b628ecd --- /dev/null +++ b/mk/debug-test.sh @@ -0,0 +1,11 @@ +#!/usr/bin/env bash + +set -eu -o pipefail + +test=$1 + +dir="$(dirname "${BASH_SOURCE[0]}")" +source "$dir/common-test.sh" + +(init_test) +run_test_proper diff --git a/mk/disable-tests.mk b/mk/disable-tests.mk new file mode 100644 index 000000000..f72f84412 --- /dev/null +++ b/mk/disable-tests.mk @@ -0,0 +1,12 @@ +# This file is only active for `./configure --disable-tests`. +# Running `make check` or `make installcheck` would indicate a mistake in the +# caller. + +installcheck: + @echo "Tests are disabled. Configure without '--disable-tests', or avoid calling 'make installcheck'." + @exit 1 + +# This currently has little effect. +check: + @echo "Tests are disabled. Configure without '--disable-tests', or avoid calling 'make check'." + @exit 1 diff --git a/mk/libraries.mk b/mk/libraries.mk index 6541775f3..1bc73d7f7 100644 --- a/mk/libraries.mk +++ b/mk/libraries.mk @@ -67,6 +67,7 @@ define build-library $(1)_LDFLAGS_USE := $(1)_LDFLAGS_USE_INSTALLED := + $(1)_LIB_CLOSURE := $(1) $$(eval $$(call create-dir, $$(_d))) @@ -125,13 +126,15 @@ define build-library $(1)_PATH := $$(_d)/$$($(1)_NAME).a $$($(1)_PATH): $$($(1)_OBJS) | $$(_d)/ - +$$(trace-ld) $(LD) -Ur -o $$(_d)/$$($(1)_NAME).o $$^ + $$(trace-ld) $(LD) $$(ifndef $(HOST_DARWIN),-U) -r -o $$(_d)/$$($(1)_NAME).o $$^ $$(trace-ar) $(AR) crs $$@ $$(_d)/$$($(1)_NAME).o - $(1)_LDFLAGS_USE += $$($(1)_PATH) $$($(1)_LDFLAGS) + $(1)_LDFLAGS_USE += $$($(1)_PATH) $$($(1)_LDFLAGS) $$(foreach lib, $$($(1)_LIBS), $$($$(lib)_LDFLAGS_USE)) $(1)_INSTALL_PATH := $$(libdir)/$$($(1)_NAME).a + $(1)_LIB_CLOSURE += $$($(1)_LIBS) + endif $(1)_LDFLAGS_USE += $$($(1)_LDFLAGS_PROPAGATED) diff --git a/mk/programs.mk b/mk/programs.mk index 0fc1990f7..1ee1d3fa5 100644 --- a/mk/programs.mk +++ b/mk/programs.mk @@ -3,6 +3,9 @@ programs-list := # Build a program with symbolic name $(1). The program is defined by # various variables prefixed by ‘$(1)_’: # +# - $(1)_NAME: the name of the program (e.g. ‘foo’); defaults to +# $(1). +# # - $(1)_DIR: the directory where the (non-installed) program will be # placed. # @@ -23,11 +26,12 @@ programs-list := # - $(1)_INSTALL_DIR: the directory where the program will be # installed; defaults to $(bindir). define build-program + $(1)_NAME ?= $(1) _d := $(buildprefix)$$($(1)_DIR) _srcs := $$(sort $$(foreach src, $$($(1)_SOURCES), $$(src))) $(1)_OBJS := $$(addprefix $(buildprefix), $$(addsuffix .o, $$(basename $$(_srcs)))) - _libs := $$(foreach lib, $$($(1)_LIBS), $$($$(lib)_PATH)) - $(1)_PATH := $$(_d)/$(1) + _libs := $$(foreach lib, $$($(1)_LIBS), $$(foreach lib2, $$($$(lib)_LIB_CLOSURE), $$($$(lib2)_PATH))) + $(1)_PATH := $$(_d)/$$($(1)_NAME) $$(eval $$(call create-dir, $$(_d))) @@ -38,7 +42,7 @@ define build-program ifdef $(1)_INSTALL_DIR - $(1)_INSTALL_PATH := $$($(1)_INSTALL_DIR)/$(1) + $(1)_INSTALL_PATH := $$($(1)_INSTALL_DIR)/$$($(1)_NAME) $$(eval $$(call create-dir, $$($(1)_INSTALL_DIR))) @@ -54,7 +58,7 @@ define build-program else $(DESTDIR)$$($(1)_INSTALL_PATH): $$($(1)_PATH) | $(DESTDIR)$$($(1)_INSTALL_DIR)/ - install -t $(DESTDIR)$$($(1)_INSTALL_DIR) $$< + +$$(trace-install) install -t $(DESTDIR)$$($(1)_INSTALL_DIR) $$< endif endif diff --git a/mk/run-test.sh b/mk/run-test.sh new file mode 100755 index 000000000..1a1d65930 --- /dev/null +++ b/mk/run-test.sh @@ -0,0 +1,38 @@ +#!/usr/bin/env bash + +set -eu -o pipefail + +red="" +green="" +yellow="" +normal="" + +test=$1 + +dir="$(dirname "${BASH_SOURCE[0]}")" +source "$dir/common-test.sh" + +post_run_msg="ran test $test..." +if [ -t 1 ]; then + red="[31;1m" + green="[32;1m" + yellow="[33;1m" + normal="[m" +fi + +run_test () { + (init_test 2>/dev/null > /dev/null) + log="$(run_test_proper 2>&1)" && status=0 || status=$? +} + +run_test + +if [ $status -eq 0 ]; then + echo "$post_run_msg [${green}PASS$normal]" +elif [ $status -eq 99 ]; then + echo "$post_run_msg [${yellow}SKIP$normal]" +else + echo "$post_run_msg [${red}FAIL$normal]" + echo "$log" | sed 's/^/ /' + exit "$status" +fi diff --git a/mk/run_test.sh b/mk/run_test.sh deleted file mode 100755 index 7e95df2ac..000000000 --- a/mk/run_test.sh +++ /dev/null @@ -1,46 +0,0 @@ -#!/bin/sh - -set -u - -red="" -green="" -yellow="" -normal="" - -post_run_msg="ran test $1..." -if [ -t 1 ]; then - red="[31;1m" - green="[32;1m" - yellow="[33;1m" - normal="[m" -fi - -run_test () { - (cd tests && env ${TESTS_ENVIRONMENT} init.sh 2>/dev/null > /dev/null) - log="$(cd $(dirname $1) && env ${TESTS_ENVIRONMENT} $(basename $1) 2>&1)" - status=$? -} - -run_test "$1" - -# Hack: Retry the test if it fails with “unexpected EOF reading a line” as these -# appear randomly without anyone knowing why. -# See https://github.com/NixOS/nix/issues/3605 for more info -if [[ $status -ne 0 && $status -ne 99 && \ - "$(uname)" == "Darwin" && \ - "$log" =~ "unexpected EOF reading a line" \ -]]; then - echo "$post_run_msg [${yellow}FAIL$normal] (possibly flaky, so will be retried)" - echo "$log" | sed 's/^/ /' - run_test "$1" -fi - -if [ $status -eq 0 ]; then - echo "$post_run_msg [${green}PASS$normal]" -elif [ $status -eq 99 ]; then - echo "$post_run_msg [${yellow}SKIP$normal]" -else - echo "$post_run_msg [${red}FAIL$normal]" - echo "$log" | sed 's/^/ /' - exit "$status" -fi diff --git a/mk/tests.mk b/mk/tests.mk index a2e30a378..3ebbd86e3 100644 --- a/mk/tests.mk +++ b/mk/tests.mk @@ -8,7 +8,11 @@ define run-install-test .PHONY: $1.test $1.test: $1 $(test-deps) - @env TEST_NAME=$(basename $1) TESTS_ENVIRONMENT="$(tests-environment)" mk/run_test.sh $1 < /dev/null + @env BASH=$(bash) $(bash) mk/run-test.sh $1 < /dev/null + + .PHONY: $1.test-debug + $1.test-debug: $1 $(test-deps) + @env BASH=$(bash) $(bash) mk/debug-test.sh $1 < /dev/null endef diff --git a/perl/Makefile b/perl/Makefile index 708f86882..c2c95f255 100644 --- a/perl/Makefile +++ b/perl/Makefile @@ -1,6 +1,6 @@ makefiles = local.mk -GLOBAL_CXXFLAGS += -g -Wall -std=c++17 -I ../src +GLOBAL_CXXFLAGS += -g -Wall -std=c++2a -I ../src -include Makefile.config diff --git a/perl/lib/Nix/Store.xs b/perl/lib/Nix/Store.xs index 54ad1680c..de91dc28d 100644 --- a/perl/lib/Nix/Store.xs +++ b/perl/lib/Nix/Store.xs @@ -26,6 +26,7 @@ static ref<Store> store() static std::shared_ptr<Store> _store; if (!_store) { try { + initLibStore(); loadConfFile(); settings.lockCPU = false; _store = openStore(); diff --git a/scripts/install-darwin-multi-user.sh b/scripts/install-darwin-multi-user.sh index afaa6783b..5111a5dde 100644 --- a/scripts/install-darwin-multi-user.sh +++ b/scripts/install-darwin-multi-user.sh @@ -167,7 +167,7 @@ poly_user_shell_get() { } poly_user_shell_set() { - _sudo "in order to give $1 a safe home directory" \ + _sudo "in order to give $1 a safe shell" \ /usr/bin/dscl . -create "/Users/$1" "UserShell" "$2" } diff --git a/scripts/install-multi-user.sh b/scripts/install-multi-user.sh index 9a18280ef..7c66538b0 100644 --- a/scripts/install-multi-user.sh +++ b/scripts/install-multi-user.sh @@ -37,6 +37,19 @@ readonly PROFILE_TARGETS=("/etc/bashrc" "/etc/profile.d/nix.sh" "/etc/zshrc" "/e readonly PROFILE_BACKUP_SUFFIX=".backup-before-nix" readonly PROFILE_NIX_FILE="$NIX_ROOT/var/nix/profiles/default/etc/profile.d/nix-daemon.sh" +# Fish has different syntax than zsh/bash, treat it separate +readonly PROFILE_FISH_SUFFIX="conf.d/nix.fish" +readonly PROFILE_FISH_PREFIXES=( + # each of these are common values of $__fish_sysconf_dir, + # under which Fish will look for a file named + # $PROFILE_FISH_SUFFIX. + "/etc/fish" # standard + "/usr/local/etc/fish" # their installer .pkg for macOS + "/opt/homebrew/etc/fish" # homebrew + "/opt/local/etc/fish" # macports +) +readonly PROFILE_NIX_FILE_FISH="$NIX_ROOT/var/nix/profiles/default/etc/profile.d/nix-daemon.fish" + readonly NIX_INSTALLED_NIX="@nix@" readonly NIX_INSTALLED_CACERT="@cacert@" #readonly NIX_INSTALLED_NIX="/nix/store/j8dbv5w6jl34caywh2ygdy88knx1mdf7-nix-2.3.6" @@ -45,7 +58,7 @@ readonly EXTRACTED_NIX_PATH="$(dirname "$0")" readonly ROOT_HOME=~root -if [ -t 0 ]; then +if [ -t 0 ] && [ -z "${NIX_INSTALLER_YES:-}" ]; then readonly IS_HEADLESS='no' else readonly IS_HEADLESS='yes' @@ -59,14 +72,35 @@ headless() { fi } +is_root() { + if [ "$EUID" -eq 0 ]; then + return 0 + else + return 1 + fi +} + +is_os_linux() { + if [ "$(uname -s)" = "Linux" ]; then + return 0 + else + return 1 + fi +} + +is_os_darwin() { + if [ "$(uname -s)" = "Darwin" ]; then + return 0 + else + return 1 + fi +} + contact_us() { - echo "You can open an issue at https://github.com/nixos/nix/issues" + echo "You can open an issue at" + echo "https://github.com/NixOS/nix/issues/new?labels=installer&template=installer.md" echo "" - echo "Or feel free to contact the team:" - echo " - Matrix: #nix:nixos.org" - echo " - IRC: in #nixos on irc.libera.chat" - echo " - twitter: @nixos_org" - echo " - forum: https://discourse.nixos.org" + echo "Or get in touch with the community: https://nixos.org/community" } get_help() { echo "We'd love to help if you need it." @@ -102,7 +136,7 @@ EOF cat <<EOF $step. Delete the files Nix added to your system: - sudo rm -rf /etc/nix $NIX_ROOT $ROOT_HOME/.nix-profile $ROOT_HOME/.nix-defexpr $ROOT_HOME/.nix-channels $HOME/.nix-profile $HOME/.nix-defexpr $HOME/.nix-channels + sudo rm -rf "/etc/nix" "$NIX_ROOT" "$ROOT_HOME/.nix-profile" "$ROOT_HOME/.nix-defexpr" "$ROOT_HOME/.nix-channels" "$ROOT_HOME/.local/state/nix" "$ROOT_HOME/.cache/nix" "$HOME/.nix-profile" "$HOME/.nix-defexpr" "$HOME/.nix-channels" "$HOME/.local/state/nix" "$HOME/.cache/nix" and that is it. @@ -313,14 +347,23 @@ __sudo() { _sudo() { local expl="$1" shift - if ! headless; then + if ! headless || is_root; then __sudo "$expl" "$*" >&2 fi - sudo "$@" + + if is_root; then + env "$@" + else + sudo "$@" + fi } +# Ensure that $TMPDIR exists if defined. +if [[ -n "${TMPDIR:-}" ]] && [[ ! -d "${TMPDIR:-}" ]]; then + mkdir -m 0700 -p "${TMPDIR:-}" +fi -readonly SCRATCH=$(mktemp -d "${TMPDIR:-/tmp/}tmp.XXXXXXXXXX") +readonly SCRATCH=$(mktemp -d) finish_cleanup() { rm -rf "$SCRATCH" } @@ -329,7 +372,7 @@ finish_fail() { finish_cleanup failure <<EOF -Jeeze, something went wrong. If you can take all the output and open +Oh no, something went wrong. If you can take all the output and open an issue, we'd love to fix the problem so nobody else has this issue. :( @@ -423,7 +466,7 @@ EOF fi done - if [ "$(uname -s)" = "Linux" ] && [ ! -e /run/systemd/system ]; then + if is_os_linux && [ ! -e /run/systemd/system ]; then warning <<EOF We did not detect systemd on your system. With a multi-user install without systemd you will have to manually configure your init system to @@ -532,7 +575,7 @@ EOF # to extract _just_ the user's note, instead it is prefixed with # some plist junk. This was causing the user note to always be set, # even if there was no reason for it. - if ! poly_user_note_get "$username" | grep -q "Nix build user $coreid"; then + if poly_user_note_get "$username" | grep -q "Nix build user $coreid"; then row " Note" "Nix build user $coreid" else poly_user_note_set "$username" "Nix build user $coreid" @@ -640,7 +683,7 @@ place_channel_configuration() { check_selinux() { if command -v getenforce > /dev/null 2>&1; then - if ! [ "$(getenforce)" = "Disabled" ]; then + if [ "$(getenforce)" = "Enforcing" ]; then failure <<EOF Nix does not work with selinux enabled yet! see https://github.com/NixOS/nix/issues/2374 @@ -777,7 +820,7 @@ EOF fi _sudo "to load data for the first time in to the Nix Database" \ - "$NIX_INSTALLED_NIX/bin/nix-store" --load-db < ./.reginfo + HOME="$ROOT_HOME" "$NIX_INSTALLED_NIX/bin/nix-store" --load-db < ./.reginfo echo " Just finished getting the nix database ready." ) @@ -795,6 +838,19 @@ fi EOF } +# Fish has differing syntax +fish_source_lines() { + cat <<EOF + +# Nix +if test -e '$PROFILE_NIX_FILE_FISH' + . '$PROFILE_NIX_FILE_FISH' +end +# End Nix + +EOF +} + configure_shell_profile() { task "Setting up shell profiles: ${PROFILE_TARGETS[*]}" for profile_target in "${PROFILE_TARGETS[@]}"; do @@ -816,6 +872,27 @@ configure_shell_profile() { tee -a "$profile_target" fi done + + task "Setting up shell profiles for Fish with with ${PROFILE_FISH_SUFFIX} inside ${PROFILE_FISH_PREFIXES[*]}" + for fish_prefix in "${PROFILE_FISH_PREFIXES[@]}"; do + if [ ! -d "$fish_prefix" ]; then + # this specific prefix (ie: /etc/fish) is very likely to exist + # if Fish is installed with this sysconfdir. + continue + fi + + profile_target="${fish_prefix}/${PROFILE_FISH_SUFFIX}" + conf_dir=$(dirname "$profile_target") + if [ ! -d "$conf_dir" ]; then + _sudo "create $conf_dir for our Fish hook" \ + mkdir "$conf_dir" + fi + + fish_source_lines \ + | _sudo "write nix-daemon settings to $profile_target" \ + tee "$profile_target" + done + # TODO: should we suggest '. $PROFILE_NIX_FILE'? It would get them on # their way less disruptively, but a counter-argument is that they won't # immediately notice if something didn't get set up right? @@ -865,24 +942,14 @@ EOF install -m 0664 "$SCRATCH/nix.conf" /etc/nix/nix.conf } -main() { - # TODO: I've moved this out of validate_starting_assumptions so we - # can fail faster in this case. Sourcing install-darwin... now runs - # `touch /` to detect Read-only root, but it could update times on - # pre-Catalina macOS if run as root user. - if [ "$EUID" -eq 0 ]; then - failure <<EOF -Please do not run this script with root privileges. I will call sudo -when I need to. -EOF - fi +main() { check_selinux - if [ "$(uname -s)" = "Darwin" ]; then + if is_os_darwin; then # shellcheck source=./install-darwin-multi-user.sh . "$EXTRACTED_NIX_PATH/install-darwin-multi-user.sh" - elif [ "$(uname -s)" = "Linux" ]; then + elif is_os_linux; then # shellcheck source=./install-systemd-multi-user.sh . "$EXTRACTED_NIX_PATH/install-systemd-multi-user.sh" # most of this works on non-systemd distros also else @@ -890,7 +957,10 @@ EOF fi welcome_to_nix - chat_about_sudo + + if ! is_root; then + chat_about_sudo + fi cure_artifacts # TODO: there's a tension between cure and validate. I moved the diff --git a/scripts/install-nix-from-closure.sh b/scripts/install-nix-from-closure.sh index cd3cf6670..794622530 100644 --- a/scripts/install-nix-from-closure.sh +++ b/scripts/install-nix-from-closure.sh @@ -71,6 +71,8 @@ while [ $# -gt 0 ]; do # # intentional tail space # ACTIONS="${ACTIONS}uninstall " # ;; + --yes) + export NIX_INSTALLER_YES=1;; --no-channel-add) export NIX_INSTALLER_NO_CHANNEL_ADD=1;; --daemon-user-count) @@ -90,7 +92,7 @@ while [ $# -gt 0 ]; do shift;; *) { - echo "Nix Installer [--daemon|--no-daemon] [--daemon-user-count INT] [--no-channel-add] [--no-modify-profile] [--nix-extra-conf-file FILE]" + echo "Nix Installer [--daemon|--no-daemon] [--daemon-user-count INT] [--yes] [--no-channel-add] [--no-modify-profile] [--nix-extra-conf-file FILE]" echo "Choose installation method." echo "" @@ -104,6 +106,8 @@ while [ $# -gt 0 ]; do echo " trivial to uninstall." echo " (default)" echo "" + echo " --yes: Run the script non-interactively, accepting all prompts." + echo "" echo " --no-channel-add: Don't add any channels. nixpkgs-unstable is installed by default." echo "" echo " --no-modify-profile: Don't modify the user profile to automatically load nix." @@ -184,6 +188,8 @@ fi # shellcheck source=./nix-profile.sh.in . "$nix/etc/profile.d/nix.sh" +NIX_LINK="$HOME/.nix-profile" + if ! "$nix/bin/nix-env" -i "$nix"; then echo "$0: unable to install Nix into your default profile" >&2 exit 1 @@ -192,7 +198,7 @@ fi # Install an SSL certificate bundle. if [ -z "$NIX_SSL_CERT_FILE" ] || ! [ -f "$NIX_SSL_CERT_FILE" ]; then "$nix/bin/nix-env" -i "$cacert" - export NIX_SSL_CERT_FILE="$HOME/.nix-profile/etc/ssl/certs/ca-bundle.crt" + export NIX_SSL_CERT_FILE="$NIX_LINK/etc/ssl/certs/ca-bundle.crt" fi # Subscribe the user to the Nixpkgs channel and fetch it. @@ -209,31 +215,50 @@ if [ -z "$NIX_INSTALLER_NO_CHANNEL_ADD" ]; then fi added= -p=$HOME/.nix-profile/etc/profile.d/nix.sh +p= +p_sh=$NIX_LINK/etc/profile.d/nix.sh +p_fish=$NIX_LINK/etc/profile.d/nix.fish if [ -z "$NIX_INSTALLER_NO_MODIFY_PROFILE" ]; then # Make the shell source nix.sh during login. for i in .bash_profile .bash_login .profile; do fn="$HOME/$i" if [ -w "$fn" ]; then - if ! grep -q "$p" "$fn"; then + if ! grep -q "$p_sh" "$fn"; then echo "modifying $fn..." >&2 - printf '\nif [ -e %s ]; then . %s; fi # added by Nix installer\n' "$p" "$p" >> "$fn" + printf '\nif [ -e %s ]; then . %s; fi # added by Nix installer\n' "$p_sh" "$p_sh" >> "$fn" fi added=1 + p=${p_sh} break fi done for i in .zshenv .zshrc; do fn="$HOME/$i" if [ -w "$fn" ]; then - if ! grep -q "$p" "$fn"; then + if ! grep -q "$p_sh" "$fn"; then echo "modifying $fn..." >&2 - printf '\nif [ -e %s ]; then . %s; fi # added by Nix installer\n' "$p" "$p" >> "$fn" + printf '\nif [ -e %s ]; then . %s; fi # added by Nix installer\n' "$p_sh" "$p_sh" >> "$fn" fi added=1 + p=${p_sh} break fi done + + if [ -d "$HOME/.config/fish" ]; then + fishdir=$HOME/.config/fish/conf.d + if [ ! -d "$fishdir" ]; then + mkdir -p "$fishdir" + fi + + fn="$fishdir/nix.fish" + echo "placing $fn..." >&2 + printf '\nif test -e %s; . %s; end # added by Nix installer\n' "$p_fish" "$p_fish" > "$fn" + added=1 + p=${p_fish} + fi +else + p=${p_sh} fi if [ -z "$added" ]; then diff --git a/scripts/install-systemd-multi-user.sh b/scripts/install-systemd-multi-user.sh index 62397127a..7dd567747 100755 --- a/scripts/install-systemd-multi-user.sh +++ b/scripts/install-systemd-multi-user.sh @@ -24,12 +24,17 @@ $1 EOF } +escape_systemd_env() { + temp_var="${1//\'/\\\'}" + echo "${temp_var//\%/%%}" +} + # Gather all non-empty proxy environment variables into a string create_systemd_proxy_env() { vars="http_proxy https_proxy ftp_proxy no_proxy HTTP_PROXY HTTPS_PROXY FTP_PROXY NO_PROXY" for v in $vars; do if [ "x${!v:-}" != "x" ]; then - echo "Environment=${v}=${!v}" + echo "Environment=${v}=$(escape_systemd_env ${!v})" fi done } diff --git a/scripts/install.in b/scripts/install.in index af5f71080..7d2e52b26 100755 --- a/scripts/install.in +++ b/scripts/install.in @@ -40,12 +40,12 @@ case "$(uname -s).$(uname -m)" in path=@tarballPath_aarch64-linux@ system=aarch64-linux ;; - Linux.armv6l_linux) + Linux.armv6l) hash=@tarballHash_armv6l-linux@ path=@tarballPath_armv6l-linux@ system=armv6l-linux ;; - Linux.armv7l_linux) + Linux.armv7l) hash=@tarballHash_armv7l-linux@ path=@tarballPath_armv7l-linux@ system=armv7l-linux diff --git a/scripts/local.mk b/scripts/local.mk index b8477178e..46255e432 100644 --- a/scripts/local.mk +++ b/scripts/local.mk @@ -6,6 +6,8 @@ noinst-scripts += $(nix_noinst_scripts) profiledir = $(sysconfdir)/profile.d $(eval $(call install-file-as, $(d)/nix-profile.sh, $(profiledir)/nix.sh, 0644)) +$(eval $(call install-file-as, $(d)/nix-profile.fish, $(profiledir)/nix.fish, 0644)) $(eval $(call install-file-as, $(d)/nix-profile-daemon.sh, $(profiledir)/nix-daemon.sh, 0644)) +$(eval $(call install-file-as, $(d)/nix-profile-daemon.fish, $(profiledir)/nix-daemon.fish, 0644)) clean-files += $(nix_noinst_scripts) diff --git a/scripts/nix-profile-daemon.fish.in b/scripts/nix-profile-daemon.fish.in new file mode 100644 index 000000000..400696812 --- /dev/null +++ b/scripts/nix-profile-daemon.fish.in @@ -0,0 +1,49 @@ +function add_path --argument-names new_path + if type -q fish_add_path + # fish 3.2.0 or newer + fish_add_path --prepend --global $new_path + else + # older versions of fish + if not contains $new_path $fish_user_paths + set --global fish_user_paths $new_path $fish_user_paths + end + end +end + +# Only execute this file once per shell. +if test -n "$__ETC_PROFILE_NIX_SOURCED" + exit +end + +set __ETC_PROFILE_NIX_SOURCED 1 + +set --export NIX_PROFILES "@localstatedir@/nix/profiles/default $HOME/.nix-profile" + +# Set $NIX_SSL_CERT_FILE so that Nixpkgs applications like curl work. +if test -n "$NIX_SSH_CERT_FILE" + : # Allow users to override the NIX_SSL_CERT_FILE +else if test -e /etc/ssl/certs/ca-certificates.crt # NixOS, Ubuntu, Debian, Gentoo, Arch + set --export NIX_SSL_CERT_FILE /etc/ssl/certs/ca-certificates.crt +else if test -e /etc/ssl/ca-bundle.pem # openSUSE Tumbleweed + set --export NIX_SSL_CERT_FILE /etc/ssl/ca-bundle.pem +else if test -e /etc/ssl/certs/ca-bundle.crt # Old NixOS + set --export NIX_SSL_CERT_FILE /etc/ssl/certs/ca-bundle.crt +else if test -e /etc/pki/tls/certs/ca-bundle.crt # Fedora, CentOS + set --export NIX_SSL_CERT_FILE /etc/pki/tls/certs/ca-bundle.crt +else if test -e "$NIX_LINK/etc/ssl/certs/ca-bundle.crt" # fall back to cacert in Nix profile + set --export NIX_SSL_CERT_FILE "$NIX_LINK/etc/ssl/certs/ca-bundle.crt" +else if test -e "$NIX_LINK/etc/ca-bundle.crt" # old cacert in Nix profile + set --export NIX_SSL_CERT_FILE "$NIX_LINK/etc/ca-bundle.crt" +else + # Fall back to what is in the nix profiles, favouring whatever is defined last. + for i in $NIX_PROFILES + if test -e "$i/etc/ssl/certs/ca-bundle.crt" + set --export NIX_SSL_CERT_FILE "$i/etc/ssl/certs/ca-bundle.crt" + end + end +end + +add_path "@localstatedir@/nix/profiles/default/bin" +add_path "$HOME/.nix-profile/bin" + +functions -e add_path diff --git a/scripts/nix-profile-daemon.sh.in b/scripts/nix-profile-daemon.sh.in index 0a47571ac..8cfd3149e 100644 --- a/scripts/nix-profile-daemon.sh.in +++ b/scripts/nix-profile-daemon.sh.in @@ -2,7 +2,33 @@ if [ -n "${__ETC_PROFILE_NIX_SOURCED:-}" ]; then return; fi __ETC_PROFILE_NIX_SOURCED=1 -export NIX_PROFILES="@localstatedir@/nix/profiles/default $HOME/.nix-profile" +NIX_LINK=$HOME/.nix-profile +if [ -n "${XDG_STATE_HOME-}" ]; then + NIX_LINK_NEW="$XDG_STATE_HOME/nix/profile" +else + NIX_LINK_NEW=$HOME/.local/state/nix/profile +fi +if [ -e "$NIX_LINK_NEW" ]; then + NIX_LINK="$NIX_LINK_NEW" +else + if [ -t 2 ] && [ -e "$NIX_LINK_NEW" ]; then + warning="\033[1;35mwarning:\033[0m" + printf "$warning Both %s and legacy %s exist; using the latter.\n" "$NIX_LINK_NEW" "$NIX_LINK" 1>&2 + if [ "$(realpath "$NIX_LINK")" = "$(realpath "$NIX_LINK_NEW")" ]; then + printf " Since the profiles match, you can safely delete either of them.\n" 1>&2 + else + # This should be an exceptionally rare occasion: the only way to get it would be to + # 1. Update to newer Nix; + # 2. Remove .nix-profile; + # 3. Set the $NIX_LINK_NEW to something other than the default user profile; + # 4. Roll back to older Nix. + # If someone did all that, they can probably figure out how to migrate the profile. + printf "$warning Profiles do not match. You should manually migrate from %s to %s.\n" "$NIX_LINK" "$NIX_LINK_NEW" 1>&2 + fi + fi +fi + +export NIX_PROFILES="@localstatedir@/nix/profiles/default $NIX_LINK" # Set $NIX_SSL_CERT_FILE so that Nixpkgs applications like curl work. if [ -n "${NIX_SSL_CERT_FILE:-}" ]; then @@ -34,4 +60,5 @@ else unset -f check_nix_profiles fi -export PATH="$HOME/.nix-profile/bin:@localstatedir@/nix/profiles/default/bin:$PATH" +export PATH="$NIX_LINK/bin:@localstatedir@/nix/profiles/default/bin:$PATH" +unset NIX_LINK diff --git a/scripts/nix-profile.fish.in b/scripts/nix-profile.fish.in new file mode 100644 index 000000000..731498c76 --- /dev/null +++ b/scripts/nix-profile.fish.in @@ -0,0 +1,51 @@ +function add_path --argument-names new_path + if type -q fish_add_path + # fish 3.2.0 or newer + fish_add_path --prepend --global $new_path + else + # older versions of fish + if not contains $new_path $fish_user_paths + set --global fish_user_paths $new_path $fish_user_paths + end + end +end + +if test -n "$HOME" && test -n "$USER" + + # Set up the per-user profile. + + set NIX_LINK $HOME/.nix-profile + + # Set up environment. + # This part should be kept in sync with nixpkgs:nixos/modules/programs/environment.nix + set --export NIX_PROFILES "@localstatedir@/nix/profiles/default $HOME/.nix-profile" + + # Set $NIX_SSL_CERT_FILE so that Nixpkgs applications like curl work. + if test -n "$NIX_SSH_CERT_FILE" + : # Allow users to override the NIX_SSL_CERT_FILE + else if test -e /etc/ssl/certs/ca-certificates.crt # NixOS, Ubuntu, Debian, Gentoo, Arch + set --export NIX_SSL_CERT_FILE /etc/ssl/certs/ca-certificates.crt + else if test -e /etc/ssl/ca-bundle.pem # openSUSE Tumbleweed + set --export NIX_SSL_CERT_FILE /etc/ssl/ca-bundle.pem + else if test -e /etc/ssl/certs/ca-bundle.crt # Old NixOS + set --export NIX_SSL_CERT_FILE /etc/ssl/certs/ca-bundle.crt + else if test -e /etc/pki/tls/certs/ca-bundle.crt # Fedora, CentOS + set --export NIX_SSL_CERT_FILE /etc/pki/tls/certs/ca-bundle.crt + else if test -e "$NIX_LINK/etc/ssl/certs/ca-bundle.crt" # fall back to cacert in Nix profile + set --export NIX_SSL_CERT_FILE "$NIX_LINK/etc/ssl/certs/ca-bundle.crt" + else if test -e "$NIX_LINK/etc/ca-bundle.crt" # old cacert in Nix profile + set --export NIX_SSL_CERT_FILE "$NIX_LINK/etc/ca-bundle.crt" + end + + # Only use MANPATH if it is already set. In general `man` will just simply + # pick up `.nix-profile/share/man` because is it close to `.nix-profile/bin` + # which is in the $PATH. For more info, run `manpath -d`. + if set --query MANPATH + set --export --prepend --path MANPATH "$NIX_LINK/share/man" + end + + add_path "$NIX_LINK/bin" + set --erase NIX_LINK +end + +functions -e add_path diff --git a/scripts/nix-profile.sh.in b/scripts/nix-profile.sh.in index 45cbcbe74..c4d60cf37 100644 --- a/scripts/nix-profile.sh.in +++ b/scripts/nix-profile.sh.in @@ -1,13 +1,36 @@ if [ -n "$HOME" ] && [ -n "$USER" ]; then # Set up the per-user profile. - # This part should be kept in sync with nixpkgs:nixos/modules/programs/shell.nix - NIX_LINK=$HOME/.nix-profile + NIX_LINK="$HOME/.nix-profile" + if [ -n "${XDG_STATE_HOME-}" ]; then + NIX_LINK_NEW="$XDG_STATE_HOME/nix/profile" + else + NIX_LINK_NEW="$HOME/.local/state/nix/profile" + fi + if [ -e "$NIX_LINK_NEW" ]; then + NIX_LINK="$NIX_LINK_NEW" + else + if [ -t 2 ] && [ -e "$NIX_LINK_NEW" ]; then + warning="\033[1;35mwarning:\033[0m" + printf "$warning Both %s and legacy %s exist; using the latter.\n" "$NIX_LINK_NEW" "$NIX_LINK" 1>&2 + if [ "$(realpath "$NIX_LINK")" = "$(realpath "$NIX_LINK_NEW")" ]; then + printf " Since the profiles match, you can safely delete either of them.\n" 1>&2 + else + # This should be an exceptionally rare occasion: the only way to get it would be to + # 1. Update to newer Nix; + # 2. Remove .nix-profile; + # 3. Set the $NIX_LINK_NEW to something other than the default user profile; + # 4. Roll back to older Nix. + # If someone did all that, they can probably figure out how to migrate the profile. + printf "$warning Profiles do not match. You should manually migrate from %s to %s.\n" "$NIX_LINK" "$NIX_LINK_NEW" 1>&2 + fi + fi + fi # Set up environment. # This part should be kept in sync with nixpkgs:nixos/modules/programs/environment.nix - export NIX_PROFILES="@localstatedir@/nix/profiles/default $HOME/.nix-profile" + export NIX_PROFILES="@localstatedir@/nix/profiles/default $NIX_LINK" # Set $NIX_SSL_CERT_FILE so that Nixpkgs applications like curl work. if [ -e /etc/ssl/certs/ca-certificates.crt ]; then # NixOS, Ubuntu, Debian, Gentoo, Arch @@ -32,5 +55,5 @@ if [ -n "$HOME" ] && [ -n "$USER" ]; then fi export PATH="$NIX_LINK/bin:$PATH" - unset NIX_LINK + unset NIX_LINK NIX_LINK_NEW fi diff --git a/src/build-remote/build-remote.cc b/src/build-remote/build-remote.cc index ff8ba2724..63e3e3fa9 100644 --- a/src/build-remote/build-remote.cc +++ b/src/build-remote/build-remote.cc @@ -72,6 +72,7 @@ static int main_build_remote(int argc, char * * argv) settings.set(name, value); } + auto maxBuildJobs = settings.maxBuildJobs; settings.maxBuildJobs.set("1"); // hack to make tests with local?root= work initPlugins(); @@ -112,10 +113,14 @@ static int main_build_remote(int argc, char * * argv) drvPath = store->parseStorePath(readString(source)); auto requiredFeatures = readStrings<std::set<std::string>>(source); - auto canBuildLocally = amWilling + /* It would be possible to build locally after some builds clear out, + so don't show the warning now: */ + bool couldBuildLocally = maxBuildJobs > 0 && ( neededSystem == settings.thisSystem || settings.extraPlatforms.get().count(neededSystem) > 0) && allSupportedLocally(*store, requiredFeatures); + /* It's possible to build this locally right now: */ + bool canBuildLocally = amWilling && couldBuildLocally; /* Error ignored here, will be caught later */ mkdir(currentLoad.c_str(), 0777); @@ -186,12 +191,12 @@ static int main_build_remote(int argc, char * * argv) // build the hint template. std::string errorText = "Failed to find a machine for remote build!\n" - "derivation: %s\nrequired (system, features): (%s, %s)"; + "derivation: %s\nrequired (system, features): (%s, [%s])"; errorText += "\n%s available machines:"; errorText += "\n(systems, maxjobs, supportedFeatures, mandatoryFeatures)"; for (unsigned int i = 0; i < machines.size(); ++i) - errorText += "\n(%s, %s, %s, %s)"; + errorText += "\n([%s], %s, [%s], [%s])"; // add the template values. std::string drvstr; @@ -214,7 +219,7 @@ static int main_build_remote(int argc, char * * argv) % concatStringsSep<StringSet>(", ", m.supportedFeatures) % concatStringsSep<StringSet>(", ", m.mandatoryFeatures); - printMsg(canBuildLocally ? lvlChatty : lvlWarn, error); + printMsg(couldBuildLocally ? lvlChatty : lvlWarn, error.str()); std::cerr << "# decline\n"; } diff --git a/src/libcmd/command.cc b/src/libcmd/command.cc index 14bb27936..bedf11e2c 100644 --- a/src/libcmd/command.cc +++ b/src/libcmd/command.cc @@ -4,6 +4,7 @@ #include "derivations.hh" #include "nixexpr.hh" #include "profiles.hh" +#include "repl.hh" #include <nlohmann/json.hpp> @@ -88,7 +89,8 @@ EvalCommand::EvalCommand() { addFlag({ .longName = "debugger", - .description = "start an interactive environment if evaluation fails", + .description = "Start an interactive environment if evaluation fails.", + .category = MixEvalArgs::category, .handler = {&startReplOnEvalErrors, true}, }); } @@ -120,12 +122,22 @@ ref<EvalState> EvalCommand::getEvalState() ; if (startReplOnEvalErrors) { - evalState->debugRepl = &runRepl; + evalState->debugRepl = &AbstractNixRepl::runSimple; }; } return ref<EvalState>(evalState); } +MixOperateOnOptions::MixOperateOnOptions() +{ + addFlag({ + .longName = "derivation", + .description = "Operate on the [store derivation](../../glossary.md#gloss-store-derivation) rather than its outputs.", + .category = installablesCategory, + .handler = {&operateOn, OperateOn::Derivation}, + }); +} + BuiltPathsCommand::BuiltPathsCommand(bool recursive) : recursive(recursive) { @@ -153,7 +165,7 @@ BuiltPathsCommand::BuiltPathsCommand(bool recursive) }); } -void BuiltPathsCommand::run(ref<Store> store) +void BuiltPathsCommand::run(ref<Store> store, Installables && installables) { BuiltPaths paths; if (all) { @@ -199,7 +211,7 @@ void StorePathsCommand::run(ref<Store> store, BuiltPaths && paths) run(store, std::move(sorted)); } -void StorePathCommand::run(ref<Store> store, std::vector<StorePath> && storePaths) +void StorePathCommand::run(ref<Store> store, StorePaths && storePaths) { if (storePaths.size() != 1) throw UsageError("this command requires exactly one store path"); @@ -207,25 +219,11 @@ void StorePathCommand::run(ref<Store> store, std::vector<StorePath> && storePath run(store, *storePaths.begin()); } -Strings editorFor(const Path & file, uint32_t line) -{ - auto editor = getEnv("EDITOR").value_or("cat"); - auto args = tokenizeString<Strings>(editor); - if (line > 0 && ( - editor.find("emacs") != std::string::npos || - editor.find("nano") != std::string::npos || - editor.find("vim") != std::string::npos || - editor.find("kak") != std::string::npos)) - args.push_back(fmt("+%d", line)); - args.push_back(file); - return args; -} - MixProfile::MixProfile() { addFlag({ .longName = "profile", - .description = "The profile to update.", + .description = "The profile to operate on.", .labels = {"path"}, .handler = {&profile}, .completer = completePath @@ -248,7 +246,7 @@ void MixProfile::updateProfile(const BuiltPaths & buildables) { if (!profile) return; - std::vector<StorePath> result; + StorePaths result; for (auto & buildable : buildables) { std::visit(overloaded { diff --git a/src/libcmd/command.hh b/src/libcmd/command.hh index 3b4b40981..874ca3249 100644 --- a/src/libcmd/command.hh +++ b/src/libcmd/command.hh @@ -22,13 +22,16 @@ static constexpr Command::Category catSecondary = 100; static constexpr Command::Category catUtility = 101; static constexpr Command::Category catNixInstallation = 102; -static constexpr auto installablesCategory = "Options that change the interpretation of installables"; +static constexpr auto installablesCategory = "Options that change the interpretation of [installables](@docroot@/command-ref/new-cli/nix.md#installables)"; struct NixMultiCommand : virtual MultiCommand, virtual Command { nlohmann::json toJSON() override; }; +// For the overloaded run methods +#pragma GCC diagnostic ignored "-Woverloaded-virtual" + /* A command that requires a Nix store. */ struct StoreCommand : virtual Command { @@ -94,17 +97,13 @@ struct SourceExprCommand : virtual Args, MixFlakeOptions { std::optional<Path> file; std::optional<std::string> expr; - bool readOnlyMode = false; - - // FIXME: move this; not all commands (e.g. 'nix run') use it. - OperateOn operateOn = OperateOn::Output; - SourceExprCommand(bool supportReadOnlyMode = false); + SourceExprCommand(); - std::vector<std::shared_ptr<Installable>> parseInstallables( + Installables parseInstallables( ref<Store> store, std::vector<std::string> ss); - std::shared_ptr<Installable> parseInstallable( + ref<Installable> parseInstallable( ref<Store> store, const std::string & installable); virtual Strings getDefaultFlakeAttrPaths(); @@ -114,34 +113,48 @@ struct SourceExprCommand : virtual Args, MixFlakeOptions void completeInstallable(std::string_view prefix); }; -/* A command that operates on a list of "installables", which can be - store paths, attribute paths, Nix expressions, etc. */ -struct InstallablesCommand : virtual Args, SourceExprCommand +struct MixReadOnlyOption : virtual Args { - std::vector<std::shared_ptr<Installable>> installables; + MixReadOnlyOption(); +}; - InstallablesCommand(); +/* Like InstallablesCommand but the installables are not loaded */ +struct RawInstallablesCommand : virtual Args, SourceExprCommand +{ + RawInstallablesCommand(); - void prepare() override; - Installables load(); + virtual void run(ref<Store> store, std::vector<std::string> && rawInstallables) = 0; - virtual bool useDefaultInstallables() { return true; } + void run(ref<Store> store) override; + + // FIXME make const after CmdRepl's override is fixed up + virtual void applyDefaultInstallables(std::vector<std::string> & rawInstallables); + + bool readFromStdIn = false; std::vector<std::string> getFlakesForCompletion() override; -protected: +private: + + std::vector<std::string> rawInstallables; +}; +/* A command that operates on a list of "installables", which can be + store paths, attribute paths, Nix expressions, etc. */ +struct InstallablesCommand : RawInstallablesCommand +{ + virtual void run(ref<Store> store, Installables && installables) = 0; - std::vector<std::string> _installables; + void run(ref<Store> store, std::vector<std::string> && rawInstallables) override; }; /* A command that operates on exactly one "installable" */ struct InstallableCommand : virtual Args, SourceExprCommand { - std::shared_ptr<Installable> installable; + InstallableCommand(); - InstallableCommand(bool supportReadOnlyMode = false); + virtual void run(ref<Store> store, ref<Installable> installable) = 0; - void prepare() override; + void run(ref<Store> store) override; std::vector<std::string> getFlakesForCompletion() override { @@ -153,8 +166,15 @@ private: std::string _installable{"."}; }; +struct MixOperateOnOptions : virtual Args +{ + OperateOn operateOn = OperateOn::Output; + + MixOperateOnOptions(); +}; + /* A command that operates on zero or more store paths. */ -struct BuiltPathsCommand : public InstallablesCommand +struct BuiltPathsCommand : InstallablesCommand, virtual MixOperateOnOptions { private: @@ -169,22 +189,18 @@ public: BuiltPathsCommand(bool recursive = false); - using StoreCommand::run; - virtual void run(ref<Store> store, BuiltPaths && paths) = 0; - void run(ref<Store> store) override; + void run(ref<Store> store, Installables && installables) override; - bool useDefaultInstallables() override { return !all; } + void applyDefaultInstallables(std::vector<std::string> & rawInstallables) override; }; struct StorePathsCommand : public BuiltPathsCommand { StorePathsCommand(bool recursive = false); - using BuiltPathsCommand::run; - - virtual void run(ref<Store> store, std::vector<StorePath> && storePaths) = 0; + virtual void run(ref<Store> store, StorePaths && storePaths) = 0; void run(ref<Store> store, BuiltPaths && paths) override; }; @@ -192,11 +208,9 @@ struct StorePathsCommand : public BuiltPathsCommand /* A command that operates on exactly one store path. */ struct StorePathCommand : public StorePathsCommand { - using StorePathsCommand::run; - virtual void run(ref<Store> store, const StorePath & storePath) = 0; - void run(ref<Store> store, std::vector<StorePath> && storePaths) override; + void run(ref<Store> store, StorePaths && storePaths) override; }; /* A helper class for registering commands globally. */ @@ -227,10 +241,6 @@ static RegisterCommand registerCommand2(std::vector<std::string> && name) return RegisterCommand(std::move(name), [](){ return make_ref<T>(); }); } -/* Helper function to generate args that invoke $EDITOR on - filename:lineno. */ -Strings editorFor(const Path & file, uint32_t line); - struct MixProfile : virtual StoreCommand { std::optional<Path> profile; @@ -280,8 +290,4 @@ void printClosureDiff( const StorePath & afterPath, std::string_view indent); - -void runRepl( - ref<EvalState> evalState, - const ValMap & extraEnv); } diff --git a/src/libcmd/common-eval-args.cc b/src/libcmd/common-eval-args.cc index 5b6e82388..908127b4d 100644 --- a/src/libcmd/common-eval-args.cc +++ b/src/libcmd/common-eval-args.cc @@ -13,8 +13,6 @@ namespace nix { MixEvalArgs::MixEvalArgs() { - auto category = "Common evaluation options"; - addFlag({ .longName = "arg", .description = "Pass the value *expr* as the argument *name* to Nix functions.", @@ -34,7 +32,77 @@ MixEvalArgs::MixEvalArgs() addFlag({ .longName = "include", .shortName = 'I', - .description = "Add *path* to the list of locations used to look up `<...>` file names.", + .description = R"( + Add *path* to the Nix search path. The Nix search path is + initialized from the colon-separated [`NIX_PATH`](@docroot@/command-ref/env-common.md#env-NIX_PATH) environment + variable, and is used to look up the location of Nix expressions using [paths](@docroot@/language/values.md#type-path) enclosed in angle + brackets (i.e., `<nixpkgs>`). + + For instance, passing + + ``` + -I /home/eelco/Dev + -I /etc/nixos + ``` + + will cause Nix to look for paths relative to `/home/eelco/Dev` and + `/etc/nixos`, in that order. This is equivalent to setting the + `NIX_PATH` environment variable to + + ``` + /home/eelco/Dev:/etc/nixos + ``` + + It is also possible to match paths against a prefix. For example, + passing + + ``` + -I nixpkgs=/home/eelco/Dev/nixpkgs-branch + -I /etc/nixos + ``` + + will cause Nix to search for `<nixpkgs/path>` in + `/home/eelco/Dev/nixpkgs-branch/path` and `/etc/nixos/nixpkgs/path`. + + If a path in the Nix search path starts with `http://` or `https://`, + it is interpreted as the URL of a tarball that will be downloaded and + unpacked to a temporary location. The tarball must consist of a single + top-level directory. For example, passing + + ``` + -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/master.tar.gz + ``` + + tells Nix to download and use the current contents of the `master` + branch in the `nixpkgs` repository. + + The URLs of the tarballs from the official `nixos.org` channels + (see [the manual page for `nix-channel`](../nix-channel.md)) can be + abbreviated as `channel:<channel-name>`. For instance, the + following two flags are equivalent: + + ``` + -I nixpkgs=channel:nixos-21.05 + -I nixpkgs=https://nixos.org/channels/nixos-21.05/nixexprs.tar.xz + ``` + + You can also fetch source trees using [flake URLs](./nix3-flake.md#url-like-syntax) and add them to the + search path. For instance, + + ``` + -I nixpkgs=flake:nixpkgs + ``` + + specifies that the prefix `nixpkgs` shall refer to the source tree + downloaded from the `nixpkgs` entry in the flake registry. Similarly, + + ``` + -I nixpkgs=flake:github:NixOS/nixpkgs/nixos-22.05 + ``` + + makes `<nixpkgs>` refer to a particular branch of the + `NixOS/nixpkgs` repository on GitHub. + )", .category = category, .labels = {"path"}, .handler = {[&](std::string s) { searchPath.push_back(s); }} @@ -91,14 +159,25 @@ Bindings * MixEvalArgs::getAutoArgs(EvalState & state) Path lookupFileArg(EvalState & state, std::string_view s) { - if (isUri(s)) { - return state.store->toRealPath( - fetchers::downloadTarball( - state.store, resolveUri(s), "source", false).first.storePath); - } else if (s.size() > 2 && s.at(0) == '<' && s.at(s.size() - 1) == '>') { + if (EvalSettings::isPseudoUrl(s)) { + auto storePath = fetchers::downloadTarball( + state.store, EvalSettings::resolvePseudoUrl(s), "source", false).first.storePath; + return state.store->toRealPath(storePath); + } + + else if (hasPrefix(s, "flake:")) { + settings.requireExperimentalFeature(Xp::Flakes); + auto flakeRef = parseFlakeRef(std::string(s.substr(6)), {}, true, false); + auto storePath = flakeRef.resolve(state.store).fetchTree(state.store).first.storePath; + return state.store->toRealPath(storePath); + } + + else if (s.size() > 2 && s.at(0) == '<' && s.at(s.size() - 1) == '>') { Path p(s.substr(1, s.size() - 2)); return state.findFile(p); - } else + } + + else return absPath(std::string(s)); } diff --git a/src/libcmd/common-eval-args.hh b/src/libcmd/common-eval-args.hh index 03fa226aa..1ec800613 100644 --- a/src/libcmd/common-eval-args.hh +++ b/src/libcmd/common-eval-args.hh @@ -10,6 +10,8 @@ class Bindings; struct MixEvalArgs : virtual Args { + static constexpr auto category = "Common evaluation options"; + MixEvalArgs(); Bindings * getAutoArgs(EvalState & state); diff --git a/src/libcmd/editor-for.cc b/src/libcmd/editor-for.cc new file mode 100644 index 000000000..f674f32bd --- /dev/null +++ b/src/libcmd/editor-for.cc @@ -0,0 +1,20 @@ +#include "util.hh" +#include "editor-for.hh" + +namespace nix { + +Strings editorFor(const Path & file, uint32_t line) +{ + auto editor = getEnv("EDITOR").value_or("cat"); + auto args = tokenizeString<Strings>(editor); + if (line > 0 && ( + editor.find("emacs") != std::string::npos || + editor.find("nano") != std::string::npos || + editor.find("vim") != std::string::npos || + editor.find("kak") != std::string::npos)) + args.push_back(fmt("+%d", line)); + args.push_back(file); + return args; +} + +} diff --git a/src/libcmd/editor-for.hh b/src/libcmd/editor-for.hh new file mode 100644 index 000000000..8fbd08792 --- /dev/null +++ b/src/libcmd/editor-for.hh @@ -0,0 +1,11 @@ +#pragma once + +#include "types.hh" + +namespace nix { + +/* Helper function to generate args that invoke $EDITOR on + filename:lineno. */ +Strings editorFor(const Path & file, uint32_t line); + +} diff --git a/src/libcmd/installable-attr-path.cc b/src/libcmd/installable-attr-path.cc new file mode 100644 index 000000000..d9377f0d6 --- /dev/null +++ b/src/libcmd/installable-attr-path.cc @@ -0,0 +1,109 @@ +#include "globals.hh" +#include "installable-attr-path.hh" +#include "outputs-spec.hh" +#include "util.hh" +#include "command.hh" +#include "attr-path.hh" +#include "common-eval-args.hh" +#include "derivations.hh" +#include "eval-inline.hh" +#include "eval.hh" +#include "get-drvs.hh" +#include "store-api.hh" +#include "shared.hh" +#include "flake/flake.hh" +#include "eval-cache.hh" +#include "url.hh" +#include "registry.hh" +#include "build-result.hh" + +#include <regex> +#include <queue> + +#include <nlohmann/json.hpp> + +namespace nix { + +InstallableAttrPath::InstallableAttrPath( + ref<EvalState> state, + SourceExprCommand & cmd, + Value * v, + const std::string & attrPath, + ExtendedOutputsSpec extendedOutputsSpec) + : InstallableValue(state) + , cmd(cmd) + , v(allocRootValue(v)) + , attrPath(attrPath) + , extendedOutputsSpec(std::move(extendedOutputsSpec)) +{ } + +std::pair<Value *, PosIdx> InstallableAttrPath::toValue(EvalState & state) +{ + auto [vRes, pos] = findAlongAttrPath(state, attrPath, *cmd.getAutoArgs(state), **v); + state.forceValue(*vRes, pos); + return {vRes, pos}; +} + +DerivedPathsWithInfo InstallableAttrPath::toDerivedPaths() +{ + auto v = toValue(*state).first; + + Bindings & autoArgs = *cmd.getAutoArgs(*state); + + DrvInfos drvInfos; + getDerivations(*state, *v, "", autoArgs, drvInfos, false); + + // Backward compatibility hack: group results by drvPath. This + // helps keep .all output together. + std::map<StorePath, OutputsSpec> byDrvPath; + + for (auto & drvInfo : drvInfos) { + auto drvPath = drvInfo.queryDrvPath(); + if (!drvPath) + throw Error("'%s' is not a derivation", what()); + + auto newOutputs = std::visit(overloaded { + [&](const ExtendedOutputsSpec::Default & d) -> OutputsSpec { + std::set<std::string> outputsToInstall; + for (auto & output : drvInfo.queryOutputs(false, true)) + outputsToInstall.insert(output.first); + return OutputsSpec::Names { std::move(outputsToInstall) }; + }, + [&](const ExtendedOutputsSpec::Explicit & e) -> OutputsSpec { + return e; + }, + }, extendedOutputsSpec.raw()); + + auto [iter, didInsert] = byDrvPath.emplace(*drvPath, newOutputs); + + if (!didInsert) + iter->second = iter->second.union_(newOutputs); + } + + DerivedPathsWithInfo res; + for (auto & [drvPath, outputs] : byDrvPath) + res.push_back({ + .path = DerivedPath::Built { + .drvPath = drvPath, + .outputs = outputs, + }, + }); + + return res; +} + +InstallableAttrPath InstallableAttrPath::parse( + ref<EvalState> state, + SourceExprCommand & cmd, + Value * v, + std::string_view prefix, + ExtendedOutputsSpec extendedOutputsSpec) +{ + return { + state, cmd, v, + prefix == "." ? "" : std::string { prefix }, + extendedOutputsSpec + }; +} + +} diff --git a/src/libcmd/installable-attr-path.hh b/src/libcmd/installable-attr-path.hh new file mode 100644 index 000000000..c06132ec8 --- /dev/null +++ b/src/libcmd/installable-attr-path.hh @@ -0,0 +1,56 @@ +#include "globals.hh" +#include "installable-value.hh" +#include "outputs-spec.hh" +#include "util.hh" +#include "command.hh" +#include "attr-path.hh" +#include "common-eval-args.hh" +#include "derivations.hh" +#include "eval-inline.hh" +#include "eval.hh" +#include "get-drvs.hh" +#include "store-api.hh" +#include "shared.hh" +#include "eval-cache.hh" +#include "url.hh" +#include "registry.hh" +#include "build-result.hh" + +#include <regex> +#include <queue> + +#include <nlohmann/json.hpp> + +namespace nix { + +class InstallableAttrPath : public InstallableValue +{ + SourceExprCommand & cmd; + RootValue v; + std::string attrPath; + ExtendedOutputsSpec extendedOutputsSpec; + + InstallableAttrPath( + ref<EvalState> state, + SourceExprCommand & cmd, + Value * v, + const std::string & attrPath, + ExtendedOutputsSpec extendedOutputsSpec); + + std::string what() const override { return attrPath; }; + + std::pair<Value *, PosIdx> toValue(EvalState & state) override; + + DerivedPathsWithInfo toDerivedPaths() override; + +public: + + static InstallableAttrPath parse( + ref<EvalState> state, + SourceExprCommand & cmd, + Value * v, + std::string_view prefix, + ExtendedOutputsSpec extendedOutputsSpec); +}; + +} diff --git a/src/libcmd/installable-derived-path.cc b/src/libcmd/installable-derived-path.cc new file mode 100644 index 000000000..729dc7d31 --- /dev/null +++ b/src/libcmd/installable-derived-path.cc @@ -0,0 +1,67 @@ +#include "installable-derived-path.hh" +#include "derivations.hh" + +namespace nix { + +std::string InstallableDerivedPath::what() const +{ + return derivedPath.to_string(*store); +} + +DerivedPathsWithInfo InstallableDerivedPath::toDerivedPaths() +{ + return {{.path = derivedPath, .info = {} }}; +} + +std::optional<StorePath> InstallableDerivedPath::getStorePath() +{ + return std::visit(overloaded { + [&](const DerivedPath::Built & bfd) { + return bfd.drvPath; + }, + [&](const DerivedPath::Opaque & bo) { + return bo.path; + }, + }, derivedPath.raw()); +} + +InstallableDerivedPath InstallableDerivedPath::parse( + ref<Store> store, + std::string_view prefix, + ExtendedOutputsSpec extendedOutputsSpec) +{ + auto derivedPath = std::visit(overloaded { + // If the user did not use ^, we treat the output more + // liberally: we accept a symlink chain or an actual + // store path. + [&](const ExtendedOutputsSpec::Default &) -> DerivedPath { + auto storePath = store->followLinksToStorePath(prefix); + // Remove this prior to stabilizing the new CLI. + if (storePath.isDerivation()) { + auto oldDerivedPath = DerivedPath::Built { + .drvPath = storePath, + .outputs = OutputsSpec::All { }, + }; + warn( + "The interpretation of store paths arguments ending in `.drv` recently changed. If this command is now failing try again with '%s'", + oldDerivedPath.to_string(*store)); + }; + return DerivedPath::Opaque { + .path = std::move(storePath), + }; + }, + // If the user did use ^, we just do exactly what is written. + [&](const ExtendedOutputsSpec::Explicit & outputSpec) -> DerivedPath { + return DerivedPath::Built { + .drvPath = store->parseStorePath(prefix), + .outputs = outputSpec, + }; + }, + }, extendedOutputsSpec.raw()); + return InstallableDerivedPath { + store, + std::move(derivedPath), + }; +} + +} diff --git a/src/libcmd/installable-derived-path.hh b/src/libcmd/installable-derived-path.hh new file mode 100644 index 000000000..042878b91 --- /dev/null +++ b/src/libcmd/installable-derived-path.hh @@ -0,0 +1,28 @@ +#pragma once + +#include "installables.hh" + +namespace nix { + +struct InstallableDerivedPath : Installable +{ + ref<Store> store; + DerivedPath derivedPath; + + InstallableDerivedPath(ref<Store> store, DerivedPath && derivedPath) + : store(store), derivedPath(std::move(derivedPath)) + { } + + std::string what() const override; + + DerivedPathsWithInfo toDerivedPaths() override; + + std::optional<StorePath> getStorePath() override; + + static InstallableDerivedPath parse( + ref<Store> store, + std::string_view prefix, + ExtendedOutputsSpec extendedOutputsSpec); +}; + +} diff --git a/src/libcmd/installable-flake.cc b/src/libcmd/installable-flake.cc new file mode 100644 index 000000000..7b0cc376d --- /dev/null +++ b/src/libcmd/installable-flake.cc @@ -0,0 +1,235 @@ +#include "globals.hh" +#include "installable-flake.hh" +#include "installable-derived-path.hh" +#include "outputs-spec.hh" +#include "util.hh" +#include "command.hh" +#include "attr-path.hh" +#include "common-eval-args.hh" +#include "derivations.hh" +#include "eval-inline.hh" +#include "eval.hh" +#include "get-drvs.hh" +#include "store-api.hh" +#include "shared.hh" +#include "flake/flake.hh" +#include "eval-cache.hh" +#include "url.hh" +#include "registry.hh" +#include "build-result.hh" + +#include <regex> +#include <queue> + +#include <nlohmann/json.hpp> + +namespace nix { + +std::vector<std::string> InstallableFlake::getActualAttrPaths() +{ + std::vector<std::string> res; + + for (auto & prefix : prefixes) + res.push_back(prefix + *attrPaths.begin()); + + for (auto & s : attrPaths) + res.push_back(s); + + return res; +} + +Value * InstallableFlake::getFlakeOutputs(EvalState & state, const flake::LockedFlake & lockedFlake) +{ + auto vFlake = state.allocValue(); + + callFlake(state, lockedFlake, *vFlake); + + auto aOutputs = vFlake->attrs->get(state.symbols.create("outputs")); + assert(aOutputs); + + state.forceValue(*aOutputs->value, [&]() { return aOutputs->value->determinePos(noPos); }); + + return aOutputs->value; +} + +static std::string showAttrPaths(const std::vector<std::string> & paths) +{ + std::string s; + for (const auto & [n, i] : enumerate(paths)) { + if (n > 0) s += n + 1 == paths.size() ? " or " : ", "; + s += '\''; s += i; s += '\''; + } + return s; +} + +InstallableFlake::InstallableFlake( + SourceExprCommand * cmd, + ref<EvalState> state, + FlakeRef && flakeRef, + std::string_view fragment, + ExtendedOutputsSpec extendedOutputsSpec, + Strings attrPaths, + Strings prefixes, + const flake::LockFlags & lockFlags) + : InstallableValue(state), + flakeRef(flakeRef), + attrPaths(fragment == "" ? attrPaths : Strings{(std::string) fragment}), + prefixes(fragment == "" ? Strings{} : prefixes), + extendedOutputsSpec(std::move(extendedOutputsSpec)), + lockFlags(lockFlags) +{ + if (cmd && cmd->getAutoArgs(*state)->size()) + throw UsageError("'--arg' and '--argstr' are incompatible with flakes"); +} + +DerivedPathsWithInfo InstallableFlake::toDerivedPaths() +{ + Activity act(*logger, lvlTalkative, actUnknown, fmt("evaluating derivation '%s'", what())); + + auto attr = getCursor(*state); + + auto attrPath = attr->getAttrPathStr(); + + if (!attr->isDerivation()) { + + // FIXME: use eval cache? + auto v = attr->forceValue(); + + if (v.type() == nPath) { + PathSet context; + auto storePath = state->copyPathToStore(context, Path(v.path)); + return {{ + .path = DerivedPath::Opaque { + .path = std::move(storePath), + } + }}; + } + + else if (v.type() == nString) { + PathSet context; + auto s = state->forceString(v, context, noPos, fmt("while evaluating the flake output attribute '%s'", attrPath)); + auto storePath = state->store->maybeParseStorePath(s); + if (storePath && context.count(std::string(s))) { + return {{ + .path = DerivedPath::Opaque { + .path = std::move(*storePath), + } + }}; + } else + throw Error("flake output attribute '%s' evaluates to the string '%s' which is not a store path", attrPath, s); + } + + else + throw Error("flake output attribute '%s' is not a derivation or path", attrPath); + } + + auto drvPath = attr->forceDerivation(); + + std::optional<NixInt> priority; + + if (attr->maybeGetAttr(state->sOutputSpecified)) { + } else if (auto aMeta = attr->maybeGetAttr(state->sMeta)) { + if (auto aPriority = aMeta->maybeGetAttr("priority")) + priority = aPriority->getInt(); + } + + return {{ + .path = DerivedPath::Built { + .drvPath = std::move(drvPath), + .outputs = std::visit(overloaded { + [&](const ExtendedOutputsSpec::Default & d) -> OutputsSpec { + std::set<std::string> outputsToInstall; + if (auto aOutputSpecified = attr->maybeGetAttr(state->sOutputSpecified)) { + if (aOutputSpecified->getBool()) { + if (auto aOutputName = attr->maybeGetAttr("outputName")) + outputsToInstall = { aOutputName->getString() }; + } + } else if (auto aMeta = attr->maybeGetAttr(state->sMeta)) { + if (auto aOutputsToInstall = aMeta->maybeGetAttr("outputsToInstall")) + for (auto & s : aOutputsToInstall->getListOfStrings()) + outputsToInstall.insert(s); + } + + if (outputsToInstall.empty()) + outputsToInstall.insert("out"); + + return OutputsSpec::Names { std::move(outputsToInstall) }; + }, + [&](const ExtendedOutputsSpec::Explicit & e) -> OutputsSpec { + return e; + }, + }, extendedOutputsSpec.raw()), + }, + .info = { + .priority = priority, + .originalRef = flakeRef, + .resolvedRef = getLockedFlake()->flake.lockedRef, + .attrPath = attrPath, + .extendedOutputsSpec = extendedOutputsSpec, + } + }}; +} + +std::pair<Value *, PosIdx> InstallableFlake::toValue(EvalState & state) +{ + return {&getCursor(state)->forceValue(), noPos}; +} + +std::vector<ref<eval_cache::AttrCursor>> +InstallableFlake::getCursors(EvalState & state) +{ + auto evalCache = openEvalCache(state, getLockedFlake()); + + auto root = evalCache->getRoot(); + + std::vector<ref<eval_cache::AttrCursor>> res; + + Suggestions suggestions; + auto attrPaths = getActualAttrPaths(); + + for (auto & attrPath : attrPaths) { + debug("trying flake output attribute '%s'", attrPath); + + auto attr = root->findAlongAttrPath(parseAttrPath(state, attrPath)); + if (attr) { + res.push_back(ref(*attr)); + } else { + suggestions += attr.getSuggestions(); + } + } + + if (res.size() == 0) + throw Error( + suggestions, + "flake '%s' does not provide attribute %s", + flakeRef, + showAttrPaths(attrPaths)); + + return res; +} + +std::shared_ptr<flake::LockedFlake> InstallableFlake::getLockedFlake() const +{ + if (!_lockedFlake) { + flake::LockFlags lockFlagsApplyConfig = lockFlags; + lockFlagsApplyConfig.applyNixConfig = true; + _lockedFlake = std::make_shared<flake::LockedFlake>(lockFlake(*state, flakeRef, lockFlagsApplyConfig)); + } + return _lockedFlake; +} + +FlakeRef InstallableFlake::nixpkgsFlakeRef() const +{ + auto lockedFlake = getLockedFlake(); + + if (auto nixpkgsInput = lockedFlake->lockFile.findInput({"nixpkgs"})) { + if (auto lockedNode = std::dynamic_pointer_cast<const flake::LockedNode>(nixpkgsInput)) { + debug("using nixpkgs flake '%s'", lockedNode->lockedRef); + return std::move(lockedNode->lockedRef); + } + } + + return Installable::nixpkgsFlakeRef(); +} + +} diff --git a/src/libcmd/installable-flake.hh b/src/libcmd/installable-flake.hh new file mode 100644 index 000000000..c75765086 --- /dev/null +++ b/src/libcmd/installable-flake.hh @@ -0,0 +1,50 @@ +#pragma once + +#include "installable-value.hh" + +namespace nix { + +struct InstallableFlake : InstallableValue +{ + FlakeRef flakeRef; + Strings attrPaths; + Strings prefixes; + ExtendedOutputsSpec extendedOutputsSpec; + const flake::LockFlags & lockFlags; + mutable std::shared_ptr<flake::LockedFlake> _lockedFlake; + + InstallableFlake( + SourceExprCommand * cmd, + ref<EvalState> state, + FlakeRef && flakeRef, + std::string_view fragment, + ExtendedOutputsSpec extendedOutputsSpec, + Strings attrPaths, + Strings prefixes, + const flake::LockFlags & lockFlags); + + std::string what() const override { return flakeRef.to_string() + "#" + *attrPaths.begin(); } + + std::vector<std::string> getActualAttrPaths(); + + Value * getFlakeOutputs(EvalState & state, const flake::LockedFlake & lockedFlake); + + DerivedPathsWithInfo toDerivedPaths() override; + + std::pair<Value *, PosIdx> toValue(EvalState & state) override; + + /* Get a cursor to every attrpath in getActualAttrPaths() + that exists. However if none exists, throw an exception. */ + std::vector<ref<eval_cache::AttrCursor>> + getCursors(EvalState & state) override; + + std::shared_ptr<flake::LockedFlake> getLockedFlake() const; + + FlakeRef nixpkgsFlakeRef() const override; +}; + +ref<eval_cache::EvalCache> openEvalCache( + EvalState & state, + std::shared_ptr<flake::LockedFlake> lockedFlake); + +} diff --git a/src/libcmd/installable-value.hh b/src/libcmd/installable-value.hh new file mode 100644 index 000000000..c6cdc4797 --- /dev/null +++ b/src/libcmd/installable-value.hh @@ -0,0 +1,14 @@ +#pragma once + +#include "installables.hh" + +namespace nix { + +struct InstallableValue : Installable +{ + ref<EvalState> state; + + InstallableValue(ref<EvalState> state) : state(state) {} +}; + +} diff --git a/src/libcmd/installables.cc b/src/libcmd/installables.cc index 59162c4df..5cbf26b88 100644 --- a/src/libcmd/installables.cc +++ b/src/libcmd/installables.cc @@ -1,5 +1,9 @@ #include "globals.hh" #include "installables.hh" +#include "installable-derived-path.hh" +#include "installable-attr-path.hh" +#include "installable-flake.hh" +#include "outputs-spec.hh" #include "util.hh" #include "command.hh" #include "attr-path.hh" @@ -143,13 +147,13 @@ void MixFlakeOptions::completionHook() completeFlakeInput(*prefix); } -SourceExprCommand::SourceExprCommand(bool supportReadOnlyMode) +SourceExprCommand::SourceExprCommand() { addFlag({ .longName = "file", .shortName = 'f', .description = - "Interpret installables as attribute paths relative to the Nix expression stored in *file*. " + "Interpret [*installables*](@docroot@/command-ref/new-cli/nix.md#installables) as attribute paths relative to the Nix expression stored in *file*. " "If *file* is the character -, then a Nix expression will be read from standard input. " "Implies `--impure`.", .category = installablesCategory, @@ -160,29 +164,23 @@ SourceExprCommand::SourceExprCommand(bool supportReadOnlyMode) addFlag({ .longName = "expr", - .description = "Interpret installables as attribute paths relative to the Nix expression *expr*.", + .description = "Interpret [*installables*](@docroot@/command-ref/new-cli/nix.md#installables) as attribute paths relative to the Nix expression *expr*.", .category = installablesCategory, .labels = {"expr"}, .handler = {&expr} }); +} +MixReadOnlyOption::MixReadOnlyOption() +{ addFlag({ - .longName = "derivation", - .description = "Operate on the store derivation rather than its outputs.", - .category = installablesCategory, - .handler = {&operateOn, OperateOn::Derivation}, + .longName = "read-only", + .description = + "Do not instantiate each evaluated derivation. " + "This improves performance, but can cause errors when accessing " + "store paths of derivations during evaluation.", + .handler = {&settings.readOnlyMode, true}, }); - - if (supportReadOnlyMode) { - addFlag({ - .longName = "read-only", - .description = - "Do not instantiate each evaluated derivation. " - "This improves performance, but can cause errors when accessing " - "store paths of derivations during evaluation.", - .handler = {&readOnlyMode, true}, - }); - } } Strings SourceExprCommand::getDefaultFlakeAttrPaths() @@ -207,55 +205,59 @@ Strings SourceExprCommand::getDefaultFlakeAttrPathPrefixes() void SourceExprCommand::completeInstallable(std::string_view prefix) { - if (file) { - completionType = ctAttrs; + try { + if (file) { + completionType = ctAttrs; - evalSettings.pureEval = false; - auto state = getEvalState(); - Expr *e = state->parseExprFromFile( - resolveExprPath(state->checkSourcePath(lookupFileArg(*state, *file))) - ); + evalSettings.pureEval = false; + auto state = getEvalState(); + Expr *e = state->parseExprFromFile( + resolveExprPath(state->checkSourcePath(lookupFileArg(*state, *file))) + ); - Value root; - state->eval(e, root); + Value root; + state->eval(e, root); - auto autoArgs = getAutoArgs(*state); + auto autoArgs = getAutoArgs(*state); - std::string prefix_ = std::string(prefix); - auto sep = prefix_.rfind('.'); - std::string searchWord; - if (sep != std::string::npos) { - searchWord = prefix_.substr(sep + 1, std::string::npos); - prefix_ = prefix_.substr(0, sep); - } else { - searchWord = prefix_; - prefix_ = ""; - } + std::string prefix_ = std::string(prefix); + auto sep = prefix_.rfind('.'); + std::string searchWord; + if (sep != std::string::npos) { + searchWord = prefix_.substr(sep + 1, std::string::npos); + prefix_ = prefix_.substr(0, sep); + } else { + searchWord = prefix_; + prefix_ = ""; + } - auto [v, pos] = findAlongAttrPath(*state, prefix_, *autoArgs, root); - Value &v1(*v); - state->forceValue(v1, pos); - Value v2; - state->autoCallFunction(*autoArgs, v1, v2); - - if (v2.type() == nAttrs) { - for (auto & i : *v2.attrs) { - std::string name = state->symbols[i.name]; - if (name.find(searchWord) == 0) { - if (prefix_ == "") - completions->add(name); - else - completions->add(prefix_ + "." + name); + auto [v, pos] = findAlongAttrPath(*state, prefix_, *autoArgs, root); + Value &v1(*v); + state->forceValue(v1, pos); + Value v2; + state->autoCallFunction(*autoArgs, v1, v2); + + if (v2.type() == nAttrs) { + for (auto & i : *v2.attrs) { + std::string name = state->symbols[i.name]; + if (name.find(searchWord) == 0) { + if (prefix_ == "") + completions->add(name); + else + completions->add(prefix_ + "." + name); + } } } + } else { + completeFlakeRefWithFragment( + getEvalState(), + lockFlags, + getDefaultFlakeAttrPathPrefixes(), + getDefaultFlakeAttrPaths(), + prefix); } - } else { - completeFlakeRefWithFragment( - getEvalState(), - lockFlags, - getDefaultFlakeAttrPathPrefixes(), - getDefaultFlakeAttrPaths(), - prefix); + } catch (EvalError&) { + // Don't want eval errors to mess-up with the completion engine, so let's just swallow them } } @@ -354,7 +356,7 @@ void completeFlakeRef(ref<Store> store, std::string_view prefix) } } -DerivedPath Installable::toDerivedPath() +DerivedPathWithInfo Installable::toDerivedPath() { auto buildables = toDerivedPaths(); if (buildables.size() != 1) @@ -374,10 +376,9 @@ Installable::getCursors(EvalState & state) ref<eval_cache::AttrCursor> Installable::getCursor(EvalState & state) { - auto cursors = getCursors(state); - if (cursors.empty()) - throw Error("cannot find flake attribute '%s'", what()); - return cursors[0]; + /* Although getCursors should return at least one element, in case it doesn't, + bound check to avoid an undefined behavior for vector[0] */ + return getCursors(state).at(0); } static StorePath getDeriver( @@ -392,169 +393,6 @@ static StorePath getDeriver( return *derivers.begin(); } -struct InstallableStorePath : Installable -{ - ref<Store> store; - StorePath storePath; - - InstallableStorePath(ref<Store> store, StorePath && storePath) - : store(store), storePath(std::move(storePath)) { } - - std::string what() const override { return store->printStorePath(storePath); } - - DerivedPaths toDerivedPaths() override - { - if (storePath.isDerivation()) { - auto drv = store->readDerivation(storePath); - return { - DerivedPath::Built { - .drvPath = storePath, - .outputs = drv.outputNames(), - } - }; - } else { - return { - DerivedPath::Opaque { - .path = storePath, - } - }; - } - } - - StorePathSet toDrvPaths(ref<Store> store) override - { - if (storePath.isDerivation()) { - return {storePath}; - } else { - return {getDeriver(store, *this, storePath)}; - } - } - - std::optional<StorePath> getStorePath() override - { - return storePath; - } -}; - -DerivedPaths InstallableValue::toDerivedPaths() -{ - DerivedPaths res; - - std::map<StorePath, std::set<std::string>> drvsToOutputs; - RealisedPath::Set drvsToCopy; - - // Group by derivation, helps with .all in particular - for (auto & drv : toDerivations()) { - for (auto & outputName : drv.outputsToInstall) - drvsToOutputs[drv.drvPath].insert(outputName); - drvsToCopy.insert(drv.drvPath); - } - - for (auto & i : drvsToOutputs) - res.push_back(DerivedPath::Built { i.first, i.second }); - - return res; -} - -StorePathSet InstallableValue::toDrvPaths(ref<Store> store) -{ - StorePathSet res; - for (auto & drv : toDerivations()) - res.insert(drv.drvPath); - return res; -} - -struct InstallableAttrPath : InstallableValue -{ - SourceExprCommand & cmd; - RootValue v; - std::string attrPath; - OutputsSpec outputsSpec; - - InstallableAttrPath( - ref<EvalState> state, - SourceExprCommand & cmd, - Value * v, - const std::string & attrPath, - OutputsSpec outputsSpec) - : InstallableValue(state) - , cmd(cmd) - , v(allocRootValue(v)) - , attrPath(attrPath) - , outputsSpec(std::move(outputsSpec)) - { } - - std::string what() const override { return attrPath; } - - std::pair<Value *, PosIdx> toValue(EvalState & state) override - { - auto [vRes, pos] = findAlongAttrPath(state, attrPath, *cmd.getAutoArgs(state), **v); - state.forceValue(*vRes, pos); - return {vRes, pos}; - } - - virtual std::vector<InstallableValue::DerivationInfo> toDerivations() override; -}; - -std::vector<InstallableValue::DerivationInfo> InstallableAttrPath::toDerivations() -{ - auto v = toValue(*state).first; - - Bindings & autoArgs = *cmd.getAutoArgs(*state); - - DrvInfos drvInfos; - getDerivations(*state, *v, "", autoArgs, drvInfos, false); - - std::vector<DerivationInfo> res; - for (auto & drvInfo : drvInfos) { - auto drvPath = drvInfo.queryDrvPath(); - if (!drvPath) - throw Error("'%s' is not a derivation", what()); - - std::set<std::string> outputsToInstall; - - if (auto outputNames = std::get_if<OutputNames>(&outputsSpec)) - outputsToInstall = *outputNames; - else - for (auto & output : drvInfo.queryOutputs(false, std::get_if<DefaultOutputs>(&outputsSpec))) - outputsToInstall.insert(output.first); - - res.push_back(DerivationInfo { - .drvPath = *drvPath, - .outputsToInstall = std::move(outputsToInstall) - }); - } - - return res; -} - -std::vector<std::string> InstallableFlake::getActualAttrPaths() -{ - std::vector<std::string> res; - - for (auto & prefix : prefixes) - res.push_back(prefix + *attrPaths.begin()); - - for (auto & s : attrPaths) - res.push_back(s); - - return res; -} - -Value * InstallableFlake::getFlakeOutputs(EvalState & state, const flake::LockedFlake & lockedFlake) -{ - auto vFlake = state.allocValue(); - - callFlake(state, lockedFlake, *vFlake); - - auto aOutputs = vFlake->attrs->get(state.symbols.create("outputs")); - assert(aOutputs); - - state.forceValue(*aOutputs->value, [&]() { return aOutputs->value->determinePos(noPos); }); - - return aOutputs->value; -} - ref<eval_cache::EvalCache> openEvalCache( EvalState & state, std::shared_ptr<flake::LockedFlake> lockedFlake) @@ -575,7 +413,7 @@ ref<eval_cache::EvalCache> openEvalCache( auto vFlake = state.allocValue(); flake::callFlake(state, *lockedFlake, *vFlake); - state.forceAttrs(*vFlake, noPos); + state.forceAttrs(*vFlake, noPos, "while parsing cached flake data"); auto aOutputs = vFlake->attrs->get(state.symbols.create("outputs")); assert(aOutputs); @@ -584,183 +422,10 @@ ref<eval_cache::EvalCache> openEvalCache( }); } -static std::string showAttrPaths(const std::vector<std::string> & paths) -{ - std::string s; - for (const auto & [n, i] : enumerate(paths)) { - if (n > 0) s += n + 1 == paths.size() ? " or " : ", "; - s += '\''; s += i; s += '\''; - } - return s; -} - -InstallableFlake::InstallableFlake( - SourceExprCommand * cmd, - ref<EvalState> state, - FlakeRef && flakeRef, - std::string_view fragment, - OutputsSpec outputsSpec, - Strings attrPaths, - Strings prefixes, - const flake::LockFlags & lockFlags) - : InstallableValue(state), - flakeRef(flakeRef), - attrPaths(fragment == "" ? attrPaths : Strings{(std::string) fragment}), - prefixes(fragment == "" ? Strings{} : prefixes), - outputsSpec(std::move(outputsSpec)), - lockFlags(lockFlags) -{ - if (cmd && cmd->getAutoArgs(*state)->size()) - throw UsageError("'--arg' and '--argstr' are incompatible with flakes"); -} - -std::tuple<std::string, FlakeRef, InstallableValue::DerivationInfo> InstallableFlake::toDerivation() -{ - auto attr = getCursor(*state); - - auto attrPath = attr->getAttrPathStr(); - - if (!attr->isDerivation()) - throw Error("flake output attribute '%s' is not a derivation", attrPath); - - auto drvPath = attr->forceDerivation(); - - std::set<std::string> outputsToInstall; - std::optional<NixInt> priority; - - if (auto aOutputSpecified = attr->maybeGetAttr(state->sOutputSpecified)) { - if (aOutputSpecified->getBool()) { - if (auto aOutputName = attr->maybeGetAttr("outputName")) - outputsToInstall = { aOutputName->getString() }; - } - } - - else if (auto aMeta = attr->maybeGetAttr(state->sMeta)) { - if (auto aOutputsToInstall = aMeta->maybeGetAttr("outputsToInstall")) - for (auto & s : aOutputsToInstall->getListOfStrings()) - outputsToInstall.insert(s); - if (auto aPriority = aMeta->maybeGetAttr("priority")) - priority = aPriority->getInt(); - } - - if (outputsToInstall.empty() || std::get_if<AllOutputs>(&outputsSpec)) { - outputsToInstall.clear(); - if (auto aOutputs = attr->maybeGetAttr(state->sOutputs)) - for (auto & s : aOutputs->getListOfStrings()) - outputsToInstall.insert(s); - } - - if (outputsToInstall.empty()) - outputsToInstall.insert("out"); - - if (auto outputNames = std::get_if<OutputNames>(&outputsSpec)) - outputsToInstall = *outputNames; - - auto drvInfo = DerivationInfo { - .drvPath = std::move(drvPath), - .outputsToInstall = std::move(outputsToInstall), - .priority = priority, - }; - - return {attrPath, getLockedFlake()->flake.lockedRef, std::move(drvInfo)}; -} - -std::vector<InstallableValue::DerivationInfo> InstallableFlake::toDerivations() -{ - std::vector<DerivationInfo> res; - res.push_back(std::get<2>(toDerivation())); - return res; -} - -std::pair<Value *, PosIdx> InstallableFlake::toValue(EvalState & state) -{ - return {&getCursor(state)->forceValue(), noPos}; -} - -std::vector<ref<eval_cache::AttrCursor>> -InstallableFlake::getCursors(EvalState & state) -{ - auto evalCache = openEvalCache(state, - std::make_shared<flake::LockedFlake>(lockFlake(state, flakeRef, lockFlags))); - - auto root = evalCache->getRoot(); - - std::vector<ref<eval_cache::AttrCursor>> res; - - for (auto & attrPath : getActualAttrPaths()) { - auto attr = root->findAlongAttrPath(parseAttrPath(state, attrPath)); - if (attr) res.push_back(ref(*attr)); - } - - return res; -} - -ref<eval_cache::AttrCursor> InstallableFlake::getCursor(EvalState & state) -{ - auto lockedFlake = getLockedFlake(); - - auto cache = openEvalCache(state, lockedFlake); - auto root = cache->getRoot(); - - Suggestions suggestions; - - auto attrPaths = getActualAttrPaths(); - - for (auto & attrPath : attrPaths) { - debug("trying flake output attribute '%s'", attrPath); - - auto attrOrSuggestions = root->findAlongAttrPath( - parseAttrPath(state, attrPath), - true - ); - - if (!attrOrSuggestions) { - suggestions += attrOrSuggestions.getSuggestions(); - continue; - } - - return *attrOrSuggestions; - } - - throw Error( - suggestions, - "flake '%s' does not provide attribute %s", - flakeRef, - showAttrPaths(attrPaths)); -} - -std::shared_ptr<flake::LockedFlake> InstallableFlake::getLockedFlake() const -{ - if (!_lockedFlake) { - flake::LockFlags lockFlagsApplyConfig = lockFlags; - lockFlagsApplyConfig.applyNixConfig = true; - _lockedFlake = std::make_shared<flake::LockedFlake>(lockFlake(*state, flakeRef, lockFlagsApplyConfig)); - } - return _lockedFlake; -} - -FlakeRef InstallableFlake::nixpkgsFlakeRef() const -{ - auto lockedFlake = getLockedFlake(); - - if (auto nixpkgsInput = lockedFlake->lockFile.findInput({"nixpkgs"})) { - if (auto lockedNode = std::dynamic_pointer_cast<const flake::LockedNode>(nixpkgsInput)) { - debug("using nixpkgs flake '%s'", lockedNode->lockedRef); - return std::move(lockedNode->lockedRef); - } - } - - return Installable::nixpkgsFlakeRef(); -} - -std::vector<std::shared_ptr<Installable>> SourceExprCommand::parseInstallables( +Installables SourceExprCommand::parseInstallables( ref<Store> store, std::vector<std::string> ss) { - std::vector<std::shared_ptr<Installable>> result; - - if (readOnlyMode) { - settings.readOnlyMode = true; - } + Installables result; if (file || expr) { if (file && expr) @@ -775,7 +440,8 @@ std::vector<std::shared_ptr<Installable>> SourceExprCommand::parseInstallables( if (file == "-") { auto e = state->parseStdin(); state->eval(e, *vFile); - } else if (file) + } + else if (file) state->evalFile(lookupFileArg(*state, *file), *vFile); else { auto e = state->parseExprFromString(*expr, absPath(".")); @@ -783,12 +449,11 @@ std::vector<std::shared_ptr<Installable>> SourceExprCommand::parseInstallables( } for (auto & s : ss) { - auto [prefix, outputsSpec] = parseOutputsSpec(s); + auto [prefix, extendedOutputsSpec] = ExtendedOutputsSpec::parse(s); result.push_back( - std::make_shared<InstallableAttrPath>( - state, *this, vFile, - prefix == "." ? "" : prefix, - outputsSpec)); + make_ref<InstallableAttrPath>( + InstallableAttrPath::parse( + state, *this, vFile, prefix, extendedOutputsSpec))); } } else { @@ -796,9 +461,15 @@ std::vector<std::shared_ptr<Installable>> SourceExprCommand::parseInstallables( for (auto & s : ss) { std::exception_ptr ex; - if (s.find('/') != std::string::npos) { + auto [prefix_, extendedOutputsSpec_] = ExtendedOutputsSpec::parse(s); + // To avoid clang's pedantry + auto prefix = std::move(prefix_); + auto extendedOutputsSpec = std::move(extendedOutputsSpec_); + + if (prefix.find('/') != std::string::npos) { try { - result.push_back(std::make_shared<InstallableStorePath>(store, store->followLinksToStorePath(s))); + result.push_back(make_ref<InstallableDerivedPath>( + InstallableDerivedPath::parse(store, prefix, extendedOutputsSpec))); continue; } catch (BadStorePath &) { } catch (...) { @@ -808,13 +479,13 @@ std::vector<std::shared_ptr<Installable>> SourceExprCommand::parseInstallables( } try { - auto [flakeRef, fragment, outputsSpec] = parseFlakeRefWithFragmentAndOutputsSpec(s, absPath(".")); - result.push_back(std::make_shared<InstallableFlake>( + auto [flakeRef, fragment] = parseFlakeRefWithFragment(std::string { prefix }, absPath(".")); + result.push_back(make_ref<InstallableFlake>( this, getEvalState(), std::move(flakeRef), fragment, - outputsSpec, + extendedOutputsSpec, getDefaultFlakeAttrPaths(), getDefaultFlakeAttrPathPrefixes(), lockFlags)); @@ -830,7 +501,7 @@ std::vector<std::shared_ptr<Installable>> SourceExprCommand::parseInstallables( return result; } -std::shared_ptr<Installable> SourceExprCommand::parseInstallable( +ref<Installable> SourceExprCommand::parseInstallable( ref<Store> store, const std::string & installable) { auto installables = parseInstallables(store, {installable}); @@ -838,40 +509,46 @@ std::shared_ptr<Installable> SourceExprCommand::parseInstallable( return installables.front(); } -BuiltPaths Installable::build( +std::vector<BuiltPathWithResult> Installable::build( ref<Store> evalStore, ref<Store> store, Realise mode, - const std::vector<std::shared_ptr<Installable>> & installables, + const Installables & installables, BuildMode bMode) { - BuiltPaths res; - for (auto & [_, builtPath] : build2(evalStore, store, mode, installables, bMode)) - res.push_back(builtPath); + std::vector<BuiltPathWithResult> res; + for (auto & [_, builtPathWithResult] : build2(evalStore, store, mode, installables, bMode)) + res.push_back(builtPathWithResult); return res; } -std::vector<std::pair<std::shared_ptr<Installable>, BuiltPath>> Installable::build2( +std::vector<std::pair<ref<Installable>, BuiltPathWithResult>> Installable::build2( ref<Store> evalStore, ref<Store> store, Realise mode, - const std::vector<std::shared_ptr<Installable>> & installables, + const Installables & installables, BuildMode bMode) { if (mode == Realise::Nothing) settings.readOnlyMode = true; + struct Aux + { + ExtraPathInfo info; + ref<Installable> installable; + }; + std::vector<DerivedPath> pathsToBuild; - std::map<DerivedPath, std::vector<std::shared_ptr<Installable>>> backmap; + std::map<DerivedPath, std::vector<Aux>> backmap; for (auto & i : installables) { for (auto b : i->toDerivedPaths()) { - pathsToBuild.push_back(b); - backmap[b].push_back(i); + pathsToBuild.push_back(b.path); + backmap[b.path].push_back({.info = b.info, .installable = i}); } } - std::vector<std::pair<std::shared_ptr<Installable>, BuiltPath>> res; + std::vector<std::pair<ref<Installable>, BuiltPathWithResult>> res; switch (mode) { @@ -880,42 +557,18 @@ std::vector<std::pair<std::shared_ptr<Installable>, BuiltPath>> Installable::bui printMissing(store, pathsToBuild, lvlError); for (auto & path : pathsToBuild) { - for (auto & installable : backmap[path]) { + for (auto & aux : backmap[path]) { std::visit(overloaded { [&](const DerivedPath::Built & bfd) { - OutputPathMap outputs; - auto drv = evalStore->readDerivation(bfd.drvPath); - auto outputHashes = staticOutputHashes(*evalStore, drv); // FIXME: expensive - auto drvOutputs = drv.outputsAndOptPaths(*store); - for (auto & output : bfd.outputs) { - auto outputHash = get(outputHashes, output); - if (!outputHash) - throw Error( - "the derivation '%s' doesn't have an output named '%s'", - store->printStorePath(bfd.drvPath), output); - if (settings.isExperimentalFeatureEnabled(Xp::CaDerivations)) { - DrvOutput outputId { *outputHash, output }; - auto realisation = store->queryRealisation(outputId); - if (!realisation) - throw Error( - "cannot operate on an output of the " - "unbuilt derivation '%s'", - outputId.to_string()); - outputs.insert_or_assign(output, realisation->outPath); - } else { - // If ca-derivations isn't enabled, assume that - // the output path is statically known. - auto drvOutput = get(drvOutputs, output); - assert(drvOutput); - assert(drvOutput->second); - outputs.insert_or_assign( - output, *drvOutput->second); - } - } - res.push_back({installable, BuiltPath::Built { bfd.drvPath, outputs }}); + auto outputs = resolveDerivedPath(*store, bfd, &*evalStore); + res.push_back({aux.installable, { + .path = BuiltPath::Built { bfd.drvPath, outputs }, + .info = aux.info}}); }, [&](const DerivedPath::Opaque & bo) { - res.push_back({installable, BuiltPath::Opaque { bo.path }}); + res.push_back({aux.installable, { + .path = BuiltPath::Opaque { bo.path }, + .info = aux.info}}); }, }, path.raw()); } @@ -925,22 +578,28 @@ std::vector<std::pair<std::shared_ptr<Installable>, BuiltPath>> Installable::bui case Realise::Outputs: { if (settings.printMissing) - printMissing(store, pathsToBuild, lvlInfo); + printMissing(store, pathsToBuild, lvlInfo); for (auto & buildResult : store->buildPathsWithResults(pathsToBuild, bMode, evalStore)) { if (!buildResult.success()) buildResult.rethrow(); - for (auto & installable : backmap[buildResult.path]) { + for (auto & aux : backmap[buildResult.path]) { std::visit(overloaded { [&](const DerivedPath::Built & bfd) { std::map<std::string, StorePath> outputs; for (auto & path : buildResult.builtOutputs) outputs.emplace(path.first.outputName, path.second.outPath); - res.push_back({installable, BuiltPath::Built { bfd.drvPath, outputs }}); + res.push_back({aux.installable, { + .path = BuiltPath::Built { bfd.drvPath, outputs }, + .info = aux.info, + .result = buildResult}}); }, [&](const DerivedPath::Opaque & bo) { - res.push_back({installable, BuiltPath::Opaque { bo.path }}); + res.push_back({aux.installable, { + .path = BuiltPath::Opaque { bo.path }, + .info = aux.info, + .result = buildResult}}); }, }, buildResult.path.raw()); } @@ -961,11 +620,14 @@ BuiltPaths Installable::toBuiltPaths( ref<Store> store, Realise mode, OperateOn operateOn, - const std::vector<std::shared_ptr<Installable>> & installables) + const Installables & installables) { - if (operateOn == OperateOn::Output) - return Installable::build(evalStore, store, mode, installables); - else { + if (operateOn == OperateOn::Output) { + BuiltPaths res; + for (auto & p : Installable::build(evalStore, store, mode, installables)) + res.push_back(p.path); + return res; + } else { if (mode == Realise::Nothing) settings.readOnlyMode = true; @@ -980,7 +642,7 @@ StorePathSet Installable::toStorePaths( ref<Store> evalStore, ref<Store> store, Realise mode, OperateOn operateOn, - const std::vector<std::shared_ptr<Installable>> & installables) + const Installables & installables) { StorePathSet outPaths; for (auto & path : toBuiltPaths(evalStore, store, mode, operateOn, installables)) { @@ -994,7 +656,7 @@ StorePath Installable::toStorePath( ref<Store> evalStore, ref<Store> store, Realise mode, OperateOn operateOn, - std::shared_ptr<Installable> installable) + ref<Installable> installable) { auto paths = toStorePaths(evalStore, store, mode, operateOn, {installable}); @@ -1006,7 +668,7 @@ StorePath Installable::toStorePath( StorePathSet Installable::toDerivations( ref<Store> store, - const std::vector<std::shared_ptr<Installable>> & installables, + const Installables & installables, bool useDeriver) { StorePathSet drvPaths; @@ -1015,55 +677,74 @@ StorePathSet Installable::toDerivations( for (const auto & b : i->toDerivedPaths()) std::visit(overloaded { [&](const DerivedPath::Opaque & bo) { - if (!useDeriver) - throw Error("argument '%s' did not evaluate to a derivation", i->what()); - drvPaths.insert(getDeriver(store, *i, bo.path)); + drvPaths.insert( + bo.path.isDerivation() + ? bo.path + : useDeriver + ? getDeriver(store, *i, bo.path) + : throw Error("argument '%s' did not evaluate to a derivation", i->what())); }, [&](const DerivedPath::Built & bfd) { drvPaths.insert(bfd.drvPath); }, - }, b.raw()); + }, b.path.raw()); return drvPaths; } -InstallablesCommand::InstallablesCommand() +RawInstallablesCommand::RawInstallablesCommand() { + addFlag({ + .longName = "stdin", + .description = "Read installables from the standard input.", + .handler = {&readFromStdIn, true} + }); + expectArgs({ .label = "installables", - .handler = {&_installables}, + .handler = {&rawInstallables}, .completer = {[&](size_t, std::string_view prefix) { completeInstallable(prefix); }} }); } -void InstallablesCommand::prepare() +void RawInstallablesCommand::applyDefaultInstallables(std::vector<std::string> & rawInstallables) { - installables = load(); -} - -Installables InstallablesCommand::load() { - Installables installables; - if (_installables.empty() && useDefaultInstallables()) + if (rawInstallables.empty()) { // FIXME: commands like "nix profile install" should not have a // default, probably. - _installables.push_back("."); - return parseInstallables(getStore(), _installables); + rawInstallables.push_back("."); + } } -std::vector<std::string> InstallablesCommand::getFlakesForCompletion() +void RawInstallablesCommand::run(ref<Store> store) { - if (_installables.empty()) { - if (useDefaultInstallables()) - return {"."}; - return {}; + if (readFromStdIn && !isatty(STDIN_FILENO)) { + std::string word; + while (std::cin >> word) { + rawInstallables.emplace_back(std::move(word)); + } } - return _installables; + + applyDefaultInstallables(rawInstallables); + run(store, std::move(rawInstallables)); +} + +std::vector<std::string> RawInstallablesCommand::getFlakesForCompletion() +{ + applyDefaultInstallables(rawInstallables); + return rawInstallables; } -InstallableCommand::InstallableCommand(bool supportReadOnlyMode) - : SourceExprCommand(supportReadOnlyMode) +void InstallablesCommand::run(ref<Store> store, std::vector<std::string> && rawInstallables) +{ + auto installables = parseInstallables(store, rawInstallables); + run(store, std::move(installables)); +} + +InstallableCommand::InstallableCommand() + : SourceExprCommand() { expectArgs({ .label = "installable", @@ -1075,9 +756,16 @@ InstallableCommand::InstallableCommand(bool supportReadOnlyMode) }); } -void InstallableCommand::prepare() +void InstallableCommand::run(ref<Store> store) +{ + auto installable = parseInstallable(store, _installable); + run(store, std::move(installable)); +} + +void BuiltPathsCommand::applyDefaultInstallables(std::vector<std::string> & rawInstallables) { - installable = parseInstallable(getStore(), _installable); + if (rawInstallables.empty() && !all) + rawInstallables.push_back("."); } } diff --git a/src/libcmd/installables.hh b/src/libcmd/installables.hh index 948f78919..6c2922d89 100644 --- a/src/libcmd/installables.hh +++ b/src/libcmd/installables.hh @@ -2,11 +2,12 @@ #include "util.hh" #include "path.hh" -#include "path-with-outputs.hh" +#include "outputs-spec.hh" #include "derived-path.hh" #include "eval.hh" #include "store-api.hh" #include "flake/flake.hh" +#include "build-result.hh" #include <optional> @@ -19,7 +20,7 @@ namespace eval_cache { class EvalCache; class AttrCursor; } struct App { - std::vector<StorePathWithOutputs> context; + std::vector<DerivedPath> context; Path program; // FIXME: add args, sandbox settings, metadata, ... }; @@ -51,20 +52,45 @@ enum class OperateOn { Derivation }; +struct ExtraPathInfo +{ + std::optional<NixInt> priority; + std::optional<FlakeRef> originalRef; + std::optional<FlakeRef> resolvedRef; + std::optional<std::string> attrPath; + // FIXME: merge with DerivedPath's 'outputs' field? + std::optional<ExtendedOutputsSpec> extendedOutputsSpec; +}; + +/* A derived path with any additional info that commands might + need from the derivation. */ +struct DerivedPathWithInfo +{ + DerivedPath path; + ExtraPathInfo info; +}; + +struct BuiltPathWithResult +{ + BuiltPath path; + ExtraPathInfo info; + std::optional<BuildResult> result; +}; + +typedef std::vector<DerivedPathWithInfo> DerivedPathsWithInfo; + +struct Installable; +typedef std::vector<ref<Installable>> Installables; + struct Installable { virtual ~Installable() { } virtual std::string what() const = 0; - virtual DerivedPaths toDerivedPaths() = 0; + virtual DerivedPathsWithInfo toDerivedPaths() = 0; - virtual StorePathSet toDrvPaths(ref<Store> store) - { - throw Error("'%s' cannot be converted to a derivation path", what()); - } - - DerivedPath toDerivedPath(); + DerivedPathWithInfo toDerivedPath(); UnresolvedApp toApp(EvalState & state); @@ -80,9 +106,13 @@ struct Installable return {}; } + /* Get a cursor to each value this Installable could refer to. However + if none exists, throw exception instead of returning empty vector. */ virtual std::vector<ref<eval_cache::AttrCursor>> getCursors(EvalState & state); + /* Get the first and most preferred cursor this Installable could refer + to, or throw an exception if none exists. */ virtual ref<eval_cache::AttrCursor> getCursor(EvalState & state); @@ -91,18 +121,18 @@ struct Installable return FlakeRef::fromAttrs({{"type","indirect"}, {"id", "nixpkgs"}}); } - static BuiltPaths build( + static std::vector<BuiltPathWithResult> build( ref<Store> evalStore, ref<Store> store, Realise mode, - const std::vector<std::shared_ptr<Installable>> & installables, + const Installables & installables, BuildMode bMode = bmNormal); - static std::vector<std::pair<std::shared_ptr<Installable>, BuiltPath>> build2( + static std::vector<std::pair<ref<Installable>, BuiltPathWithResult>> build2( ref<Store> evalStore, ref<Store> store, Realise mode, - const std::vector<std::shared_ptr<Installable>> & installables, + const Installables & installables, BuildMode bMode = bmNormal); static std::set<StorePath> toStorePaths( @@ -110,18 +140,18 @@ struct Installable ref<Store> store, Realise mode, OperateOn operateOn, - const std::vector<std::shared_ptr<Installable>> & installables); + const Installables & installables); static StorePath toStorePath( ref<Store> evalStore, ref<Store> store, Realise mode, OperateOn operateOn, - std::shared_ptr<Installable> installable); + ref<Installable> installable); static std::set<StorePath> toDerivations( ref<Store> store, - const std::vector<std::shared_ptr<Installable>> & installables, + const Installables & installables, bool useDeriver = false); static BuiltPaths toBuiltPaths( @@ -129,78 +159,7 @@ struct Installable ref<Store> store, Realise mode, OperateOn operateOn, - const std::vector<std::shared_ptr<Installable>> & installables); -}; - -typedef std::vector<std::shared_ptr<Installable>> Installables; - -struct InstallableValue : Installable -{ - ref<EvalState> state; - - InstallableValue(ref<EvalState> state) : state(state) {} - - struct DerivationInfo - { - StorePath drvPath; - std::set<std::string> outputsToInstall; - std::optional<NixInt> priority; - }; - - virtual std::vector<DerivationInfo> toDerivations() = 0; - - DerivedPaths toDerivedPaths() override; - - StorePathSet toDrvPaths(ref<Store> store) override; -}; - -struct InstallableFlake : InstallableValue -{ - FlakeRef flakeRef; - Strings attrPaths; - Strings prefixes; - OutputsSpec outputsSpec; - const flake::LockFlags & lockFlags; - mutable std::shared_ptr<flake::LockedFlake> _lockedFlake; - - InstallableFlake( - SourceExprCommand * cmd, - ref<EvalState> state, - FlakeRef && flakeRef, - std::string_view fragment, - OutputsSpec outputsSpec, - Strings attrPaths, - Strings prefixes, - const flake::LockFlags & lockFlags); - - std::string what() const override { return flakeRef.to_string() + "#" + *attrPaths.begin(); } - - std::vector<std::string> getActualAttrPaths(); - - Value * getFlakeOutputs(EvalState & state, const flake::LockedFlake & lockedFlake); - - std::tuple<std::string, FlakeRef, DerivationInfo> toDerivation(); - - std::vector<DerivationInfo> toDerivations() override; - - std::pair<Value *, PosIdx> toValue(EvalState & state) override; - - /* Get a cursor to every attrpath in getActualAttrPaths() that - exists. */ - std::vector<ref<eval_cache::AttrCursor>> - getCursors(EvalState & state) override; - - /* Get a cursor to the first attrpath in getActualAttrPaths() that - exists, or throw an exception with suggestions if none exists. */ - ref<eval_cache::AttrCursor> getCursor(EvalState & state) override; - - std::shared_ptr<flake::LockedFlake> getLockedFlake() const; - - FlakeRef nixpkgsFlakeRef() const override; + const Installables & installables); }; -ref<eval_cache::EvalCache> openEvalCache( - EvalState & state, - std::shared_ptr<flake::LockedFlake> lockedFlake); - } diff --git a/src/libcmd/local.mk b/src/libcmd/local.mk index 3a4de6bcb..541a7d2ba 100644 --- a/src/libcmd/local.mk +++ b/src/libcmd/local.mk @@ -6,9 +6,9 @@ libcmd_DIR := $(d) libcmd_SOURCES := $(wildcard $(d)/*.cc) -libcmd_CXXFLAGS += -I src/libutil -I src/libstore -I src/libexpr -I src/libmain -I src/libfetchers -I src/nix +libcmd_CXXFLAGS += -I src/libutil -I src/libstore -I src/libexpr -I src/libmain -I src/libfetchers -libcmd_LDFLAGS = $(EDITLINE_LIBS) -llowdown -pthread +libcmd_LDFLAGS = $(EDITLINE_LIBS) $(LOWDOWN_LIBS) -pthread libcmd_LIBS = libstore libutil libexpr libmain libfetchers diff --git a/src/libcmd/markdown.cc b/src/libcmd/markdown.cc index 71f9c8dff..668a07763 100644 --- a/src/libcmd/markdown.cc +++ b/src/libcmd/markdown.cc @@ -18,7 +18,7 @@ std::string renderMarkdownToTerminal(std::string_view markdown) .hmargin = 0, .vmargin = 0, .feat = LOWDOWN_COMMONMARK | LOWDOWN_FENCED | LOWDOWN_DEFLIST | LOWDOWN_TABLES, - .oflags = 0, + .oflags = LOWDOWN_TERM_NOLINK, }; auto doc = lowdown_doc_new(&opts); diff --git a/src/libcmd/nix-cmd.pc.in b/src/libcmd/nix-cmd.pc.in index 1761a9f41..39575f222 100644 --- a/src/libcmd/nix-cmd.pc.in +++ b/src/libcmd/nix-cmd.pc.in @@ -6,4 +6,4 @@ Name: Nix Description: Nix Package Manager Version: @PACKAGE_VERSION@ Libs: -L${libdir} -lnixcmd -Cflags: -I${includedir}/nix -std=c++17 +Cflags: -I${includedir}/nix -std=c++2a diff --git a/src/libcmd/repl.cc b/src/libcmd/repl.cc index 23df40337..e3afb1531 100644 --- a/src/libcmd/repl.cc +++ b/src/libcmd/repl.cc @@ -19,6 +19,8 @@ extern "C" { } #endif +#include "repl.hh" + #include "ansicolor.hh" #include "shared.hh" #include "eval.hh" @@ -31,10 +33,13 @@ extern "C" { #include "get-drvs.hh" #include "derivations.hh" #include "globals.hh" -#include "command.hh" +#include "flake/flake.hh" +#include "flake/lockfile.hh" +#include "editor-for.hh" #include "finally.hh" #include "markdown.hh" #include "local-fs-store.hh" +#include "progress-bar.hh" #if HAVE_BOEHMGC #define GC_INCLUDE_NEW @@ -44,18 +49,16 @@ extern "C" { namespace nix { struct NixRepl + : AbstractNixRepl #if HAVE_BOEHMGC - : gc + , gc #endif { std::string curDir; - ref<EvalState> state; - Bindings * autoArgs; size_t debugTraceIndex; Strings loadedFiles; - typedef std::vector<std::pair<Value*,std::string>> AnnotatedValues; std::function<AnnotatedValues()> getValues; const static int envSize = 32768; @@ -68,8 +71,11 @@ struct NixRepl NixRepl(const Strings & searchPath, nix::ref<Store> store,ref<EvalState> state, std::function<AnnotatedValues()> getValues); - ~NixRepl(); - void mainLoop(); + virtual ~NixRepl(); + + void mainLoop() override; + void initEnv() override; + StringSet completePrefix(const std::string & prefix); bool getLine(std::string & input, const std::string & prompt); StorePath getDerivationPath(Value & v); @@ -77,7 +83,6 @@ struct NixRepl void loadFile(const Path & path); void loadFlake(const std::string & flakeRef); - void initEnv(); void loadFiles(); void reloadFiles(); void addAttrsToScope(Value & attrs); @@ -91,7 +96,6 @@ struct NixRepl std::ostream & printValue(std::ostream & str, Value & v, unsigned int maxDepth, ValuesSeen & seen); }; - std::string removeWhitespace(std::string s) { s = chomp(s); @@ -103,7 +107,7 @@ std::string removeWhitespace(std::string s) NixRepl::NixRepl(const Strings & searchPath, nix::ref<Store> store, ref<EvalState> state, std::function<NixRepl::AnnotatedValues()> getValues) - : state(state) + : AbstractNixRepl(state) , debugTraceIndex(0) , getValues(getValues) , staticEnv(new StaticEnv(false, state->staticBaseEnv.get())) @@ -214,17 +218,15 @@ static std::ostream & showDebugTrace(std::ostream & out, const PosTable & positi out << dt.hint.str() << "\n"; // prefer direct pos, but if noPos then try the expr. - auto pos = *dt.pos - ? *dt.pos - : positions[dt.expr.getPos() ? dt.expr.getPos() : noPos]; + auto pos = dt.pos + ? dt.pos + : static_cast<std::shared_ptr<AbstractPos>>(positions[dt.expr.getPos() ? dt.expr.getPos() : noPos]); if (pos) { - printAtPos(pos, out); - - auto loc = getCodeLines(pos); - if (loc.has_value()) { + out << pos; + if (auto loc = pos->getCodeLines()) { out << "\n"; - printCodeLines(out, "", pos, *loc); + printCodeLines(out, "", *pos, *loc); out << "\n"; } } @@ -241,7 +243,11 @@ void NixRepl::mainLoop() // Allow nix-repl specific settings in .inputrc rl_readline_name = "nix-repl"; - createDirs(dirOf(historyFile)); + try { + createDirs(dirOf(historyFile)); + } catch (SysError & e) { + logWarning(e.info()); + } #ifndef READLINE el_hist_size = 1000; #endif @@ -252,6 +258,10 @@ void NixRepl::mainLoop() rl_set_list_possib_func(listPossibleCallback); #endif + /* Stop the progress bar because it interferes with the display of + the repl. */ + stopProgressBar(); + std::string input; while (true) { @@ -261,6 +271,7 @@ void NixRepl::mainLoop() // ctrl-D should exit the debugger. state->debugStop = false; state->debugQuit = true; + logger->cout(""); break; } try { @@ -375,6 +386,10 @@ StringSet NixRepl::completePrefix(const std::string & prefix) i++; } } else { + /* Temporarily disable the debugger, to avoid re-entering readline. */ + auto debug_repl = state->debugRepl; + state->debugRepl = nullptr; + Finally restoreDebug([&]() { state->debugRepl = debug_repl; }); try { /* This is an expression that should evaluate to an attribute set. Evaluate it to get the names of the @@ -385,7 +400,7 @@ StringSet NixRepl::completePrefix(const std::string & prefix) Expr * e = parseString(expr); Value v; e->eval(*state, *env, v); - state->forceAttrs(v, noPos); + state->forceAttrs(v, noPos, "while evaluating an attrset for the purpose of completion (this error should not be displayed; file an issue?)"); for (auto & i : *v.attrs) { std::string_view name = state->symbols[i.name]; @@ -575,15 +590,17 @@ bool NixRepl::processLine(std::string line) Value v; evalString(arg, v); - const auto [file, line] = [&] () -> std::pair<std::string, uint32_t> { + const auto [path, line] = [&] () -> std::pair<Path, uint32_t> { if (v.type() == nPath || v.type() == nString) { PathSet context; - auto filename = state->coerceToString(noPos, v, context).toOwned(); - state->symbols.create(filename); - return {filename, 0}; + auto path = state->coerceToPath(noPos, v, context, "while evaluating the filename to edit"); + return {path, 0}; } else if (v.isLambda()) { auto pos = state->positions[v.lambda.fun->pos]; - return {pos.file, pos.line}; + if (auto path = std::get_if<Path>(&pos.origin)) + return {*path, pos.line}; + else + throw Error("'%s' cannot be shown in an editor", pos); } else { // assume it's a derivation return findPackageFilename(*state, v, arg); @@ -591,7 +608,7 @@ bool NixRepl::processLine(std::string line) }(); // Open in EDITOR - auto args = editorFor(file, line); + auto args = editorFor(path, line); auto editor = args.front(); args.pop_front(); @@ -627,7 +644,12 @@ bool NixRepl::processLine(std::string line) Path drvPathRaw = state->store->printStorePath(drvPath); if (command == ":b" || command == ":bl") { - state->store->buildPaths({DerivedPath::Built{drvPath}}); + state->store->buildPaths({ + DerivedPath::Built { + .drvPath = drvPath, + .outputs = OutputsSpec::All { }, + }, + }); auto drv = state->store->readDerivation(drvPath); logger->cout("\nThis derivation produced the following outputs:"); for (auto & [outputName, outputPath] : state->store->queryDerivationOutputMap(drvPath)) { @@ -773,7 +795,7 @@ void NixRepl::loadFlake(const std::string & flakeRefS) flake::LockFlags { .updateLockFile = false, .useRegistries = !evalSettings.pureEval, - .allowMutable = !evalSettings.pureEval, + .allowUnlocked = !evalSettings.pureEval, }), v); addAttrsToScope(v); @@ -820,7 +842,7 @@ void NixRepl::loadFiles() void NixRepl::addAttrsToScope(Value & attrs) { - state->forceAttrs(attrs, [&]() { return attrs.determinePos(noPos); }); + state->forceAttrs(attrs, [&]() { return attrs.determinePos(noPos); }, "while evaluating an attribute set to be merged in the global scope"); if (displ + attrs.attrs->size() >= envSize) throw Error("environment full; cannot add more variables"); @@ -925,7 +947,7 @@ std::ostream & NixRepl::printValue(std::ostream & str, Value & v, unsigned int m Bindings::iterator i = v.attrs->find(state->sDrvPath); PathSet context; if (i != v.attrs->end()) - str << state->store->printStorePath(state->coerceToStorePath(i->pos, *i->value, context)); + str << state->store->printStorePath(state->coerceToStorePath(i->pos, *i->value, context, "while evaluating the drvPath of a derivation")); else str << "???"; str << "»"; @@ -1010,8 +1032,22 @@ std::ostream & NixRepl::printValue(std::ostream & str, Value & v, unsigned int m return str; } -void runRepl( - ref<EvalState>evalState, + +std::unique_ptr<AbstractNixRepl> AbstractNixRepl::create( + const Strings & searchPath, nix::ref<Store> store, ref<EvalState> state, + std::function<AnnotatedValues()> getValues) +{ + return std::make_unique<NixRepl>( + searchPath, + openStore(), + state, + getValues + ); +} + + +void AbstractNixRepl::runSimple( + ref<EvalState> evalState, const ValMap & extraEnv) { auto getValues = [&]()->NixRepl::AnnotatedValues{ @@ -1035,87 +1071,4 @@ void runRepl( repl->mainLoop(); } -struct CmdRepl : InstallablesCommand -{ - CmdRepl(){ - evalSettings.pureEval = false; - } - void prepare() - { - if (!settings.isExperimentalFeatureEnabled(Xp::ReplFlake) && !(file) && this->_installables.size() >= 1) { - warn("future versions of Nix will require using `--file` to load a file"); - if (this->_installables.size() > 1) - warn("more than one input file is not currently supported"); - auto filePath = this->_installables[0].data(); - file = std::optional(filePath); - _installables.front() = _installables.back(); - _installables.pop_back(); - } - installables = InstallablesCommand::load(); - } - std::vector<std::string> files; - Strings getDefaultFlakeAttrPaths() override - { - return {""}; - } - virtual bool useDefaultInstallables() override - { - return file.has_value() or expr.has_value(); - } - - bool forceImpureByDefault() override - { - return true; - } - - std::string description() override - { - return "start an interactive environment for evaluating Nix expressions"; - } - - std::string doc() override - { - return - #include "repl.md" - ; - } - - void run(ref<Store> store) override - { - auto state = getEvalState(); - auto getValues = [&]()->NixRepl::AnnotatedValues{ - auto installables = load(); - NixRepl::AnnotatedValues values; - for (auto & installable: installables){ - auto what = installable->what(); - if (file){ - auto [val, pos] = installable->toValue(*state); - auto what = installable->what(); - state->forceValue(*val, pos); - auto autoArgs = getAutoArgs(*state); - auto valPost = state->allocValue(); - state->autoCallFunction(*autoArgs, *val, *valPost); - state->forceValue(*valPost, pos); - values.push_back( {valPost, what }); - } else { - auto [val, pos] = installable->toValue(*state); - values.push_back( {val, what} ); - } - } - return values; - }; - auto repl = std::make_unique<NixRepl>( - searchPath, - openStore(), - state, - getValues - ); - repl->autoArgs = getAutoArgs(*repl->state); - repl->initEnv(); - repl->mainLoop(); - } -}; - -static auto rCmdRepl = registerCommand<CmdRepl>("repl"); - } diff --git a/src/libcmd/repl.hh b/src/libcmd/repl.hh new file mode 100644 index 000000000..dfccc93e7 --- /dev/null +++ b/src/libcmd/repl.hh @@ -0,0 +1,39 @@ +#pragma once + +#include "eval.hh" + +#if HAVE_BOEHMGC +#define GC_INCLUDE_NEW +#include <gc/gc_cpp.h> +#endif + +namespace nix { + +struct AbstractNixRepl +{ + ref<EvalState> state; + Bindings * autoArgs; + + AbstractNixRepl(ref<EvalState> state) + : state(state) + { } + + virtual ~AbstractNixRepl() + { } + + typedef std::vector<std::pair<Value*,std::string>> AnnotatedValues; + + static std::unique_ptr<AbstractNixRepl> create( + const Strings & searchPath, nix::ref<Store> store, ref<EvalState> state, + std::function<AnnotatedValues()> getValues); + + static void runSimple( + ref<EvalState> evalState, + const ValMap & extraEnv); + + virtual void initEnv() = 0; + + virtual void mainLoop() = 0; +}; + +} diff --git a/src/libexpr/attr-path.cc b/src/libexpr/attr-path.cc index 94ab60f9a..7c0705091 100644 --- a/src/libexpr/attr-path.cc +++ b/src/libexpr/attr-path.cc @@ -118,7 +118,7 @@ std::pair<std::string, uint32_t> findPackageFilename(EvalState & state, Value & // FIXME: is it possible to extract the Pos object instead of doing this // toString + parsing? - auto pos = state.forceString(*v2); + auto pos = state.forceString(*v2, noPos, "while evaluating the 'meta.position' attribute of a derivation"); auto colon = pos.rfind(':'); if (colon == std::string::npos) diff --git a/src/libexpr/eval-cache.cc b/src/libexpr/eval-cache.cc index 0d83b6cfe..1219b2471 100644 --- a/src/libexpr/eval-cache.cc +++ b/src/libexpr/eval-cache.cc @@ -300,7 +300,7 @@ struct AttrDb NixStringContext context; if (!queryAttribute.isNull(3)) for (auto & s : tokenizeString<std::vector<std::string>>(queryAttribute.getStr(3), ";")) - context.push_back(decodeContext(cfg, s)); + context.push_back(NixStringContextElem::parse(cfg, s)); return {{rowId, string_t{queryAttribute.getStr(2), context}}}; } case AttrType::Bool: @@ -385,7 +385,7 @@ Value & AttrCursor::getValue() if (!_value) { if (parent) { auto & vParent = parent->first->getValue(); - root->state.forceAttrs(vParent, noPos); + root->state.forceAttrs(vParent, noPos, "while searching for an attribute"); auto attr = vParent.attrs->get(parent->second); if (!attr) throw Error("attribute '%s' is unexpectedly missing", getAttrPathStr()); @@ -507,11 +507,6 @@ std::shared_ptr<AttrCursor> AttrCursor::maybeGetAttr(Symbol name, bool forceErro return nullptr; //throw TypeError("'%s' is not an attribute set", getAttrPathStr()); - for (auto & attr : *v.attrs) { - if (root->db) - root->db->setPlaceholder({cachedValue->first, attr.name}); - } - auto attr = v.attrs->get(name); if (!attr) { @@ -576,14 +571,14 @@ std::string AttrCursor::getString() debug("using cached string attribute '%s'", getAttrPathStr()); return s->first; } else - root->state.debugThrowLastTrace(TypeError("'%s' is not a string", getAttrPathStr())); + root->state.error("'%s' is not a string", getAttrPathStr()).debugThrow<TypeError>(); } } auto & v = forceValue(); if (v.type() != nString && v.type() != nPath) - root->state.debugThrowLastTrace(TypeError("'%s' is not a string but %s", getAttrPathStr(), showType(v.type()))); + root->state.error("'%s' is not a string but %s", getAttrPathStr()).debugThrow<TypeError>(); return v.type() == nString ? v.string.s : v.path; } @@ -597,7 +592,18 @@ string_t AttrCursor::getStringWithContext() if (auto s = std::get_if<string_t>(&cachedValue->second)) { bool valid = true; for (auto & c : s->second) { - if (!root->state.store->isValidPath(c.first)) { + const StorePath & path = std::visit(overloaded { + [&](const NixStringContextElem::DrvDeep & d) -> const StorePath & { + return d.drvPath; + }, + [&](const NixStringContextElem::Built & b) -> const StorePath & { + return b.drvPath; + }, + [&](const NixStringContextElem::Opaque & o) -> const StorePath & { + return o.path; + }, + }, c.raw()); + if (!root->state.store->isValidPath(path)) { valid = false; break; } @@ -607,7 +613,7 @@ string_t AttrCursor::getStringWithContext() return *s; } } else - root->state.debugThrowLastTrace(TypeError("'%s' is not a string", getAttrPathStr())); + root->state.error("'%s' is not a string", getAttrPathStr()).debugThrow<TypeError>(); } } @@ -618,7 +624,7 @@ string_t AttrCursor::getStringWithContext() else if (v.type() == nPath) return {v.path, {}}; else - root->state.debugThrowLastTrace(TypeError("'%s' is not a string but %s", getAttrPathStr(), showType(v.type()))); + root->state.error("'%s' is not a string but %s", getAttrPathStr()).debugThrow<TypeError>(); } bool AttrCursor::getBool() @@ -631,14 +637,14 @@ bool AttrCursor::getBool() debug("using cached Boolean attribute '%s'", getAttrPathStr()); return *b; } else - root->state.debugThrowLastTrace(TypeError("'%s' is not a Boolean", getAttrPathStr())); + root->state.error("'%s' is not a Boolean", getAttrPathStr()).debugThrow<TypeError>(); } } auto & v = forceValue(); if (v.type() != nBool) - root->state.debugThrowLastTrace(TypeError("'%s' is not a Boolean", getAttrPathStr())); + root->state.error("'%s' is not a Boolean", getAttrPathStr()).debugThrow<TypeError>(); return v.boolean; } @@ -650,17 +656,17 @@ NixInt AttrCursor::getInt() cachedValue = root->db->getAttr(getKey()); if (cachedValue && !std::get_if<placeholder_t>(&cachedValue->second)) { if (auto i = std::get_if<int_t>(&cachedValue->second)) { - debug("using cached Integer attribute '%s'", getAttrPathStr()); + debug("using cached integer attribute '%s'", getAttrPathStr()); return i->x; } else - throw TypeError("'%s' is not an Integer", getAttrPathStr()); + throw TypeError("'%s' is not an integer", getAttrPathStr()); } } auto & v = forceValue(); if (v.type() != nInt) - throw TypeError("'%s' is not an Integer", getAttrPathStr()); + throw TypeError("'%s' is not an integer", getAttrPathStr()); return v.integer; } @@ -690,7 +696,7 @@ std::vector<std::string> AttrCursor::getListOfStrings() std::vector<std::string> res; for (auto & elem : v.listItems()) - res.push_back(std::string(root->state.forceStringNoCtx(*elem))); + res.push_back(std::string(root->state.forceStringNoCtx(*elem, noPos, "while evaluating an attribute for caching"))); if (root->db) cachedValue = {root->db->setListOfStrings(getKey(), res), res}; @@ -708,14 +714,14 @@ std::vector<Symbol> AttrCursor::getAttrs() debug("using cached attrset attribute '%s'", getAttrPathStr()); return *attrs; } else - root->state.debugThrowLastTrace(TypeError("'%s' is not an attribute set", getAttrPathStr())); + root->state.error("'%s' is not an attribute set", getAttrPathStr()).debugThrow<TypeError>(); } } auto & v = forceValue(); if (v.type() != nAttrs) - root->state.debugThrowLastTrace(TypeError("'%s' is not an attribute set", getAttrPathStr())); + root->state.error("'%s' is not an attribute set", getAttrPathStr()).debugThrow<TypeError>(); std::vector<Symbol> attrs; for (auto & attr : *getValue().attrs) diff --git a/src/libexpr/eval-inline.hh b/src/libexpr/eval-inline.hh index f2f4ba725..f0da688db 100644 --- a/src/libexpr/eval-inline.hh +++ b/src/libexpr/eval-inline.hh @@ -103,33 +103,36 @@ void EvalState::forceValue(Value & v, Callable getPos) else if (v.isApp()) callFunction(*v.app.left, *v.app.right, v, noPos); else if (v.isBlackhole()) - throwEvalError(getPos(), "infinite recursion encountered"); + error("infinite recursion encountered").atPos(getPos()).template debugThrow<EvalError>(); } [[gnu::always_inline]] -inline void EvalState::forceAttrs(Value & v, const PosIdx pos) +inline void EvalState::forceAttrs(Value & v, const PosIdx pos, std::string_view errorCtx) { - forceAttrs(v, [&]() { return pos; }); + forceAttrs(v, [&]() { return pos; }, errorCtx); } template <typename Callable> [[gnu::always_inline]] -inline void EvalState::forceAttrs(Value & v, Callable getPos) +inline void EvalState::forceAttrs(Value & v, Callable getPos, std::string_view errorCtx) { - forceValue(v, getPos); - if (v.type() != nAttrs) - throwTypeError(getPos(), "value is %1% while a set was expected", v); + forceValue(v, noPos); + if (v.type() != nAttrs) { + PosIdx pos = getPos(); + error("value is %1% while a set was expected", showType(v)).withTrace(pos, errorCtx).debugThrow<TypeError>(); + } } [[gnu::always_inline]] -inline void EvalState::forceList(Value & v, const PosIdx pos) +inline void EvalState::forceList(Value & v, const PosIdx pos, std::string_view errorCtx) { - forceValue(v, pos); - if (!v.isList()) - throwTypeError(pos, "value is %1% while a list was expected", v); + forceValue(v, noPos); + if (!v.isList()) { + error("value is %1% while a list was expected", showType(v)).withTrace(pos, errorCtx).debugThrow<TypeError>(); + } } diff --git a/src/libexpr/eval.cc b/src/libexpr/eval.cc index e3716f217..2721b6733 100644 --- a/src/libexpr/eval.cc +++ b/src/libexpr/eval.cc @@ -7,12 +7,13 @@ #include "globals.hh" #include "eval-inline.hh" #include "filetransfer.hh" -#include "json.hh" #include "function-trace.hh" #include <algorithm> #include <chrono> +#include <iostream> #include <cstring> +#include <optional> #include <unistd.h> #include <sys/time.h> #include <sys/resource.h> @@ -21,6 +22,7 @@ #include <functional> #include <sys/resource.h> +#include <nlohmann/json.hpp> #if HAVE_BOEHMGC @@ -35,6 +37,8 @@ #endif +using json = nlohmann::json; + namespace nix { static char * allocString(size_t size) @@ -43,7 +47,7 @@ static char * allocString(size_t size) #if HAVE_BOEHMGC t = (char *) GC_MALLOC_ATOMIC(size); #else - t = malloc(size); + t = (char *) malloc(size); #endif if (!t) throw std::bad_alloc(); return t; @@ -65,26 +69,19 @@ static char * dupString(const char * s) // When there's no need to write to the string, we can optimize away empty // string allocations. -// This function handles makeImmutableStringWithLen(null, 0) by returning the -// empty string. -static const char * makeImmutableStringWithLen(const char * s, size_t size) +// This function handles makeImmutableString(std::string_view()) by returning +// the empty string. +static const char * makeImmutableString(std::string_view s) { - char * t; + const size_t size = s.size(); if (size == 0) return ""; -#if HAVE_BOEHMGC - t = GC_STRNDUP(s, size); -#else - t = strndup(s, size); -#endif - if (!t) throw std::bad_alloc(); + auto t = allocString(size + 1); + memcpy(t, s.data(), size); + t[size] = '\0'; return t; } -static inline const char * makeImmutableString(std::string_view s) { - return makeImmutableStringWithLen(s.data(), s.size()); -} - RootValue allocRootValue(Value * v) { @@ -323,7 +320,7 @@ static Symbol getName(const AttrName & name, EvalState & state, Env & env) } else { Value nameValue; name.expr->eval(state, env, nameValue); - state.forceStringNoCtx(nameValue); + state.forceStringNoCtx(nameValue, noPos, "while evaluating an attribute name"); return state.symbols.create(nameValue.string.s); } } @@ -371,7 +368,7 @@ void initGC() size = (pageSize * pages) / 4; // 25% of RAM if (size > maxSize) size = maxSize; #endif - debug(format("setting initial heap size to %1% bytes") % size); + debug("setting initial heap size to %1% bytes", size); GC_expand_hp(size); } @@ -404,7 +401,8 @@ static Strings parseNixPath(const std::string & s) } if (*p == ':') { - if (isUri(std::string(start2, s.end()))) { + auto prefix = std::string(start2, s.end()); + if (EvalSettings::isPseudoUrl(prefix) || hasPrefix(prefix, "flake:")) { ++p; while (p != s.end() && *p != ':') ++p; } @@ -418,6 +416,44 @@ static Strings parseNixPath(const std::string & s) return res; } +ErrorBuilder & ErrorBuilder::atPos(PosIdx pos) +{ + info.errPos = state.positions[pos]; + return *this; +} + +ErrorBuilder & ErrorBuilder::withTrace(PosIdx pos, const std::string_view text) +{ + info.traces.push_front(Trace{ .pos = state.positions[pos], .hint = hintformat(std::string(text)), .frame = false }); + return *this; +} + +ErrorBuilder & ErrorBuilder::withFrameTrace(PosIdx pos, const std::string_view text) +{ + info.traces.push_front(Trace{ .pos = state.positions[pos], .hint = hintformat(std::string(text)), .frame = true }); + return *this; +} + +ErrorBuilder & ErrorBuilder::withSuggestions(Suggestions & s) +{ + info.suggestions = s; + return *this; +} + +ErrorBuilder & ErrorBuilder::withFrame(const Env & env, const Expr & expr) +{ + // NOTE: This is abusing side-effects. + // TODO: check compatibility with nested debugger calls. + state.debugTraces.push_front(DebugTrace { + .pos = nullptr, + .expr = expr, + .env = env, + .hint = hintformat("Fake frame for debugging purposes"), + .isError = true + }); + return *this; +} + EvalState::EvalState( const Strings & _searchPath, @@ -472,9 +508,6 @@ EvalState::EvalState( #if HAVE_BOEHMGC , valueAllocCache(std::allocate_shared<void *>(traceable_allocator<void *>(), nullptr)) , env1AllocCache(std::allocate_shared<void *>(traceable_allocator<void *>(), nullptr)) -#else - , valueAllocCache(std::make_shared<void *>(nullptr)) - , env1AllocCache(std::make_shared<void *>(nullptr)) #endif , baseEnv(allocEnv(128)) , staticBaseEnv{std::make_shared<StaticEnv>(false, nullptr)} @@ -576,7 +609,7 @@ Path EvalState::checkSourcePath(const Path & path_) } /* Resolve symlinks. */ - debug(format("checking access to '%s'") % abspath); + debug("checking access to '%s'", abspath); Path path = canonPath(abspath, true); for (auto & i : *allowedPaths) { @@ -653,25 +686,7 @@ void EvalState::addConstant(const std::string & name, Value * v) Value * EvalState::addPrimOp(const std::string & name, size_t arity, PrimOpFun primOp) { - auto name2 = name.substr(0, 2) == "__" ? name.substr(2) : name; - auto sym = symbols.create(name2); - - /* Hack to make constants lazy: turn them into a application of - the primop to a dummy value. */ - if (arity == 0) { - auto vPrimOp = allocValue(); - vPrimOp->mkPrimOp(new PrimOp { .fun = primOp, .arity = 1, .name = name2 }); - Value v; - v.mkApp(vPrimOp, vPrimOp); - return addConstant(name, v); - } - - Value * v = allocValue(); - v->mkPrimOp(new PrimOp { .fun = primOp, .arity = arity, .name = name2 }); - staticBaseEnv->vars.emplace_back(symbols.create(name), baseEnvDispl); - baseEnv.values[baseEnvDispl++] = v; - baseEnv.values[0]->attrs->push_back(Attr(sym, v)); - return v; + return addPrimOp(PrimOp { .fun = primOp, .arity = arity, .name = name }); } @@ -824,7 +839,7 @@ void EvalState::runDebugRepl(const Error * error, const Env & env, const Expr & ? std::make_unique<DebugTraceStacker>( *this, DebugTrace { - .pos = error->info().errPos ? *error->info().errPos : positions[expr.getPos()], + .pos = error->info().errPos ? error->info().errPos : static_cast<std::shared_ptr<AbstractPos>>(positions[expr.getPos()]), .expr = expr, .env = env, .hint = error->info().msg, @@ -849,189 +864,27 @@ void EvalState::runDebugRepl(const Error * error, const Env & env, const Expr & } } -/* Every "format" object (even temporary) takes up a few hundred bytes - of stack space, which is a real killer in the recursive - evaluator. So here are some helper functions for throwing - exceptions. */ -void EvalState::throwEvalError(const PosIdx pos, const char * s, Env & env, Expr & expr) -{ - debugThrow(EvalError({ - .msg = hintfmt(s), - .errPos = positions[pos] - }), env, expr); -} - -void EvalState::throwEvalError(const PosIdx pos, const char * s) -{ - debugThrowLastTrace(EvalError({ - .msg = hintfmt(s), - .errPos = positions[pos] - })); -} - -void EvalState::throwEvalError(const char * s, const std::string & s2) -{ - debugThrowLastTrace(EvalError(s, s2)); -} - -void EvalState::throwEvalError(const PosIdx pos, const Suggestions & suggestions, const char * s, - const std::string & s2, Env & env, Expr & expr) -{ - debugThrow(EvalError(ErrorInfo{ - .msg = hintfmt(s, s2), - .errPos = positions[pos], - .suggestions = suggestions, - }), env, expr); -} - -void EvalState::throwEvalError(const PosIdx pos, const char * s, const std::string & s2) -{ - debugThrowLastTrace(EvalError({ - .msg = hintfmt(s, s2), - .errPos = positions[pos] - })); -} - -void EvalState::throwEvalError(const PosIdx pos, const char * s, const std::string & s2, Env & env, Expr & expr) -{ - debugThrow(EvalError({ - .msg = hintfmt(s, s2), - .errPos = positions[pos] - }), env, expr); -} - -void EvalState::throwEvalError(const char * s, const std::string & s2, - const std::string & s3) -{ - debugThrowLastTrace(EvalError({ - .msg = hintfmt(s, s2), - .errPos = positions[noPos] - })); -} - -void EvalState::throwEvalError(const PosIdx pos, const char * s, const std::string & s2, - const std::string & s3) -{ - debugThrowLastTrace(EvalError({ - .msg = hintfmt(s, s2), - .errPos = positions[pos] - })); -} - -void EvalState::throwEvalError(const PosIdx pos, const char * s, const std::string & s2, - const std::string & s3, Env & env, Expr & expr) -{ - debugThrow(EvalError({ - .msg = hintfmt(s, s2), - .errPos = positions[pos] - }), env, expr); -} - -void EvalState::throwEvalError(const PosIdx p1, const char * s, const Symbol sym, const PosIdx p2, Env & env, Expr & expr) -{ - // p1 is where the error occurred; p2 is a position mentioned in the message. - debugThrow(EvalError({ - .msg = hintfmt(s, symbols[sym], positions[p2]), - .errPos = positions[p1] - }), env, expr); -} - -void EvalState::throwTypeError(const PosIdx pos, const char * s, const Value & v) -{ - debugThrowLastTrace(TypeError({ - .msg = hintfmt(s, showType(v)), - .errPos = positions[pos] - })); -} - -void EvalState::throwTypeError(const PosIdx pos, const char * s, const Value & v, Env & env, Expr & expr) -{ - debugThrow(TypeError({ - .msg = hintfmt(s, showType(v)), - .errPos = positions[pos] - }), env, expr); -} - -void EvalState::throwTypeError(const PosIdx pos, const char * s) -{ - debugThrowLastTrace(TypeError({ - .msg = hintfmt(s), - .errPos = positions[pos] - })); -} - -void EvalState::throwTypeError(const PosIdx pos, const char * s, const ExprLambda & fun, - const Symbol s2, Env & env, Expr &expr) -{ - debugThrow(TypeError({ - .msg = hintfmt(s, fun.showNamePos(*this), symbols[s2]), - .errPos = positions[pos] - }), env, expr); -} - -void EvalState::throwTypeError(const PosIdx pos, const Suggestions & suggestions, const char * s, - const ExprLambda & fun, const Symbol s2, Env & env, Expr &expr) -{ - debugThrow(TypeError(ErrorInfo { - .msg = hintfmt(s, fun.showNamePos(*this), symbols[s2]), - .errPos = positions[pos], - .suggestions = suggestions, - }), env, expr); -} - -void EvalState::throwTypeError(const char * s, const Value & v, Env & env, Expr &expr) -{ - debugThrow(TypeError({ - .msg = hintfmt(s, showType(v)), - .errPos = positions[expr.getPos()], - }), env, expr); -} - -void EvalState::throwAssertionError(const PosIdx pos, const char * s, const std::string & s1, Env & env, Expr &expr) -{ - debugThrow(AssertionError({ - .msg = hintfmt(s, s1), - .errPos = positions[pos] - }), env, expr); -} - -void EvalState::throwUndefinedVarError(const PosIdx pos, const char * s, const std::string & s1, Env & env, Expr &expr) -{ - debugThrow(UndefinedVarError({ - .msg = hintfmt(s, s1), - .errPos = positions[pos] - }), env, expr); -} - -void EvalState::throwMissingArgumentError(const PosIdx pos, const char * s, const std::string & s1, Env & env, Expr &expr) -{ - debugThrow(MissingArgumentError({ - .msg = hintfmt(s, s1), - .errPos = positions[pos] - }), env, expr); -} - void EvalState::addErrorTrace(Error & e, const char * s, const std::string & s2) const { - e.addTrace(std::nullopt, s, s2); + e.addTrace(nullptr, s, s2); } -void EvalState::addErrorTrace(Error & e, const PosIdx pos, const char * s, const std::string & s2) const +void EvalState::addErrorTrace(Error & e, const PosIdx pos, const char * s, const std::string & s2, bool frame) const { - e.addTrace(positions[pos], s, s2); + e.addTrace(positions[pos], hintfmt(s, s2), frame); } static std::unique_ptr<DebugTraceStacker> makeDebugTraceStacker( EvalState & state, Expr & expr, Env & env, - std::optional<ErrPos> pos, + std::shared_ptr<AbstractPos> && pos, const char * s, const std::string & s2) { return std::make_unique<DebugTraceStacker>(state, DebugTrace { - .pos = pos, + .pos = std::move(pos), .expr = expr, .env = env, .hint = hintfmt(s, s2), @@ -1095,7 +948,7 @@ inline Value * EvalState::lookupVar(Env * env, const ExprVar & var, bool noEval) if (env->type == Env::HasWithExpr) { if (noEval) return 0; Value * v = allocValue(); - evalAttrs(*env->up, (Expr *) env->values[0], *v); + evalAttrs(*env->up, (Expr *) env->values[0], *v, noPos, "<borked>"); env->values[0] = v; env->type = Env::HasWithAttrs; } @@ -1105,7 +958,7 @@ inline Value * EvalState::lookupVar(Env * env, const ExprVar & var, bool noEval) return j->value; } if (!env->prevWith) - throwUndefinedVarError(var.pos, "undefined variable '%1%'", symbols[var.name], *env, const_cast<ExprVar&>(var)); + error("undefined variable '%1%'", symbols[var.name]).atPos(var.pos).withFrame(*env, var).debugThrow<UndefinedVarError>(); for (size_t l = env->prevWith; l; --l, env = env->up) ; } } @@ -1137,9 +990,9 @@ void EvalState::mkThunk_(Value & v, Expr * expr) void EvalState::mkPos(Value & v, PosIdx p) { auto pos = positions[p]; - if (!pos.file.empty()) { + if (auto path = std::get_if<Path>(&pos.origin)) { auto attrs = buildBindings(3); - attrs.alloc(sFile).mkString(pos.file); + attrs.alloc(sFile).mkString(*path); attrs.alloc(sLine).mkInt(pos.line); attrs.alloc(sColumn).mkInt(pos.column); v.mkAttrs(attrs); @@ -1247,7 +1100,7 @@ void EvalState::cacheFile( *this, *e, this->baseEnv, - e->getPos() ? std::optional(ErrPos(positions[e->getPos()])) : std::nullopt, + e->getPos() ? static_cast<std::shared_ptr<AbstractPos>>(positions[e->getPos()]) : nullptr, "while evaluating the file '%1%':", resolvedPath) : nullptr; @@ -1255,7 +1108,7 @@ void EvalState::cacheFile( // computation. if (mustBeTrivial && !(dynamic_cast<ExprAttrs *>(e))) - throw EvalError("file '%s' must be an attribute set", path); + error("file '%s' must be an attribute set", path).debugThrow<EvalError>(); eval(e, v); } catch (Error & e) { addErrorTrace(e, "while evaluating the file '%1%':", resolvedPath); @@ -1273,31 +1126,31 @@ void EvalState::eval(Expr * e, Value & v) } -inline bool EvalState::evalBool(Env & env, Expr * e) +inline bool EvalState::evalBool(Env & env, Expr * e, const PosIdx pos, std::string_view errorCtx) { - Value v; - e->eval(*this, env, v); - if (v.type() != nBool) - throwTypeError(noPos, "value is %1% while a Boolean was expected", v, env, *e); - return v.boolean; -} - - -inline bool EvalState::evalBool(Env & env, Expr * e, const PosIdx pos) -{ - Value v; - e->eval(*this, env, v); - if (v.type() != nBool) - throwTypeError(pos, "value is %1% while a Boolean was expected", v, env, *e); - return v.boolean; + try { + Value v; + e->eval(*this, env, v); + if (v.type() != nBool) + error("value is %1% while a Boolean was expected", showType(v)).withFrame(env, *e).debugThrow<TypeError>(); + return v.boolean; + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; + } } -inline void EvalState::evalAttrs(Env & env, Expr * e, Value & v) +inline void EvalState::evalAttrs(Env & env, Expr * e, Value & v, const PosIdx pos, std::string_view errorCtx) { - e->eval(*this, env, v); - if (v.type() != nAttrs) - throwTypeError(noPos, "value is %1% while a set was expected", v, env, *e); + try { + e->eval(*this, env, v); + if (v.type() != nAttrs) + error("value is %1% while a set was expected", showType(v)).withFrame(env, *e).debugThrow<TypeError>(); + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; + } } @@ -1370,7 +1223,7 @@ void ExprAttrs::eval(EvalState & state, Env & env, Value & v) Hence we need __overrides.) */ if (hasOverrides) { Value * vOverrides = (*v.attrs)[overrides->second.displ].value; - state.forceAttrs(*vOverrides, [&]() { return vOverrides->determinePos(noPos); }); + state.forceAttrs(*vOverrides, [&]() { return vOverrides->determinePos(noPos); }, "while evaluating the `__overrides` attribute"); Bindings * newBnds = state.allocBindings(v.attrs->capacity() + vOverrides->attrs->size()); for (auto & i : *v.attrs) newBnds->push_back(i); @@ -1398,11 +1251,11 @@ void ExprAttrs::eval(EvalState & state, Env & env, Value & v) state.forceValue(nameVal, i.pos); if (nameVal.type() == nNull) continue; - state.forceStringNoCtx(nameVal); + state.forceStringNoCtx(nameVal, i.pos, "while evaluating the name of a dynamic attribute"); auto nameSym = state.symbols.create(nameVal.string.s); Bindings::iterator j = v.attrs->find(nameSym); if (j != v.attrs->end()) - state.throwEvalError(i.pos, "dynamic attribute '%1%' already defined at %2%", nameSym, j->pos, env, *this); + state.error("dynamic attribute '%1%' already defined at %2%", state.symbols[nameSym], state.positions[j->pos]).atPos(i.pos).withFrame(env, *this).debugThrow<EvalError>(); i.valueExpr->setName(nameSym); /* Keep sorted order so find can catch duplicates */ @@ -1499,15 +1352,14 @@ void ExprSelect::eval(EvalState & state, Env & env, Value & v) return; } } else { - state.forceAttrs(*vAttrs, pos); + state.forceAttrs(*vAttrs, pos, "while selecting an attribute"); if ((j = vAttrs->attrs->find(name)) == vAttrs->attrs->end()) { std::set<std::string> allAttrNames; for (auto & attr : *vAttrs->attrs) allAttrNames.insert(state.symbols[attr.name]); - state.throwEvalError( - pos, - Suggestions::bestMatches(allAttrNames, state.symbols[name]), - "attribute '%1%' missing", state.symbols[name], env, *this); + auto suggestions = Suggestions::bestMatches(allAttrNames, state.symbols[name]); + state.error("attribute '%1%' missing", state.symbols[name]) + .atPos(pos).withSuggestions(suggestions).withFrame(env, *this).debugThrow<EvalError>(); } } vAttrs = j->value; @@ -1518,10 +1370,13 @@ void ExprSelect::eval(EvalState & state, Env & env, Value & v) state.forceValue(*vAttrs, (pos2 ? pos2 : this->pos ) ); } catch (Error & e) { - auto pos2r = state.positions[pos2]; - if (pos2 && pos2r.file != state.derivationNixPath) - state.addErrorTrace(e, pos2, "while evaluating the attribute '%1%'", - showAttrPath(state, env, attrPath)); + if (pos2) { + auto pos2r = state.positions[pos2]; + auto origin = std::get_if<Path>(&pos2r.origin); + if (!(origin && *origin == state.derivationNixPath)) + state.addErrorTrace(e, pos2, "while evaluating the attribute '%1%'", + showAttrPath(state, env, attrPath)); + } throw; } @@ -1599,7 +1454,12 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & if (!lambda.hasFormals()) env2.values[displ++] = args[0]; else { - forceAttrs(*args[0], pos); + try { + forceAttrs(*args[0], lambda.pos, "while evaluating the value passed for the lambda argument"); + } catch (Error & e) { + if (pos) e.addTrace(positions[pos], "from call site"); + throw; + } if (lambda.arg) env2.values[displ++] = args[0]; @@ -1611,8 +1471,15 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & for (auto & i : lambda.formals->formals) { auto j = args[0]->attrs->get(i.name); if (!j) { - if (!i.def) throwTypeError(pos, "%1% called without required argument '%2%'", - lambda, i.name, *fun.lambda.env, lambda); + if (!i.def) { + error("function '%1%' called without required argument '%2%'", + (lambda.name ? std::string(symbols[lambda.name]) : "anonymous lambda"), + symbols[i.name]) + .atPos(lambda.pos) + .withTrace(pos, "from call site") + .withFrame(*fun.lambda.env, lambda) + .debugThrow<TypeError>(); + } env2.values[displ++] = i.def->maybeThunk(*this, env2); } else { attrsUsed++; @@ -1630,11 +1497,15 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & std::set<std::string> formalNames; for (auto & formal : lambda.formals->formals) formalNames.insert(symbols[formal.name]); - throwTypeError( - pos, - Suggestions::bestMatches(formalNames, symbols[i.name]), - "%1% called with unexpected argument '%2%'", - lambda, i.name, *fun.lambda.env, lambda); + auto suggestions = Suggestions::bestMatches(formalNames, symbols[i.name]); + error("function '%1%' called with unexpected argument '%2%'", + (lambda.name ? std::string(symbols[lambda.name]) : "anonymous lambda"), + symbols[i.name]) + .atPos(lambda.pos) + .withTrace(pos, "from call site") + .withSuggestions(suggestions) + .withFrame(*fun.lambda.env, lambda) + .debugThrow<TypeError>(); } abort(); // can't happen } @@ -1648,7 +1519,7 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & auto dts = debugRepl ? makeDebugTraceStacker( *this, *lambda.body, env2, positions[lambda.pos], - "while evaluating %s", + "while calling %s", lambda.name ? concatStrings("'", symbols[lambda.name], "'") : "anonymous lambda") @@ -1657,11 +1528,15 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & lambda.body->eval(*this, env2, vCur); } catch (Error & e) { if (loggerSettings.showTrace.get()) { - addErrorTrace(e, lambda.pos, "while evaluating %s", - (lambda.name - ? concatStrings("'", symbols[lambda.name], "'") - : "anonymous lambda")); - addErrorTrace(e, pos, "from call site%s", ""); + addErrorTrace( + e, + lambda.pos, + "while calling %s", + lambda.name + ? concatStrings("'", symbols[lambda.name], "'") + : "anonymous lambda", + true); + if (pos) addErrorTrace(e, pos, "from call site%s", "", true); } throw; } @@ -1680,9 +1555,17 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & return; } else { /* We have all the arguments, so call the primop. */ + auto name = vCur.primOp->name; + nrPrimOpCalls++; - if (countCalls) primOpCalls[vCur.primOp->name]++; - vCur.primOp->fun(*this, pos, args, vCur); + if (countCalls) primOpCalls[name]++; + + try { + vCur.primOp->fun(*this, noPos, args, vCur); + } catch (Error & e) { + addErrorTrace(e, pos, "while calling the '%1%' builtin", name); + throw; + } nrArgs -= argsLeft; args += argsLeft; @@ -1717,9 +1600,20 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & for (size_t i = 0; i < argsLeft; ++i) vArgs[argsDone + i] = args[i]; + auto name = primOp->primOp->name; nrPrimOpCalls++; - if (countCalls) primOpCalls[primOp->primOp->name]++; - primOp->primOp->fun(*this, pos, vArgs, vCur); + if (countCalls) primOpCalls[name]++; + + try { + // TODO: + // 1. Unify this and above code. Heavily redundant. + // 2. Create a fake env (arg1, arg2, etc.) and a fake expr (arg1: arg2: etc: builtins.name arg1 arg2 etc) + // so the debugger allows to inspect the wrong parameters passed to the builtin. + primOp->primOp->fun(*this, noPos, vArgs, vCur); + } catch (Error & e) { + addErrorTrace(e, pos, "while calling the '%1%' builtin", name); + throw; + } nrArgs -= argsLeft; args += argsLeft; @@ -1732,14 +1626,18 @@ void EvalState::callFunction(Value & fun, size_t nrArgs, Value * * args, Value & heap-allocate a copy and use that instead. */ Value * args2[] = {allocValue(), args[0]}; *args2[0] = vCur; - /* !!! Should we use the attr pos here? */ - callFunction(*functor->value, 2, args2, vCur, pos); + try { + callFunction(*functor->value, 2, args2, vCur, functor->pos); + } catch (Error & e) { + e.addTrace(positions[pos], "while calling a functor (an attribute set with a '__functor' attribute)"); + throw; + } nrArgs--; args++; } else - throwTypeError(pos, "attempt to call something which is not a function but %1%", vCur); + error("attempt to call something which is not a function but %1%", showType(vCur)).atPos(pos).debugThrow<TypeError>(); } vRes = vCur; @@ -1803,13 +1701,12 @@ void EvalState::autoCallFunction(Bindings & args, Value & fun, Value & res) if (j != args.end()) { attrs.insert(*j); } else if (!i.def) { - throwMissingArgumentError(i.pos, R"(cannot evaluate a function that has an argument without a value ('%1%') - + error(R"(cannot evaluate a function that has an argument without a value ('%1%') Nix attempted to evaluate a function as a top level expression; in this case it must have its arguments supplied either by default values, or passed explicitly with '--arg' or '--argstr'. See -https://nixos.org/manual/nix/stable/expressions/language-constructs.html#functions.)", symbols[i.name], - *fun.lambda.env, *fun.lambda.fun); +https://nixos.org/manual/nix/stable/language/constructs.html#functions.)", symbols[i.name]) + .atPos(i.pos).withFrame(*fun.lambda.env, *fun.lambda.fun).debugThrow<MissingArgumentError>(); } } } @@ -1832,16 +1729,17 @@ void ExprWith::eval(EvalState & state, Env & env, Value & v) void ExprIf::eval(EvalState & state, Env & env, Value & v) { - (state.evalBool(env, cond, pos) ? then : else_)->eval(state, env, v); + // We cheat in the parser, and pass the position of the condition as the position of the if itself. + (state.evalBool(env, cond, pos, "while evaluating a branch condition") ? then : else_)->eval(state, env, v); } void ExprAssert::eval(EvalState & state, Env & env, Value & v) { - if (!state.evalBool(env, cond, pos)) { + if (!state.evalBool(env, cond, pos, "in the condition of the assert statement")) { std::ostringstream out; cond->show(state.symbols, out); - state.throwAssertionError(pos, "assertion '%1%' failed", out.str(), env, *this); + state.error("assertion '%1%' failed", out.str()).atPos(pos).withFrame(env, *this).debugThrow<AssertionError>(); } body->eval(state, env, v); } @@ -1849,7 +1747,7 @@ void ExprAssert::eval(EvalState & state, Env & env, Value & v) void ExprOpNot::eval(EvalState & state, Env & env, Value & v) { - v.mkBool(!state.evalBool(env, e)); + v.mkBool(!state.evalBool(env, e, noPos, "in the argument of the not operator")); // XXX: FIXME: ! } @@ -1857,7 +1755,7 @@ void ExprOpEq::eval(EvalState & state, Env & env, Value & v) { Value v1; e1->eval(state, env, v1); Value v2; e2->eval(state, env, v2); - v.mkBool(state.eqValues(v1, v2)); + v.mkBool(state.eqValues(v1, v2, pos, "while testing two values for equality")); } @@ -1865,33 +1763,33 @@ void ExprOpNEq::eval(EvalState & state, Env & env, Value & v) { Value v1; e1->eval(state, env, v1); Value v2; e2->eval(state, env, v2); - v.mkBool(!state.eqValues(v1, v2)); + v.mkBool(!state.eqValues(v1, v2, pos, "while testing two values for inequality")); } void ExprOpAnd::eval(EvalState & state, Env & env, Value & v) { - v.mkBool(state.evalBool(env, e1, pos) && state.evalBool(env, e2, pos)); + v.mkBool(state.evalBool(env, e1, pos, "in the left operand of the AND (&&) operator") && state.evalBool(env, e2, pos, "in the right operand of the AND (&&) operator")); } void ExprOpOr::eval(EvalState & state, Env & env, Value & v) { - v.mkBool(state.evalBool(env, e1, pos) || state.evalBool(env, e2, pos)); + v.mkBool(state.evalBool(env, e1, pos, "in the left operand of the OR (||) operator") || state.evalBool(env, e2, pos, "in the right operand of the OR (||) operator")); } void ExprOpImpl::eval(EvalState & state, Env & env, Value & v) { - v.mkBool(!state.evalBool(env, e1, pos) || state.evalBool(env, e2, pos)); + v.mkBool(!state.evalBool(env, e1, pos, "in the left operand of the IMPL (->) operator") || state.evalBool(env, e2, pos, "in the right operand of the IMPL (->) operator")); } void ExprOpUpdate::eval(EvalState & state, Env & env, Value & v) { Value v1, v2; - state.evalAttrs(env, e1, v1); - state.evalAttrs(env, e2, v2); + state.evalAttrs(env, e1, v1, pos, "in the left operand of the update (//) operator"); + state.evalAttrs(env, e2, v2, pos, "in the right operand of the update (//) operator"); state.nrOpUpdates++; @@ -1930,18 +1828,18 @@ void ExprOpConcatLists::eval(EvalState & state, Env & env, Value & v) Value v1; e1->eval(state, env, v1); Value v2; e2->eval(state, env, v2); Value * lists[2] = { &v1, &v2 }; - state.concatLists(v, 2, lists, pos); + state.concatLists(v, 2, lists, pos, "while evaluating one of the elements to concatenate"); } -void EvalState::concatLists(Value & v, size_t nrLists, Value * * lists, const PosIdx pos) +void EvalState::concatLists(Value & v, size_t nrLists, Value * * lists, const PosIdx pos, std::string_view errorCtx) { nrListConcats++; Value * nonEmpty = 0; size_t len = 0; for (size_t n = 0; n < nrLists; ++n) { - forceList(*lists[n], pos); + forceList(*lists[n], pos, errorCtx); auto l = lists[n]->listSize(); len += l; if (l) nonEmpty = lists[n]; @@ -2018,20 +1916,22 @@ void ExprConcatStrings::eval(EvalState & state, Env & env, Value & v) nf = n; nf += vTmp.fpoint; } else - state.throwEvalError(i_pos, "cannot add %1% to an integer", showType(vTmp), env, *this); + state.error("cannot add %1% to an integer", showType(vTmp)).atPos(i_pos).withFrame(env, *this).debugThrow<EvalError>(); } else if (firstType == nFloat) { if (vTmp.type() == nInt) { nf += vTmp.integer; } else if (vTmp.type() == nFloat) { nf += vTmp.fpoint; } else - state.throwEvalError(i_pos, "cannot add %1% to a float", showType(vTmp), env, *this); + state.error("cannot add %1% to a float", showType(vTmp)).atPos(i_pos).withFrame(env, *this).debugThrow<EvalError>(); } else { if (s.empty()) s.reserve(es->size()); /* skip canonization of first path, which would only be not canonized in the first place if it's coming from a ./${foo} type path */ - auto part = state.coerceToString(i_pos, vTmp, context, false, firstType == nString, !first); + auto part = state.coerceToString(i_pos, vTmp, context, + "while evaluating a path segment", + false, firstType == nString, !first); sSize += part->size(); s.emplace_back(std::move(part)); } @@ -2045,7 +1945,7 @@ void ExprConcatStrings::eval(EvalState & state, Env & env, Value & v) v.mkFloat(nf); else if (firstType == nPath) { if (!context.empty()) - state.throwEvalError(pos, "a string that refers to a store path cannot be appended to a path", env, *this); + state.error("a string that refers to a store path cannot be appended to a path").atPos(pos).withFrame(env, *this).debugThrow<EvalError>(); v.mkPath(canonPath(str())); } else v.mkStringMove(c_str(), context); @@ -2095,33 +1995,47 @@ void EvalState::forceValueDeep(Value & v) } -NixInt EvalState::forceInt(Value & v, const PosIdx pos) +NixInt EvalState::forceInt(Value & v, const PosIdx pos, std::string_view errorCtx) { - forceValue(v, pos); - if (v.type() != nInt) - throwTypeError(pos, "value is %1% while an integer was expected", v); - - return v.integer; + try { + forceValue(v, pos); + if (v.type() != nInt) + error("value is %1% while an integer was expected", showType(v)).debugThrow<TypeError>(); + return v.integer; + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; + } } -NixFloat EvalState::forceFloat(Value & v, const PosIdx pos) +NixFloat EvalState::forceFloat(Value & v, const PosIdx pos, std::string_view errorCtx) { - forceValue(v, pos); - if (v.type() == nInt) - return v.integer; - else if (v.type() != nFloat) - throwTypeError(pos, "value is %1% while a float was expected", v); - return v.fpoint; + try { + forceValue(v, pos); + if (v.type() == nInt) + return v.integer; + else if (v.type() != nFloat) + error("value is %1% while a float was expected", showType(v)).debugThrow<TypeError>(); + return v.fpoint; + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; + } } -bool EvalState::forceBool(Value & v, const PosIdx pos) +bool EvalState::forceBool(Value & v, const PosIdx pos, std::string_view errorCtx) { - forceValue(v, pos); - if (v.type() != nBool) - throwTypeError(pos, "value is %1% while a Boolean was expected", v); - return v.boolean; + try { + forceValue(v, pos); + if (v.type() != nBool) + error("value is %1% while a Boolean was expected", showType(v)).debugThrow<TypeError>(); + return v.boolean; + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; + } } @@ -2131,42 +2045,30 @@ bool EvalState::isFunctor(Value & fun) } -void EvalState::forceFunction(Value & v, const PosIdx pos) -{ - forceValue(v, pos); - if (v.type() != nFunction && !isFunctor(v)) - throwTypeError(pos, "value is %1% while a function was expected", v); -} - - -std::string_view EvalState::forceString(Value & v, const PosIdx pos) +void EvalState::forceFunction(Value & v, const PosIdx pos, std::string_view errorCtx) { - forceValue(v, pos); - if (v.type() != nString) { - throwTypeError(pos, "value is %1% while a string was expected", v); + try { + forceValue(v, pos); + if (v.type() != nFunction && !isFunctor(v)) + error("value is %1% while a function was expected", showType(v)).debugThrow<TypeError>(); + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; } - return v.string.s; } -/* Decode a context string ‘!<name>!<path>’ into a pair <path, - name>. */ -NixStringContextElem decodeContext(const Store & store, std::string_view s) +std::string_view EvalState::forceString(Value & v, const PosIdx pos, std::string_view errorCtx) { - if (s.at(0) == '!') { - size_t index = s.find("!", 1); - return { - store.parseStorePath(s.substr(index + 1)), - std::string(s.substr(1, index - 1)), - }; - } else - return { - store.parseStorePath( - s.at(0) == '/' - ? s - : s.substr(1)), - "", - }; + try { + forceValue(v, pos); + if (v.type() != nString) + error("value is %1% while a string was expected", showType(v)).debugThrow<TypeError>(); + return v.string.s; + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; + } } @@ -2184,29 +2086,24 @@ NixStringContext Value::getContext(const Store & store) assert(internalType == tString); if (string.context) for (const char * * p = string.context; *p; ++p) - res.push_back(decodeContext(store, *p)); + res.push_back(NixStringContextElem::parse(store, *p)); return res; } -std::string_view EvalState::forceString(Value & v, PathSet & context, const PosIdx pos) +std::string_view EvalState::forceString(Value & v, PathSet & context, const PosIdx pos, std::string_view errorCtx) { - auto s = forceString(v, pos); + auto s = forceString(v, pos, errorCtx); copyContext(v, context); return s; } -std::string_view EvalState::forceStringNoCtx(Value & v, const PosIdx pos) +std::string_view EvalState::forceStringNoCtx(Value & v, const PosIdx pos, std::string_view errorCtx) { - auto s = forceString(v, pos); + auto s = forceString(v, pos, errorCtx); if (v.string.context) { - if (pos) - throwEvalError(pos, "the string '%1%' is not allowed to refer to a store path (such as '%2%')", - v.string.s, v.string.context[0]); - else - throwEvalError("the string '%1%' is not allowed to refer to a store path (such as '%2%')", - v.string.s, v.string.context[0]); + error("the string '%1%' is not allowed to refer to a store path (such as '%2%')", v.string.s, v.string.context[0]).withTrace(pos, errorCtx).debugThrow<EvalError>(); } return s; } @@ -2230,14 +2127,16 @@ std::optional<std::string> EvalState::tryAttrsToString(const PosIdx pos, Value & if (i != v.attrs->end()) { Value v1; callFunction(*i->value, v, v1, pos); - return coerceToString(pos, v1, context, coerceMore, copyToStore).toOwned(); + return coerceToString(pos, v1, context, + "while evaluating the result of the `__toString` attribute", + coerceMore, copyToStore).toOwned(); } return {}; } -BackedStringView EvalState::coerceToString(const PosIdx pos, Value & v, PathSet & context, - bool coerceMore, bool copyToStore, bool canonicalizePath) +BackedStringView EvalState::coerceToString(const PosIdx pos, Value &v, PathSet &context, + std::string_view errorCtx, bool coerceMore, bool copyToStore, bool canonicalizePath) { forceValue(v, pos); @@ -2251,7 +2150,7 @@ BackedStringView EvalState::coerceToString(const PosIdx pos, Value & v, PathSet if (canonicalizePath) path = canonPath(*path); if (copyToStore) - path = copyPathToStore(context, std::move(path).toOwned()); + path = store->printStorePath(copyPathToStore(context, std::move(path).toOwned())); return path; } @@ -2260,13 +2159,23 @@ BackedStringView EvalState::coerceToString(const PosIdx pos, Value & v, PathSet if (maybeString) return std::move(*maybeString); auto i = v.attrs->find(sOutPath); - if (i == v.attrs->end()) - throwTypeError(pos, "cannot coerce a set to a string"); - return coerceToString(pos, *i->value, context, coerceMore, copyToStore); + if (i == v.attrs->end()) { + error("cannot coerce %1% to a string", showType(v)) + .withTrace(pos, errorCtx) + .debugThrow<TypeError>(); + } + return coerceToString(pos, *i->value, context, errorCtx, + coerceMore, copyToStore, canonicalizePath); } - if (v.type() == nExternal) - return v.external->coerceToString(positions[pos], context, coerceMore, copyToStore); + if (v.type() == nExternal) { + try { + return v.external->coerceToString(positions[pos], context, coerceMore, copyToStore); + } catch (Error & e) { + e.addTrace(nullptr, errorCtx); + throw; + } + } if (coerceMore) { /* Note that `false' is represented as an empty string for @@ -2280,7 +2189,14 @@ BackedStringView EvalState::coerceToString(const PosIdx pos, Value & v, PathSet if (v.isList()) { std::string result; for (auto [n, v2] : enumerate(v.listItems())) { - result += *coerceToString(pos, *v2, context, coerceMore, copyToStore); + try { + result += *coerceToString(noPos, *v2, context, + "while evaluating one element of the list", + coerceMore, copyToStore, canonicalizePath); + } catch (Error & e) { + e.addTrace(positions[pos], errorCtx); + throw; + } if (n < v.listSize() - 1 /* !!! not quite correct */ && (!v2->isList() || v2->listSize() != 0)) @@ -2290,56 +2206,55 @@ BackedStringView EvalState::coerceToString(const PosIdx pos, Value & v, PathSet } } - throwTypeError(pos, "cannot coerce %1% to a string", v); + error("cannot coerce %1% to a string", showType(v)) + .withTrace(pos, errorCtx) + .debugThrow<TypeError>(); } -std::string EvalState::copyPathToStore(PathSet & context, const Path & path) +StorePath EvalState::copyPathToStore(PathSet & context, const Path & path) { if (nix::isDerivation(path)) - throwEvalError("file names are not allowed to end in '%1%'", drvExtension); - - Path dstPath; - auto i = srcToStore.find(path); - if (i != srcToStore.end()) - dstPath = store->printStorePath(i->second); - else { - auto p = settings.readOnlyMode + error("file names are not allowed to end in '%1%'", drvExtension).debugThrow<EvalError>(); + + auto dstPath = [&]() -> StorePath + { + auto i = srcToStore.find(path); + if (i != srcToStore.end()) return i->second; + + auto dstPath = settings.readOnlyMode ? store->computeStorePathForPath(std::string(baseNameOf(path)), checkSourcePath(path)).first : store->addToStore(std::string(baseNameOf(path)), checkSourcePath(path), FileIngestionMethod::Recursive, htSHA256, defaultPathFilter, repair); - dstPath = store->printStorePath(p); - allowPath(p); - srcToStore.insert_or_assign(path, std::move(p)); - printMsg(lvlChatty, "copied source '%1%' -> '%2%'", path, dstPath); - } + allowPath(dstPath); + srcToStore.insert_or_assign(path, dstPath); + printMsg(lvlChatty, "copied source '%1%' -> '%2%'", path, store->printStorePath(dstPath)); + return dstPath; + }(); - context.insert(dstPath); + context.insert(store->printStorePath(dstPath)); return dstPath; } -Path EvalState::coerceToPath(const PosIdx pos, Value & v, PathSet & context) +Path EvalState::coerceToPath(const PosIdx pos, Value & v, PathSet & context, std::string_view errorCtx) { - auto path = coerceToString(pos, v, context, false, false).toOwned(); + auto path = coerceToString(pos, v, context, errorCtx, false, false, true).toOwned(); if (path == "" || path[0] != '/') - throwEvalError(pos, "string '%1%' doesn't represent an absolute path", path); + error("string '%1%' doesn't represent an absolute path", path).withTrace(pos, errorCtx).debugThrow<EvalError>(); return path; } -StorePath EvalState::coerceToStorePath(const PosIdx pos, Value & v, PathSet & context) +StorePath EvalState::coerceToStorePath(const PosIdx pos, Value & v, PathSet & context, std::string_view errorCtx) { - auto path = coerceToString(pos, v, context, false, false).toOwned(); + auto path = coerceToString(pos, v, context, errorCtx, false, false, true).toOwned(); if (auto storePath = store->maybeParseStorePath(path)) return *storePath; - throw EvalError({ - .msg = hintfmt("path '%1%' is not in the Nix store", path), - .errPos = positions[pos] - }); + error("path '%1%' is not in the Nix store", path).withTrace(pos, errorCtx).debugThrow<EvalError>(); } -bool EvalState::eqValues(Value & v1, Value & v2) +bool EvalState::eqValues(Value & v1, Value & v2, const PosIdx pos, std::string_view errorCtx) { forceValue(v1, noPos); forceValue(v2, noPos); @@ -2359,7 +2274,6 @@ bool EvalState::eqValues(Value & v1, Value & v2) if (v1.type() != v2.type()) return false; switch (v1.type()) { - case nInt: return v1.integer == v2.integer; @@ -2378,7 +2292,7 @@ bool EvalState::eqValues(Value & v1, Value & v2) case nList: if (v1.listSize() != v2.listSize()) return false; for (size_t n = 0; n < v1.listSize(); ++n) - if (!eqValues(*v1.listElems()[n], *v2.listElems()[n])) return false; + if (!eqValues(*v1.listElems()[n], *v2.listElems()[n], pos, errorCtx)) return false; return true; case nAttrs: { @@ -2388,7 +2302,7 @@ bool EvalState::eqValues(Value & v1, Value & v2) Bindings::iterator i = v1.attrs->find(sOutPath); Bindings::iterator j = v2.attrs->find(sOutPath); if (i != v1.attrs->end() && j != v2.attrs->end()) - return eqValues(*i->value, *j->value); + return eqValues(*i->value, *j->value, pos, errorCtx); } if (v1.attrs->size() != v2.attrs->size()) return false; @@ -2396,7 +2310,7 @@ bool EvalState::eqValues(Value & v1, Value & v2) /* Otherwise, compare the attributes one by one. */ Bindings::iterator i, j; for (i = v1.attrs->begin(), j = v2.attrs->begin(); i != v1.attrs->end(); ++i, ++j) - if (i->name != j->name || !eqValues(*i->value, *j->value)) + if (i->name != j->name || !eqValues(*i->value, *j->value, pos, errorCtx)) return false; return true; @@ -2413,9 +2327,7 @@ bool EvalState::eqValues(Value & v1, Value & v2) return v1.fpoint == v2.fpoint; default: - throwEvalError("cannot compare %1% with %2%", - showType(v1), - showType(v2)); + error("cannot compare %1% with %2%", showType(v1), showType(v2)).withTrace(pos, errorCtx).debugThrow<EvalError>(); } } @@ -2441,97 +2353,99 @@ void EvalState::printStats() std::fstream fs; if (outPath != "-") fs.open(outPath, std::fstream::out); - JSONObject topObj(outPath == "-" ? std::cerr : fs, true); - topObj.attr("cpuTime",cpuTime); - { - auto envs = topObj.object("envs"); - envs.attr("number", nrEnvs); - envs.attr("elements", nrValuesInEnvs); - envs.attr("bytes", bEnvs); - } - { - auto lists = topObj.object("list"); - lists.attr("elements", nrListElems); - lists.attr("bytes", bLists); - lists.attr("concats", nrListConcats); - } - { - auto values = topObj.object("values"); - values.attr("number", nrValues); - values.attr("bytes", bValues); - } - { - auto syms = topObj.object("symbols"); - syms.attr("number", symbols.size()); - syms.attr("bytes", symbols.totalSize()); - } - { - auto sets = topObj.object("sets"); - sets.attr("number", nrAttrsets); - sets.attr("bytes", bAttrsets); - sets.attr("elements", nrAttrsInAttrsets); - } - { - auto sizes = topObj.object("sizes"); - sizes.attr("Env", sizeof(Env)); - sizes.attr("Value", sizeof(Value)); - sizes.attr("Bindings", sizeof(Bindings)); - sizes.attr("Attr", sizeof(Attr)); - } - topObj.attr("nrOpUpdates", nrOpUpdates); - topObj.attr("nrOpUpdateValuesCopied", nrOpUpdateValuesCopied); - topObj.attr("nrThunks", nrThunks); - topObj.attr("nrAvoided", nrAvoided); - topObj.attr("nrLookups", nrLookups); - topObj.attr("nrPrimOpCalls", nrPrimOpCalls); - topObj.attr("nrFunctionCalls", nrFunctionCalls); + json topObj = json::object(); + topObj["cpuTime"] = cpuTime; + topObj["envs"] = { + {"number", nrEnvs}, + {"elements", nrValuesInEnvs}, + {"bytes", bEnvs}, + }; + topObj["list"] = { + {"elements", nrListElems}, + {"bytes", bLists}, + {"concats", nrListConcats}, + }; + topObj["values"] = { + {"number", nrValues}, + {"bytes", bValues}, + }; + topObj["symbols"] = { + {"number", symbols.size()}, + {"bytes", symbols.totalSize()}, + }; + topObj["sets"] = { + {"number", nrAttrsets}, + {"bytes", bAttrsets}, + {"elements", nrAttrsInAttrsets}, + }; + topObj["sizes"] = { + {"Env", sizeof(Env)}, + {"Value", sizeof(Value)}, + {"Bindings", sizeof(Bindings)}, + {"Attr", sizeof(Attr)}, + }; + topObj["nrOpUpdates"] = nrOpUpdates; + topObj["nrOpUpdateValuesCopied"] = nrOpUpdateValuesCopied; + topObj["nrThunks"] = nrThunks; + topObj["nrAvoided"] = nrAvoided; + topObj["nrLookups"] = nrLookups; + topObj["nrPrimOpCalls"] = nrPrimOpCalls; + topObj["nrFunctionCalls"] = nrFunctionCalls; #if HAVE_BOEHMGC - { - auto gc = topObj.object("gc"); - gc.attr("heapSize", heapSize); - gc.attr("totalBytes", totalBytes); - } + topObj["gc"] = { + {"heapSize", heapSize}, + {"totalBytes", totalBytes}, + }; #endif if (countCalls) { + topObj["primops"] = primOpCalls; { - auto obj = topObj.object("primops"); - for (auto & i : primOpCalls) - obj.attr(i.first, i.second); - } - { - auto list = topObj.list("functions"); + auto& list = topObj["functions"]; + list = json::array(); for (auto & [fun, count] : functionCalls) { - auto obj = list.object(); + json obj = json::object(); if (fun->name) - obj.attr("name", (std::string_view) symbols[fun->name]); + obj["name"] = (std::string_view) symbols[fun->name]; else - obj.attr("name", nullptr); + obj["name"] = nullptr; if (auto pos = positions[fun->pos]) { - obj.attr("file", (std::string_view) pos.file); - obj.attr("line", pos.line); - obj.attr("column", pos.column); + if (auto path = std::get_if<Path>(&pos.origin)) + obj["file"] = *path; + obj["line"] = pos.line; + obj["column"] = pos.column; } - obj.attr("count", count); + obj["count"] = count; + list.push_back(obj); } } { - auto list = topObj.list("attributes"); + auto list = topObj["attributes"]; + list = json::array(); for (auto & i : attrSelects) { - auto obj = list.object(); + json obj = json::object(); if (auto pos = positions[i.first]) { - obj.attr("file", (const std::string &) pos.file); - obj.attr("line", pos.line); - obj.attr("column", pos.column); + if (auto path = std::get_if<Path>(&pos.origin)) + obj["file"] = *path; + obj["line"] = pos.line; + obj["column"] = pos.column; } - obj.attr("count", i.second); + obj["count"] = i.second; + list.push_back(obj); } } } if (getEnv("NIX_SHOW_SYMBOLS").value_or("0") != "0") { - auto list = topObj.list("symbols"); - symbols.dump([&](const std::string & s) { list.elem(s); }); + // XXX: overrides earlier assignment + topObj["symbols"] = json::array(); + auto &list = topObj["symbols"]; + symbols.dump([&](const std::string & s) { list.emplace_back(s); }); + } + if (outPath == "-") { + std::cerr << topObj.dump(2) << std::endl; + } else { + fs << topObj.dump(2) << std::endl; } } } @@ -2540,8 +2454,7 @@ void EvalState::printStats() std::string ExternalValueBase::coerceToString(const Pos & pos, PathSet & context, bool copyMore, bool copyToStore) const { throw TypeError({ - .msg = hintfmt("cannot coerce %1% to a string", showType()), - .errPos = pos + .msg = hintfmt("cannot coerce %1% to a string", showType()) }); } @@ -2577,7 +2490,7 @@ Strings EvalSettings::getDefaultNixPath() }; if (!evalSettings.restrictEval && !evalSettings.pureEval) { - add(getHome() + "/.nix-defexpr/channels"); + add(settings.useXDGBaseDirectories ? getStateDir() + "/nix/defexpr/channels" : getHome() + "/.nix-defexpr/channels"); add(settings.nixStateDir + "/profiles/per-user/root/channels/nixpkgs", "nixpkgs"); add(settings.nixStateDir + "/profiles/per-user/root/channels"); } @@ -2585,6 +2498,23 @@ Strings EvalSettings::getDefaultNixPath() return res; } +bool EvalSettings::isPseudoUrl(std::string_view s) +{ + if (s.compare(0, 8, "channel:") == 0) return true; + size_t pos = s.find("://"); + if (pos == std::string::npos) return false; + std::string scheme(s, 0, pos); + return scheme == "http" || scheme == "https" || scheme == "file" || scheme == "channel" || scheme == "git" || scheme == "s3" || scheme == "ssh"; +} + +std::string EvalSettings::resolvePseudoUrl(std::string_view url) +{ + if (hasPrefix(url, "channel:")) + return "https://nixos.org/channels/" + std::string(url.substr(8)) + "/nixexprs.tar.xz"; + else + return std::string(url); +} + EvalSettings evalSettings; static GlobalConfig::Register rEvalSettings(&evalSettings); diff --git a/src/libexpr/eval.hh b/src/libexpr/eval.hh index f07f15d43..e4d5906bd 100644 --- a/src/libexpr/eval.hh +++ b/src/libexpr/eval.hh @@ -60,7 +60,6 @@ void copyContext(const Value & v, PathSet & context); typedef std::map<Path, StorePath> SrcToStore; -std::ostream & printValue(const EvalState & state, std::ostream & str, const Value & v); std::string printValue(const EvalState & state, const Value & v); std::ostream & operator << (std::ostream & os, const ValueType t); @@ -78,7 +77,7 @@ struct RegexCache; std::shared_ptr<RegexCache> makeRegexCache(); struct DebugTrace { - std::optional<ErrPos> pos; + std::shared_ptr<AbstractPos> pos; const Expr & expr; const Env & env; hintformat hint; @@ -87,6 +86,43 @@ struct DebugTrace { void debugError(Error * e, Env & env, Expr & expr); +class ErrorBuilder +{ + private: + EvalState & state; + ErrorInfo info; + + ErrorBuilder(EvalState & s, ErrorInfo && i): state(s), info(i) { } + + public: + template<typename... Args> + [[nodiscard, gnu::noinline]] + static ErrorBuilder * create(EvalState & s, const Args & ... args) + { + return new ErrorBuilder(s, ErrorInfo { .msg = hintfmt(args...) }); + } + + [[nodiscard, gnu::noinline]] + ErrorBuilder & atPos(PosIdx pos); + + [[nodiscard, gnu::noinline]] + ErrorBuilder & withTrace(PosIdx pos, const std::string_view text); + + [[nodiscard, gnu::noinline]] + ErrorBuilder & withFrameTrace(PosIdx pos, const std::string_view text); + + [[nodiscard, gnu::noinline]] + ErrorBuilder & withSuggestions(Suggestions & s); + + [[nodiscard, gnu::noinline]] + ErrorBuilder & withFrame(const Env & e, const Expr & ex); + + template<class ErrorType> + [[gnu::noinline, gnu::noreturn]] + void debugThrow(); +}; + + class EvalState : public std::enable_shared_from_this<EvalState> { public: @@ -146,29 +182,38 @@ public: template<class E> [[gnu::noinline, gnu::noreturn]] - void debugThrow(E && error, const Env & env, const Expr & expr) + void debugThrowLastTrace(E && error) { - if (debugRepl) - runDebugRepl(&error, env, expr); - - throw std::move(error); + debugThrow(error, nullptr, nullptr); } template<class E> [[gnu::noinline, gnu::noreturn]] - void debugThrowLastTrace(E && e) + void debugThrow(E && error, const Env * env, const Expr * expr) { - // Call this in the situation where Expr and Env are inaccessible. - // The debugger will start in the last context that's in the - // DebugTrace stack. - if (debugRepl && !debugTraces.empty()) { - const DebugTrace & last = debugTraces.front(); - runDebugRepl(&e, last.env, last.expr); + if (debugRepl && ((env && expr) || !debugTraces.empty())) { + if (!env || !expr) { + const DebugTrace & last = debugTraces.front(); + env = &last.env; + expr = &last.expr; + } + runDebugRepl(&error, *env, *expr); } - throw std::move(e); + throw std::move(error); } + // This is dangerous, but gets in line with the idea that error creation and + // throwing should not allocate on the stack of hot functions. + // as long as errors are immediately thrown, it works. + ErrorBuilder * errorBuilder; + + template<typename... Args> + [[nodiscard, gnu::noinline]] + ErrorBuilder & error(const Args & ... args) { + errorBuilder = ErrorBuilder::create(*this, args...); + return *errorBuilder; + } private: SrcToStore srcToStore; @@ -283,8 +328,8 @@ public: /* Evaluation the expression, then verify that it has the expected type. */ inline bool evalBool(Env & env, Expr * e); - inline bool evalBool(Env & env, Expr * e, const PosIdx pos); - inline void evalAttrs(Env & env, Expr * e, Value & v); + inline bool evalBool(Env & env, Expr * e, const PosIdx pos, std::string_view errorCtx); + inline void evalAttrs(Env & env, Expr * e, Value & v, const PosIdx pos, std::string_view errorCtx); /* If `v' is a thunk, enter it and overwrite `v' with the result of the evaluation of the thunk. If `v' is a delayed function @@ -300,89 +345,25 @@ public: void forceValueDeep(Value & v); /* Force `v', and then verify that it has the expected type. */ - NixInt forceInt(Value & v, const PosIdx pos); - NixFloat forceFloat(Value & v, const PosIdx pos); - bool forceBool(Value & v, const PosIdx pos); + NixInt forceInt(Value & v, const PosIdx pos, std::string_view errorCtx); + NixFloat forceFloat(Value & v, const PosIdx pos, std::string_view errorCtx); + bool forceBool(Value & v, const PosIdx pos, std::string_view errorCtx); - void forceAttrs(Value & v, const PosIdx pos); + void forceAttrs(Value & v, const PosIdx pos, std::string_view errorCtx); template <typename Callable> - inline void forceAttrs(Value & v, Callable getPos); + inline void forceAttrs(Value & v, Callable getPos, std::string_view errorCtx); - inline void forceList(Value & v, const PosIdx pos); - void forceFunction(Value & v, const PosIdx pos); // either lambda or primop - std::string_view forceString(Value & v, const PosIdx pos = noPos); - std::string_view forceString(Value & v, PathSet & context, const PosIdx pos = noPos); - std::string_view forceStringNoCtx(Value & v, const PosIdx pos = noPos); - - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx pos, const char * s); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx pos, const char * s, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const char * s, const std::string & s2); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx pos, const char * s, const std::string & s2); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const char * s, const std::string & s2, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx pos, const char * s, const std::string & s2, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const char * s, const std::string & s2, const std::string & s3, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx pos, const char * s, const std::string & s2, const std::string & s3, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx pos, const char * s, const std::string & s2, const std::string & s3); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const char * s, const std::string & s2, const std::string & s3); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx pos, const Suggestions & suggestions, const char * s, const std::string & s2, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwEvalError(const PosIdx p1, const char * s, const Symbol sym, const PosIdx p2, - Env & env, Expr & expr); - - [[gnu::noinline, gnu::noreturn]] - void throwTypeError(const PosIdx pos, const char * s, const Value & v); - [[gnu::noinline, gnu::noreturn]] - void throwTypeError(const PosIdx pos, const char * s, const Value & v, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwTypeError(const PosIdx pos, const char * s); - [[gnu::noinline, gnu::noreturn]] - void throwTypeError(const PosIdx pos, const char * s, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwTypeError(const PosIdx pos, const char * s, const ExprLambda & fun, const Symbol s2, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwTypeError(const PosIdx pos, const Suggestions & suggestions, const char * s, const ExprLambda & fun, const Symbol s2, - Env & env, Expr & expr); - [[gnu::noinline, gnu::noreturn]] - void throwTypeError(const char * s, const Value & v, - Env & env, Expr & expr); - - [[gnu::noinline, gnu::noreturn]] - void throwAssertionError(const PosIdx pos, const char * s, const std::string & s1, - Env & env, Expr & expr); - - [[gnu::noinline, gnu::noreturn]] - void throwUndefinedVarError(const PosIdx pos, const char * s, const std::string & s1, - Env & env, Expr & expr); - - [[gnu::noinline, gnu::noreturn]] - void throwMissingArgumentError(const PosIdx pos, const char * s, const std::string & s1, - Env & env, Expr & expr); + inline void forceList(Value & v, const PosIdx pos, std::string_view errorCtx); + void forceFunction(Value & v, const PosIdx pos, std::string_view errorCtx); // either lambda or primop + std::string_view forceString(Value & v, const PosIdx pos, std::string_view errorCtx); + std::string_view forceString(Value & v, PathSet & context, const PosIdx pos, std::string_view errorCtx); + std::string_view forceStringNoCtx(Value & v, const PosIdx pos, std::string_view errorCtx); [[gnu::noinline]] void addErrorTrace(Error & e, const char * s, const std::string & s2) const; [[gnu::noinline]] - void addErrorTrace(Error & e, const PosIdx pos, const char * s, const std::string & s2) const; + void addErrorTrace(Error & e, const PosIdx pos, const char * s, const std::string & s2, bool frame = false) const; public: /* Return true iff the value `v' denotes a derivation (i.e. a @@ -397,18 +378,19 @@ public: booleans and lists to a string. If `copyToStore' is set, referenced paths are copied to the Nix store as a side effect. */ BackedStringView coerceToString(const PosIdx pos, Value & v, PathSet & context, + std::string_view errorCtx, bool coerceMore = false, bool copyToStore = true, bool canonicalizePath = true); - std::string copyPathToStore(PathSet & context, const Path & path); + StorePath copyPathToStore(PathSet & context, const Path & path); /* Path coercion. Converts strings, paths and derivations to a path. The result is guaranteed to be a canonicalised, absolute path. Nothing is copied to the store. */ - Path coerceToPath(const PosIdx pos, Value & v, PathSet & context); + Path coerceToPath(const PosIdx pos, Value & v, PathSet & context, std::string_view errorCtx); /* Like coerceToPath, but the result must be a store path. */ - StorePath coerceToStorePath(const PosIdx pos, Value & v, PathSet & context); + StorePath coerceToStorePath(const PosIdx pos, Value & v, PathSet & context, std::string_view errorCtx); public: @@ -457,14 +439,18 @@ private: friend struct ExprAttrs; friend struct ExprLet; - Expr * parse(char * text, size_t length, FileOrigin origin, const PathView path, - const PathView basePath, std::shared_ptr<StaticEnv> & staticEnv); + Expr * parse( + char * text, + size_t length, + Pos::Origin origin, + Path basePath, + std::shared_ptr<StaticEnv> & staticEnv); public: /* Do a deep equality test between two values. That is, list elements and attributes are compared recursively. */ - bool eqValues(Value & v1, Value & v2); + bool eqValues(Value & v1, Value & v2, const PosIdx pos, std::string_view errorCtx); bool isFunctor(Value & fun); @@ -499,7 +485,7 @@ public: void mkThunk_(Value & v, Expr * expr); void mkPos(Value & v, PosIdx pos); - void concatLists(Value & v, size_t nrLists, Value * * lists, const PosIdx pos); + void concatLists(Value & v, size_t nrLists, Value * * lists, const PosIdx pos, std::string_view errorCtx); /* Print statistics. */ void printStats(); @@ -568,10 +554,6 @@ struct DebugTraceStacker { std::string_view showType(ValueType type); std::string showType(const Value & v); -/* Decode a context string ‘!<name>!<path>’ into a pair <path, - name>. */ -NixStringContextElem decodeContext(const Store & store, std::string_view s); - /* If `path' refers to a directory, then append "/default.nix". */ Path resolveExprPath(Path path); @@ -590,6 +572,10 @@ struct EvalSettings : Config static Strings getDefaultNixPath(); + static bool isPseudoUrl(std::string_view s); + + static std::string resolvePseudoUrl(std::string_view url); + Setting<bool> enableNativeCode{this, false, "allow-unsafe-native-code-during-evaluation", "Whether builtin functions that allow executing native code should be enabled."}; @@ -662,6 +648,13 @@ extern EvalSettings evalSettings; static const std::string corepkgsPrefix{"/__corepkgs__/"}; +template<class ErrorType> +void ErrorBuilder::debugThrow() +{ + // NOTE: We always use the -LastTrace version as we push the new trace in withFrame() + state.debugThrowLastTrace(ErrorType(info)); +} + } #include "eval-inline.hh" diff --git a/src/libexpr/fetchurl.nix b/src/libexpr/fetchurl.nix index 02531103b..9d1b61d7f 100644 --- a/src/libexpr/fetchurl.nix +++ b/src/libexpr/fetchurl.nix @@ -12,13 +12,13 @@ , executable ? false , unpack ? false , name ? baseNameOf (toString url) +, impure ? false }: -derivation { +derivation ({ builder = "builtin:fetchurl"; # New-style output content requirements. - inherit outputHashAlgo outputHash; outputHashMode = if unpack || executable then "recursive" else "flat"; inherit name url executable unpack; @@ -38,4 +38,6 @@ derivation { # To make "nix-prefetch-url" work. urls = [ url ]; -} +} // (if impure + then { __impure = true; } + else { inherit outputHashAlgo outputHash; })) diff --git a/src/libexpr/flake/call-flake.nix b/src/libexpr/flake/call-flake.nix index 932ac5e90..4beb0b0fe 100644 --- a/src/libexpr/flake/call-flake.nix +++ b/src/libexpr/flake/call-flake.nix @@ -16,7 +16,9 @@ let subdir = if key == lockFile.root then rootSubdir else node.locked.dir or ""; - flake = import (sourceInfo + (if subdir != "" then "/" else "") + subdir + "/flake.nix"); + outPath = sourceInfo + ((if subdir == "" then "" else "/") + subdir); + + flake = import (outPath + "/flake.nix"); inputs = builtins.mapAttrs (inputName: inputSpec: allNodes.${resolveInput inputSpec}) @@ -43,7 +45,21 @@ let outputs = flake.outputs (inputs // { self = result; }); - result = outputs // sourceInfo // { inherit inputs; inherit outputs; inherit sourceInfo; }; + result = + outputs + # We add the sourceInfo attribute for its metadata, as they are + # relevant metadata for the flake. However, the outPath of the + # sourceInfo does not necessarily match the outPath of the flake, + # as the flake may be in a subdirectory of a source. + # This is shadowed in the next // + // sourceInfo + // { + # This shadows the sourceInfo.outPath + inherit outPath; + + inherit inputs; inherit outputs; inherit sourceInfo; _type = "flake"; + }; + in if node.flake or true then assert builtins.isFunction flake.outputs; diff --git a/src/libexpr/flake/config.cc b/src/libexpr/flake/config.cc index 3e9d264b4..89ddbde7e 100644 --- a/src/libexpr/flake/config.cc +++ b/src/libexpr/flake/config.cc @@ -56,7 +56,7 @@ void ConfigFile::apply() auto tlname = get(trustedList, name); if (auto saved = tlname ? get(*tlname, valueS) : nullptr) { trusted = *saved; - warn("Using saved setting for '%s = %s' from ~/.local/share/nix/trusted-settings.json.", name,valueS); + printInfo("Using saved setting for '%s = %s' from ~/.local/share/nix/trusted-settings.json.", name, valueS); } else { // FIXME: filter ANSI escapes, newlines, \r, etc. if (std::tolower(logger->ask(fmt("do you want to allow configuration setting '%s' to be set to '" ANSI_RED "%s" ANSI_NORMAL "' (y/N)?", name, valueS)).value_or('n')) == 'y') { @@ -68,7 +68,7 @@ void ConfigFile::apply() } } if (!trusted) { - warn("ignoring untrusted flake configuration setting '%s'", name); + warn("ignoring untrusted flake configuration setting '%s'.\nPass '%s' to trust it", name, "--accept-flake-config"); continue; } } diff --git a/src/libexpr/flake/flake.cc b/src/libexpr/flake/flake.cc index cc9be1336..336eb274d 100644 --- a/src/libexpr/flake/flake.cc +++ b/src/libexpr/flake/flake.cc @@ -143,7 +143,7 @@ static FlakeInput parseFlakeInput(EvalState & state, } catch (Error & e) { e.addTrace( state.positions[attr.pos], - hintfmt("in flake attribute '%s'", state.symbols[attr.name])); + hintfmt("while evaluating flake attribute '%s'", state.symbols[attr.name])); throw; } } @@ -152,7 +152,7 @@ static FlakeInput parseFlakeInput(EvalState & state, try { input.ref = FlakeRef::fromAttrs(attrs); } catch (Error & e) { - e.addTrace(state.positions[pos], hintfmt("in flake input")); + e.addTrace(state.positions[pos], hintfmt("while evaluating flake input")); throw; } else { @@ -220,7 +220,7 @@ static Flake getFlake( Value vInfo; state.evalFile(flakeFile, vInfo, true); // FIXME: symlink attack - expectType(state, nAttrs, vInfo, state.positions.add({flakeFile, foFile}, 0, 0)); + expectType(state, nAttrs, vInfo, state.positions.add({flakeFile}, 1, 1)); if (auto description = vInfo.attrs->get(state.sDescription)) { expectType(state, nString, *description->value, description->pos); @@ -259,28 +259,28 @@ static Flake getFlake( if (setting.value->type() == nString) flake.config.settings.emplace( state.symbols[setting.name], - std::string(state.forceStringNoCtx(*setting.value, setting.pos))); + std::string(state.forceStringNoCtx(*setting.value, setting.pos, ""))); else if (setting.value->type() == nPath) { PathSet emptyContext = {}; flake.config.settings.emplace( state.symbols[setting.name], - state.coerceToString(setting.pos, *setting.value, emptyContext, false, true, true) .toOwned()); + state.coerceToString(setting.pos, *setting.value, emptyContext, "", false, true, true) .toOwned()); } else if (setting.value->type() == nInt) flake.config.settings.emplace( state.symbols[setting.name], - state.forceInt(*setting.value, setting.pos)); + state.forceInt(*setting.value, setting.pos, "")); else if (setting.value->type() == nBool) flake.config.settings.emplace( state.symbols[setting.name], - Explicit<bool> { state.forceBool(*setting.value, setting.pos) }); + Explicit<bool> { state.forceBool(*setting.value, setting.pos, "") }); else if (setting.value->type() == nList) { std::vector<std::string> ss; for (auto elem : setting.value->listItems()) { if (elem->type() != nString) throw TypeError("list element in flake configuration setting '%s' is %s while a string is expected", state.symbols[setting.name], showType(*setting.value)); - ss.emplace_back(state.forceStringNoCtx(*elem, setting.pos)); + ss.emplace_back(state.forceStringNoCtx(*elem, setting.pos, "")); } flake.config.settings.emplace(state.symbols[setting.name], ss); } @@ -341,7 +341,6 @@ LockedFlake lockFlake( debug("old lock file: %s", oldLockFile); - // FIXME: check whether all overrides are used. std::map<InputPath, FlakeInput> overrides; std::set<InputPath> overridesUsed, updatesUsed; @@ -354,7 +353,7 @@ LockedFlake lockFlake( std::function<void( const FlakeInputs & flakeInputs, - std::shared_ptr<Node> node, + ref<Node> node, const InputPath & inputPathPrefix, std::shared_ptr<const Node> oldNode, const InputPath & lockRootPath, @@ -363,9 +362,15 @@ LockedFlake lockFlake( computeLocks; computeLocks = [&]( + /* The inputs of this node, either from flake.nix or + flake.lock. */ const FlakeInputs & flakeInputs, - std::shared_ptr<Node> node, + /* The node whose locks are to be updated.*/ + ref<Node> node, + /* The path to this node in the lock file graph. */ const InputPath & inputPathPrefix, + /* The old node, if any, from which locks can be + copied. */ std::shared_ptr<const Node> oldNode, const InputPath & lockRootPath, const Path & parentPath, @@ -453,7 +458,7 @@ LockedFlake lockFlake( /* Copy the input from the old lock since its flakeref didn't change and there is no override from a higher level flake. */ - auto childNode = std::make_shared<LockedNode>( + auto childNode = make_ref<LockedNode>( oldLock->lockedRef, oldLock->originalRef, oldLock->isFlake); node->inputs.insert_or_assign(id, childNode); @@ -482,14 +487,14 @@ LockedFlake lockFlake( .isFlake = (*lockedNode)->isFlake, }); } else if (auto follows = std::get_if<1>(&i.second)) { - if (! trustLock) { + if (!trustLock) { // It is possible that the flake has changed, - // so we must confirm all the follows that are in the lockfile are also in the flake. + // so we must confirm all the follows that are in the lock file are also in the flake. auto overridePath(inputPath); overridePath.push_back(i.first); auto o = overrides.find(overridePath); // If the override disappeared, we have to refetch the flake, - // since some of the inputs may not be present in the lockfile. + // since some of the inputs may not be present in the lock file. if (o == overrides.end()) { mustRefetch = true; // There's no point populating the rest of the fake inputs, @@ -522,8 +527,8 @@ LockedFlake lockFlake( this input. */ debug("creating new input '%s'", inputPathS); - if (!lockFlags.allowMutable && !input.ref->input.isLocked()) - throw Error("cannot update flake input '%s' in pure mode", inputPathS); + if (!lockFlags.allowUnlocked && !input.ref->input.isLocked()) + throw Error("cannot update unlocked flake input '%s' in pure mode", inputPathS); /* Note: in case of an --override-input, we use the *original* ref (input2.ref) for the @@ -545,7 +550,7 @@ LockedFlake lockFlake( auto inputFlake = getFlake(state, localRef, useRegistries, flakeCache, inputPath); - auto childNode = std::make_shared<LockedNode>(inputFlake.lockedRef, ref); + auto childNode = make_ref<LockedNode>(inputFlake.lockedRef, ref); node->inputs.insert_or_assign(id, childNode); @@ -565,15 +570,19 @@ LockedFlake lockFlake( oldLock ? std::dynamic_pointer_cast<const Node>(oldLock) : LockFile::read( - inputFlake.sourceInfo->actualPath + "/" + inputFlake.lockedRef.subdir + "/flake.lock").root, - oldLock ? lockRootPath : inputPath, localPath, false); + inputFlake.sourceInfo->actualPath + "/" + inputFlake.lockedRef.subdir + "/flake.lock").root.get_ptr(), + oldLock ? lockRootPath : inputPath, + localPath, + false); } else { auto [sourceInfo, resolvedRef, lockedRef] = fetchOrSubstituteTree( state, *input.ref, useRegistries, flakeCache); - node->inputs.insert_or_assign(id, - std::make_shared<LockedNode>(lockedRef, ref, false)); + + auto childNode = make_ref<LockedNode>(lockedRef, ref, false); + + node->inputs.insert_or_assign(id, childNode); } } @@ -588,8 +597,13 @@ LockedFlake lockFlake( auto parentPath = canonPath(flake.sourceInfo->actualPath + "/" + flake.lockedRef.subdir, true); computeLocks( - flake.inputs, newLockFile.root, {}, - lockFlags.recreateLockFile ? nullptr : oldLockFile.root, {}, parentPath, false); + flake.inputs, + newLockFile.root, + {}, + lockFlags.recreateLockFile ? nullptr : oldLockFile.root.get_ptr(), + {}, + parentPath, + false); for (auto & i : lockFlags.inputOverrides) if (!overridesUsed.count(i.first)) @@ -612,9 +626,9 @@ LockedFlake lockFlake( if (lockFlags.writeLockFile) { if (auto sourcePath = topRef.input.getSourcePath()) { - if (!newLockFile.isImmutable()) { + if (auto unlockedInput = newLockFile.isUnlocked()) { if (fetchSettings.warnDirty) - warn("will not write lock file of flake '%s' because it has a mutable input", topRef); + warn("will not write lock file of flake '%s' because it has an unlocked input ('%s')", topRef, *unlockedInput); } else { if (!lockFlags.updateLockFile) throw Error("flake '%s' requires lock file changes but they're not allowed due to '--no-update-lock-file'", topRef); @@ -727,7 +741,7 @@ void callFlake(EvalState & state, static void prim_getFlake(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - std::string flakeRefS(state.forceStringNoCtx(*args[0], pos)); + std::string flakeRefS(state.forceStringNoCtx(*args[0], pos, "while evaluating the argument passed to builtins.getFlake")); auto flakeRef = parseFlakeRef(flakeRefS, {}, true); if (evalSettings.pureEval && !flakeRef.input.isLocked()) throw Error("cannot call 'getFlake' on unlocked flake reference '%s', at %s (use --impure to override)", flakeRefS, state.positions[pos]); @@ -738,7 +752,7 @@ static void prim_getFlake(EvalState & state, const PosIdx pos, Value * * args, V .updateLockFile = false, .writeLockFile = false, .useRegistries = !evalSettings.pureEval && fetchSettings.useRegistries, - .allowMutable = !evalSettings.pureEval, + .allowUnlocked = !evalSettings.pureEval, }), v); } diff --git a/src/libexpr/flake/flake.hh b/src/libexpr/flake/flake.hh index 524b18af1..10301d8aa 100644 --- a/src/libexpr/flake/flake.hh +++ b/src/libexpr/flake/flake.hh @@ -108,11 +108,11 @@ struct LockFlags bool applyNixConfig = false; - /* Whether mutable flake references (i.e. those without a Git + /* Whether unlocked flake references (i.e. those without a Git revision or similar) without a corresponding lock are - allowed. Mutable flake references with a lock are always + allowed. Unlocked flake references with a lock are always allowed. */ - bool allowMutable = true; + bool allowUnlocked = true; /* Whether to commit changes to flake.lock. */ bool commitLockFile = false; diff --git a/src/libexpr/flake/flakeref.cc b/src/libexpr/flake/flakeref.cc index eede493f8..08adbe0c9 100644 --- a/src/libexpr/flake/flakeref.cc +++ b/src/libexpr/flake/flakeref.cc @@ -238,15 +238,15 @@ std::pair<fetchers::Tree, FlakeRef> FlakeRef::fetchTree(ref<Store> store) const return {std::move(tree), FlakeRef(std::move(lockedInput), subdir)}; } -std::tuple<FlakeRef, std::string, OutputsSpec> parseFlakeRefWithFragmentAndOutputsSpec( +std::tuple<FlakeRef, std::string, ExtendedOutputsSpec> parseFlakeRefWithFragmentAndExtendedOutputsSpec( const std::string & url, const std::optional<Path> & baseDir, bool allowMissing, bool isFlake) { - auto [prefix, outputsSpec] = parseOutputsSpec(url); - auto [flakeRef, fragment] = parseFlakeRefWithFragment(prefix, baseDir, allowMissing, isFlake); - return {std::move(flakeRef), fragment, outputsSpec}; + auto [prefix, extendedOutputsSpec] = ExtendedOutputsSpec::parse(url); + auto [flakeRef, fragment] = parseFlakeRefWithFragment(std::string { prefix }, baseDir, allowMissing, isFlake); + return {std::move(flakeRef), fragment, extendedOutputsSpec}; } } diff --git a/src/libexpr/flake/flakeref.hh b/src/libexpr/flake/flakeref.hh index a9182f4bf..c4142fc20 100644 --- a/src/libexpr/flake/flakeref.hh +++ b/src/libexpr/flake/flakeref.hh @@ -3,7 +3,7 @@ #include "types.hh" #include "hash.hh" #include "fetchers.hh" -#include "path-with-outputs.hh" +#include "outputs-spec.hh" #include <variant> @@ -28,14 +28,14 @@ typedef std::string FlakeId; * object that fetcher generates (usually via * FlakeRef::fromAttrs(attrs) or parseFlakeRef(url) calls). * - * The actual fetch not have been performed yet (i.e. a FlakeRef may + * The actual fetch may not have been performed yet (i.e. a FlakeRef may * be lazy), but the fetcher can be invoked at any time via the * FlakeRef to ensure the store is populated with this input. */ struct FlakeRef { - /* fetcher-specific representation of the input, sufficient to + /* Fetcher-specific representation of the input, sufficient to perform the fetch operation. */ fetchers::Input input; @@ -80,7 +80,7 @@ std::pair<FlakeRef, std::string> parseFlakeRefWithFragment( std::optional<std::pair<FlakeRef, std::string>> maybeParseFlakeRefWithFragment( const std::string & url, const std::optional<Path> & baseDir = {}); -std::tuple<FlakeRef, std::string, OutputsSpec> parseFlakeRefWithFragmentAndOutputsSpec( +std::tuple<FlakeRef, std::string, ExtendedOutputsSpec> parseFlakeRefWithFragmentAndExtendedOutputsSpec( const std::string & url, const std::optional<Path> & baseDir = {}, bool allowMissing = false, diff --git a/src/libexpr/flake/lockfile.cc b/src/libexpr/flake/lockfile.cc index 60b52d578..a74e68c9c 100644 --- a/src/libexpr/flake/lockfile.cc +++ b/src/libexpr/flake/lockfile.cc @@ -31,12 +31,12 @@ FlakeRef getFlakeRef( } LockedNode::LockedNode(const nlohmann::json & json) - : lockedRef(getFlakeRef(json, "locked", "info")) + : lockedRef(getFlakeRef(json, "locked", "info")) // FIXME: remove "info" , originalRef(getFlakeRef(json, "original", nullptr)) , isFlake(json.find("flake") != json.end() ? (bool) json["flake"] : true) { if (!lockedRef.input.isLocked()) - throw Error("lockfile contains mutable lock '%s'", + throw Error("lock file contains mutable lock '%s'", fetchers::attrsToJSON(lockedRef.input.toAttrs())); } @@ -49,15 +49,15 @@ std::shared_ptr<Node> LockFile::findInput(const InputPath & path) { auto pos = root; - if (!pos) return {}; - for (auto & elem : path) { if (auto i = get(pos->inputs, elem)) { if (auto node = std::get_if<0>(&*i)) pos = *node; else if (auto follows = std::get_if<1>(&*i)) { - pos = findInput(*follows); - if (!pos) return {}; + if (auto p = findInput(*follows)) + pos = ref(p); + else + return {}; } } else return {}; @@ -72,7 +72,7 @@ LockFile::LockFile(const nlohmann::json & json, const Path & path) if (version < 5 || version > 7) throw Error("lock file '%s' has unsupported version %d", path, version); - std::unordered_map<std::string, std::shared_ptr<Node>> nodeMap; + std::map<std::string, ref<Node>> nodeMap; std::function<void(Node & node, const nlohmann::json & jsonNode)> getInputs; @@ -93,12 +93,12 @@ LockFile::LockFile(const nlohmann::json & json, const Path & path) auto jsonNode2 = nodes.find(inputKey); if (jsonNode2 == nodes.end()) throw Error("lock file references missing node '%s'", inputKey); - auto input = std::make_shared<LockedNode>(*jsonNode2); + auto input = make_ref<LockedNode>(*jsonNode2); k = nodeMap.insert_or_assign(inputKey, input).first; getInputs(*input, *jsonNode2); } - if (auto child = std::dynamic_pointer_cast<LockedNode>(k->second)) - node.inputs.insert_or_assign(i.key(), child); + if (auto child = k->second.dynamic_pointer_cast<LockedNode>()) + node.inputs.insert_or_assign(i.key(), ref(child)); else // FIXME: replace by follows node throw Error("lock file contains cycle to root node"); @@ -122,9 +122,9 @@ nlohmann::json LockFile::toJSON() const std::unordered_map<std::shared_ptr<const Node>, std::string> nodeKeys; std::unordered_set<std::string> keys; - std::function<std::string(const std::string & key, std::shared_ptr<const Node> node)> dumpNode; + std::function<std::string(const std::string & key, ref<const Node> node)> dumpNode; - dumpNode = [&](std::string key, std::shared_ptr<const Node> node) -> std::string + dumpNode = [&](std::string key, ref<const Node> node) -> std::string { auto k = nodeKeys.find(node); if (k != nodeKeys.end()) @@ -159,10 +159,11 @@ nlohmann::json LockFile::toJSON() const n["inputs"] = std::move(inputs); } - if (auto lockedNode = std::dynamic_pointer_cast<const LockedNode>(node)) { + if (auto lockedNode = node.dynamic_pointer_cast<const LockedNode>()) { n["original"] = fetchers::attrsToJSON(lockedNode->originalRef.toAttrs()); n["locked"] = fetchers::attrsToJSON(lockedNode->lockedRef.toAttrs()); - if (!lockedNode->isFlake) n["flake"] = false; + if (!lockedNode->isFlake) + n["flake"] = false; } nodes[key] = std::move(n); @@ -201,13 +202,13 @@ void LockFile::write(const Path & path) const writeFile(path, fmt("%s\n", *this)); } -bool LockFile::isImmutable() const +std::optional<FlakeRef> LockFile::isUnlocked() const { - std::unordered_set<std::shared_ptr<const Node>> nodes; + std::set<ref<const Node>> nodes; - std::function<void(std::shared_ptr<const Node> node)> visit; + std::function<void(ref<const Node> node)> visit; - visit = [&](std::shared_ptr<const Node> node) + visit = [&](ref<const Node> node) { if (!nodes.insert(node).second) return; for (auto & i : node->inputs) @@ -218,12 +219,13 @@ bool LockFile::isImmutable() const visit(root); for (auto & i : nodes) { - if (i == root) continue; - auto lockedNode = std::dynamic_pointer_cast<const LockedNode>(i); - if (lockedNode && !lockedNode->lockedRef.input.isLocked()) return false; + if (i == ref<const Node>(root)) continue; + auto node = i.dynamic_pointer_cast<const LockedNode>(); + if (node && !node->lockedRef.input.isLocked()) + return node->lockedRef; } - return true; + return {}; } bool LockFile::operator ==(const LockFile & other) const @@ -247,12 +249,12 @@ InputPath parseInputPath(std::string_view s) std::map<InputPath, Node::Edge> LockFile::getAllInputs() const { - std::unordered_set<std::shared_ptr<Node>> done; + std::set<ref<Node>> done; std::map<InputPath, Node::Edge> res; - std::function<void(const InputPath & prefix, std::shared_ptr<Node> node)> recurse; + std::function<void(const InputPath & prefix, ref<Node> node)> recurse; - recurse = [&](const InputPath & prefix, std::shared_ptr<Node> node) + recurse = [&](const InputPath & prefix, ref<Node> node) { if (!done.insert(node).second) return; diff --git a/src/libexpr/flake/lockfile.hh b/src/libexpr/flake/lockfile.hh index 96f1edc76..02e9bdfbc 100644 --- a/src/libexpr/flake/lockfile.hh +++ b/src/libexpr/flake/lockfile.hh @@ -20,7 +20,7 @@ struct LockedNode; type LockedNode. */ struct Node : std::enable_shared_from_this<Node> { - typedef std::variant<std::shared_ptr<LockedNode>, InputPath> Edge; + typedef std::variant<ref<LockedNode>, InputPath> Edge; std::map<FlakeId, Edge> inputs; @@ -47,11 +47,13 @@ struct LockedNode : Node struct LockFile { - std::shared_ptr<Node> root = std::make_shared<Node>(); + ref<Node> root = make_ref<Node>(); LockFile() {}; LockFile(const nlohmann::json & json, const Path & path); + typedef std::map<ref<const Node>, std::string> KeyMap; + nlohmann::json toJSON() const; std::string to_string() const; @@ -60,7 +62,8 @@ struct LockFile void write(const Path & path) const; - bool isImmutable() const; + /* Check whether this lock file has any unlocked inputs. */ + std::optional<FlakeRef> isUnlocked() const; bool operator ==(const LockFile & other) const; diff --git a/src/libexpr/get-drvs.cc b/src/libexpr/get-drvs.cc index 346741dd5..1602fbffb 100644 --- a/src/libexpr/get-drvs.cc +++ b/src/libexpr/get-drvs.cc @@ -51,7 +51,7 @@ std::string DrvInfo::queryName() const if (name == "" && attrs) { auto i = attrs->find(state->sName); if (i == attrs->end()) throw TypeError("derivation name missing"); - name = state->forceStringNoCtx(*i->value); + name = state->forceStringNoCtx(*i->value, noPos, "while evaluating the 'name' attribute of a derivation"); } return name; } @@ -61,7 +61,7 @@ std::string DrvInfo::querySystem() const { if (system == "" && attrs) { auto i = attrs->find(state->sSystem); - system = i == attrs->end() ? "unknown" : state->forceStringNoCtx(*i->value, i->pos); + system = i == attrs->end() ? "unknown" : state->forceStringNoCtx(*i->value, i->pos, "while evaluating the 'system' attribute of a derivation"); } return system; } @@ -75,7 +75,7 @@ std::optional<StorePath> DrvInfo::queryDrvPath() const if (i == attrs->end()) drvPath = {std::nullopt}; else - drvPath = {state->coerceToStorePath(i->pos, *i->value, context)}; + drvPath = {state->coerceToStorePath(i->pos, *i->value, context, "while evaluating the 'drvPath' attribute of a derivation")}; } return drvPath.value_or(std::nullopt); } @@ -95,7 +95,7 @@ StorePath DrvInfo::queryOutPath() const Bindings::iterator i = attrs->find(state->sOutPath); PathSet context; if (i != attrs->end()) - outPath = state->coerceToStorePath(i->pos, *i->value, context); + outPath = state->coerceToStorePath(i->pos, *i->value, context, "while evaluating the output path of a derivation"); } if (!outPath) throw UnimplementedError("CA derivations are not yet supported"); @@ -109,23 +109,23 @@ DrvInfo::Outputs DrvInfo::queryOutputs(bool withPaths, bool onlyOutputsToInstall /* Get the ‘outputs’ list. */ Bindings::iterator i; if (attrs && (i = attrs->find(state->sOutputs)) != attrs->end()) { - state->forceList(*i->value, i->pos); + state->forceList(*i->value, i->pos, "while evaluating the 'outputs' attribute of a derivation"); /* For each output... */ for (auto elem : i->value->listItems()) { - std::string output(state->forceStringNoCtx(*elem, i->pos)); + std::string output(state->forceStringNoCtx(*elem, i->pos, "while evaluating the name of an output of a derivation")); if (withPaths) { /* Evaluate the corresponding set. */ Bindings::iterator out = attrs->find(state->symbols.create(output)); if (out == attrs->end()) continue; // FIXME: throw error? - state->forceAttrs(*out->value, i->pos); + state->forceAttrs(*out->value, i->pos, "while evaluating an output of a derivation"); /* And evaluate its ‘outPath’ attribute. */ Bindings::iterator outPath = out->value->attrs->find(state->sOutPath); if (outPath == out->value->attrs->end()) continue; // FIXME: throw error? PathSet context; - outputs.emplace(output, state->coerceToStorePath(outPath->pos, *outPath->value, context)); + outputs.emplace(output, state->coerceToStorePath(outPath->pos, *outPath->value, context, "while evaluating an output path of a derivation")); } else outputs.emplace(output, std::nullopt); } @@ -137,7 +137,7 @@ DrvInfo::Outputs DrvInfo::queryOutputs(bool withPaths, bool onlyOutputsToInstall return outputs; Bindings::iterator i; - if (attrs && (i = attrs->find(state->sOutputSpecified)) != attrs->end() && state->forceBool(*i->value, i->pos)) { + if (attrs && (i = attrs->find(state->sOutputSpecified)) != attrs->end() && state->forceBool(*i->value, i->pos, "while evaluating the 'outputSpecified' attribute of a derivation")) { Outputs result; auto out = outputs.find(queryOutputName()); if (out == outputs.end()) @@ -150,7 +150,7 @@ DrvInfo::Outputs DrvInfo::queryOutputs(bool withPaths, bool onlyOutputsToInstall /* Check for `meta.outputsToInstall` and return `outputs` reduced to that. */ const Value * outTI = queryMeta("outputsToInstall"); if (!outTI) return outputs; - const auto errMsg = Error("this derivation has bad 'meta.outputsToInstall'"); + auto errMsg = Error("this derivation has bad 'meta.outputsToInstall'"); /* ^ this shows during `nix-env -i` right under the bad derivation */ if (!outTI->isList()) throw errMsg; Outputs result; @@ -169,7 +169,7 @@ std::string DrvInfo::queryOutputName() const { if (outputName == "" && attrs) { Bindings::iterator i = attrs->find(state->sOutputName); - outputName = i != attrs->end() ? state->forceStringNoCtx(*i->value) : ""; + outputName = i != attrs->end() ? state->forceStringNoCtx(*i->value, noPos, "while evaluating the output name of a derivation") : ""; } return outputName; } @@ -181,7 +181,7 @@ Bindings * DrvInfo::getMeta() if (!attrs) return 0; Bindings::iterator a = attrs->find(state->sMeta); if (a == attrs->end()) return 0; - state->forceAttrs(*a->value, a->pos); + state->forceAttrs(*a->value, a->pos, "while evaluating the 'meta' attribute of a derivation"); meta = a->value->attrs; return meta; } @@ -382,7 +382,7 @@ static void getDerivations(EvalState & state, Value & vIn, `recurseForDerivations = true' attribute. */ if (i->value->type() == nAttrs) { Bindings::iterator j = i->value->attrs->find(state.sRecurseForDerivations); - if (j != i->value->attrs->end() && state.forceBool(*j->value, j->pos)) + if (j != i->value->attrs->end() && state.forceBool(*j->value, j->pos, "while evaluating the attribute `recurseForDerivations`")) getDerivations(state, *i->value, pathPrefix2, autoArgs, drvs, done, ignoreAssertionFailures); } } diff --git a/src/libexpr/local.mk b/src/libexpr/local.mk index 016631647..2171e769b 100644 --- a/src/libexpr/local.mk +++ b/src/libexpr/local.mk @@ -6,6 +6,7 @@ libexpr_DIR := $(d) libexpr_SOURCES := \ $(wildcard $(d)/*.cc) \ + $(wildcard $(d)/value/*.cc) \ $(wildcard $(d)/primops/*.cc) \ $(wildcard $(d)/flake/*.cc) \ $(d)/lexer-tab.cc \ @@ -37,6 +38,8 @@ clean-files += $(d)/parser-tab.cc $(d)/parser-tab.hh $(d)/lexer-tab.cc $(d)/lexe $(eval $(call install-file-in, $(d)/nix-expr.pc, $(libdir)/pkgconfig, 0644)) +$(foreach i, $(wildcard src/libexpr/value/*.hh), \ + $(eval $(call install-file-in, $(i), $(includedir)/nix/value, 0644))) $(foreach i, $(wildcard src/libexpr/flake/*.hh), \ $(eval $(call install-file-in, $(i), $(includedir)/nix/flake, 0644))) diff --git a/src/libexpr/nix-expr.pc.in b/src/libexpr/nix-expr.pc.in index 80f7a492b..60ffb5dba 100644 --- a/src/libexpr/nix-expr.pc.in +++ b/src/libexpr/nix-expr.pc.in @@ -7,4 +7,4 @@ Description: Nix Package Manager Version: @PACKAGE_VERSION@ Requires: nix-store bdw-gc Libs: -L${libdir} -lnixexpr -Cflags: -I${includedir}/nix -std=c++17 +Cflags: -I${includedir}/nix -std=c++2a diff --git a/src/libexpr/nixexpr.cc b/src/libexpr/nixexpr.cc index 7c623a07d..eb6f062b4 100644 --- a/src/libexpr/nixexpr.cc +++ b/src/libexpr/nixexpr.cc @@ -8,6 +8,58 @@ namespace nix { +struct PosAdapter : AbstractPos +{ + Pos::Origin origin; + + PosAdapter(Pos::Origin origin) + : origin(std::move(origin)) + { + } + + std::optional<std::string> getSource() const override + { + return std::visit(overloaded { + [](const Pos::none_tag &) -> std::optional<std::string> { + return std::nullopt; + }, + [](const Pos::Stdin & s) -> std::optional<std::string> { + // Get rid of the null terminators added by the parser. + return std::string(s.source->c_str()); + }, + [](const Pos::String & s) -> std::optional<std::string> { + // Get rid of the null terminators added by the parser. + return std::string(s.source->c_str()); + }, + [](const Path & path) -> std::optional<std::string> { + try { + return readFile(path); + } catch (Error &) { + return std::nullopt; + } + } + }, origin); + } + + void print(std::ostream & out) const override + { + std::visit(overloaded { + [&](const Pos::none_tag &) { out << "«none»"; }, + [&](const Pos::Stdin &) { out << "«stdin»"; }, + [&](const Pos::String & s) { out << "«string»"; }, + [&](const Path & path) { out << path; } + }, origin); + } +}; + +Pos::operator std::shared_ptr<AbstractPos>() const +{ + auto pos = std::make_shared<PosAdapter>(origin); + pos->line = line; + pos->column = column; + return pos; +} + /* Displaying abstract syntax trees. */ static void showString(std::ostream & str, std::string_view s) @@ -248,24 +300,10 @@ void ExprPos::show(const SymbolTable & symbols, std::ostream & str) const std::ostream & operator << (std::ostream & str, const Pos & pos) { - if (!pos) + if (auto pos2 = (std::shared_ptr<AbstractPos>) pos) { + str << *pos2; + } else str << "undefined position"; - else - { - auto f = format(ANSI_BOLD "%1%" ANSI_NORMAL ":%2%:%3%"); - switch (pos.origin) { - case foFile: - f % (const std::string &) pos.file; - break; - case foStdin: - case foString: - f % "(string)"; - break; - default: - throw Error("unhandled Pos origin!"); - } - str << (f % pos.line % pos.column).str(); - } return str; } @@ -289,7 +327,6 @@ std::string showAttrPath(const SymbolTable & symbols, const AttrPath & attrPath) } - /* Computing levels/displacements for variables. */ void Expr::bindVars(EvalState & es, const std::shared_ptr<const StaticEnv> & env) diff --git a/src/libexpr/nixexpr.hh b/src/libexpr/nixexpr.hh index 5eb022770..4a81eaa47 100644 --- a/src/libexpr/nixexpr.hh +++ b/src/libexpr/nixexpr.hh @@ -8,7 +8,6 @@ #include "error.hh" #include "chunked-vector.hh" - namespace nix { @@ -23,15 +22,22 @@ MakeError(MissingArgumentError, EvalError); MakeError(RestrictedPathError, Error); /* Position objects. */ - struct Pos { - std::string file; - FileOrigin origin; uint32_t line; uint32_t column; + struct none_tag { }; + struct Stdin { ref<std::string> source; }; + struct String { ref<std::string> source; }; + + typedef std::variant<none_tag, Stdin, String, Path> Origin; + + Origin origin; + explicit operator bool() const { return line > 0; } + + operator std::shared_ptr<AbstractPos>() const; }; class PosIdx { @@ -47,7 +53,11 @@ public: explicit operator bool() const { return id > 0; } - bool operator<(const PosIdx other) const { return id < other.id; } + bool operator <(const PosIdx other) const { return id < other.id; } + + bool operator ==(const PosIdx other) const { return id == other.id; } + + bool operator !=(const PosIdx other) const { return id != other.id; } }; class PosTable @@ -61,13 +71,13 @@ public: // current origins.back() can be reused or not. mutable uint32_t idx = std::numeric_limits<uint32_t>::max(); - explicit Origin(uint32_t idx): idx(idx), file{}, origin{} {} + // Used for searching in PosTable::[]. + explicit Origin(uint32_t idx): idx(idx), origin{Pos::none_tag()} {} public: - const std::string file; - const FileOrigin origin; + const Pos::Origin origin; - Origin(std::string file, FileOrigin origin): file(std::move(file)), origin(origin) {} + Origin(Pos::Origin origin): origin(origin) {} }; struct Offset { @@ -107,7 +117,7 @@ public: [] (const auto & a, const auto & b) { return a.idx < b.idx; }); const auto origin = *std::prev(pastOrigin); const auto offset = offsets[idx]; - return {origin.file, origin.origin, offset.line, offset.column}; + return {offset.line, offset.column, origin.origin}; } }; @@ -176,7 +186,7 @@ struct ExprString : Expr { std::string s; Value v; - ExprString(std::string s) : s(std::move(s)) { v.mkString(this->s.data()); }; + ExprString(std::string &&s) : s(std::move(s)) { v.mkString(this->s.data()); }; Value * maybeThunk(EvalState & state, Env & env) override; COMMON_METHODS }; @@ -223,7 +233,7 @@ struct ExprSelect : Expr PosIdx pos; Expr * e, * def; AttrPath attrPath; - ExprSelect(const PosIdx & pos, Expr * e, const AttrPath & attrPath, Expr * def) : pos(pos), e(e), def(def), attrPath(attrPath) { }; + ExprSelect(const PosIdx & pos, Expr * e, const AttrPath && attrPath, Expr * def) : pos(pos), e(e), def(def), attrPath(std::move(attrPath)) { }; ExprSelect(const PosIdx & pos, Expr * e, Symbol name) : pos(pos), e(e), def(0) { attrPath.push_back(AttrName(name)); }; PosIdx getPos() const override { return pos; } COMMON_METHODS @@ -233,7 +243,7 @@ struct ExprOpHasAttr : Expr { Expr * e; AttrPath attrPath; - ExprOpHasAttr(Expr * e, const AttrPath & attrPath) : e(e), attrPath(attrPath) { }; + ExprOpHasAttr(Expr * e, const AttrPath && attrPath) : e(e), attrPath(std::move(attrPath)) { }; PosIdx getPos() const override { return e->getPos(); } COMMON_METHODS }; diff --git a/src/libexpr/parser.y b/src/libexpr/parser.y index 7c9b5a2db..0f75ed9a0 100644 --- a/src/libexpr/parser.y +++ b/src/libexpr/parser.y @@ -34,11 +34,6 @@ namespace nix { Path basePath; PosTable::Origin origin; std::optional<ErrorInfo> error; - ParseData(EvalState & state, PosTable::Origin origin) - : state(state) - , symbols(state.symbols) - , origin(std::move(origin)) - { }; }; struct ParserFormals { @@ -95,7 +90,7 @@ static void dupAttr(const EvalState & state, Symbol attr, const PosIdx pos, cons } -static void addAttr(ExprAttrs * attrs, AttrPath & attrPath, +static void addAttr(ExprAttrs * attrs, AttrPath && attrPath, Expr * e, const PosIdx pos, const nix::EvalState & state) { AttrPath::iterator i; @@ -193,7 +188,7 @@ static Formals * toFormals(ParseData & data, ParserFormals * formals, static Expr * stripIndentation(const PosIdx pos, SymbolTable & symbols, - std::vector<std::pair<PosIdx, std::variant<Expr *, StringToken>>> & es) + std::vector<std::pair<PosIdx, std::variant<Expr *, StringToken>>> && es) { if (es.empty()) return new ExprString(""); @@ -273,7 +268,7 @@ static Expr * stripIndentation(const PosIdx pos, SymbolTable & symbols, s2 = std::string(s2, 0, p + 1); } - es2->emplace_back(i->first, new ExprString(s2)); + es2->emplace_back(i->first, new ExprString(std::move(s2))); }; for (; i != es.end(); ++i, --n) { std::visit(overloaded { trimExpr, trimString }, i->second); @@ -405,21 +400,21 @@ expr_op | '-' expr_op %prec NEGATE { $$ = new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__sub")), {new ExprInt(0), $2}); } | expr_op EQ expr_op { $$ = new ExprOpEq($1, $3); } | expr_op NEQ expr_op { $$ = new ExprOpNEq($1, $3); } - | expr_op '<' expr_op { $$ = new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__lessThan")), {$1, $3}); } - | expr_op LEQ expr_op { $$ = new ExprOpNot(new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__lessThan")), {$3, $1})); } - | expr_op '>' expr_op { $$ = new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__lessThan")), {$3, $1}); } - | expr_op GEQ expr_op { $$ = new ExprOpNot(new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__lessThan")), {$1, $3})); } - | expr_op AND expr_op { $$ = new ExprOpAnd(CUR_POS, $1, $3); } - | expr_op OR expr_op { $$ = new ExprOpOr(CUR_POS, $1, $3); } - | expr_op IMPL expr_op { $$ = new ExprOpImpl(CUR_POS, $1, $3); } - | expr_op UPDATE expr_op { $$ = new ExprOpUpdate(CUR_POS, $1, $3); } - | expr_op '?' attrpath { $$ = new ExprOpHasAttr($1, *$3); } + | expr_op '<' expr_op { $$ = new ExprCall(makeCurPos(@2, data), new ExprVar(data->symbols.create("__lessThan")), {$1, $3}); } + | expr_op LEQ expr_op { $$ = new ExprOpNot(new ExprCall(makeCurPos(@2, data), new ExprVar(data->symbols.create("__lessThan")), {$3, $1})); } + | expr_op '>' expr_op { $$ = new ExprCall(makeCurPos(@2, data), new ExprVar(data->symbols.create("__lessThan")), {$3, $1}); } + | expr_op GEQ expr_op { $$ = new ExprOpNot(new ExprCall(makeCurPos(@2, data), new ExprVar(data->symbols.create("__lessThan")), {$1, $3})); } + | expr_op AND expr_op { $$ = new ExprOpAnd(makeCurPos(@2, data), $1, $3); } + | expr_op OR expr_op { $$ = new ExprOpOr(makeCurPos(@2, data), $1, $3); } + | expr_op IMPL expr_op { $$ = new ExprOpImpl(makeCurPos(@2, data), $1, $3); } + | expr_op UPDATE expr_op { $$ = new ExprOpUpdate(makeCurPos(@2, data), $1, $3); } + | expr_op '?' attrpath { $$ = new ExprOpHasAttr($1, std::move(*$3)); delete $3; } | expr_op '+' expr_op - { $$ = new ExprConcatStrings(CUR_POS, false, new std::vector<std::pair<PosIdx, Expr *>>({{makeCurPos(@1, data), $1}, {makeCurPos(@3, data), $3}})); } - | expr_op '-' expr_op { $$ = new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__sub")), {$1, $3}); } - | expr_op '*' expr_op { $$ = new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__mul")), {$1, $3}); } - | expr_op '/' expr_op { $$ = new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__div")), {$1, $3}); } - | expr_op CONCAT expr_op { $$ = new ExprOpConcatLists(CUR_POS, $1, $3); } + { $$ = new ExprConcatStrings(makeCurPos(@2, data), false, new std::vector<std::pair<PosIdx, Expr *> >({{makeCurPos(@1, data), $1}, {makeCurPos(@3, data), $3}})); } + | expr_op '-' expr_op { $$ = new ExprCall(makeCurPos(@2, data), new ExprVar(data->symbols.create("__sub")), {$1, $3}); } + | expr_op '*' expr_op { $$ = new ExprCall(makeCurPos(@2, data), new ExprVar(data->symbols.create("__mul")), {$1, $3}); } + | expr_op '/' expr_op { $$ = new ExprCall(makeCurPos(@2, data), new ExprVar(data->symbols.create("__div")), {$1, $3}); } + | expr_op CONCAT expr_op { $$ = new ExprOpConcatLists(makeCurPos(@2, data), $1, $3); } | expr_app ; @@ -436,14 +431,14 @@ expr_app expr_select : expr_simple '.' attrpath - { $$ = new ExprSelect(CUR_POS, $1, *$3, 0); } + { $$ = new ExprSelect(CUR_POS, $1, std::move(*$3), nullptr); delete $3; } | expr_simple '.' attrpath OR_KW expr_select - { $$ = new ExprSelect(CUR_POS, $1, *$3, $5); } + { $$ = new ExprSelect(CUR_POS, $1, std::move(*$3), $5); delete $3; } | /* Backwards compatibility: because Nixpkgs has a rarely used function named ‘or’, allow stuff like ‘map or [...]’. */ expr_simple OR_KW { $$ = new ExprCall(CUR_POS, $1, {new ExprVar(CUR_POS, data->symbols.create("or"))}); } - | expr_simple { $$ = $1; } + | expr_simple ; expr_simple @@ -458,9 +453,10 @@ expr_simple | FLOAT { $$ = new ExprFloat($1); } | '"' string_parts '"' { $$ = $2; } | IND_STRING_OPEN ind_string_parts IND_STRING_CLOSE { - $$ = stripIndentation(CUR_POS, data->symbols, *$2); + $$ = stripIndentation(CUR_POS, data->symbols, std::move(*$2)); + delete $2; } - | path_start PATH_END { $$ = $1; } + | path_start PATH_END | path_start string_parts_interpolated PATH_END { $2->insert($2->begin(), {makeCurPos(@1, data), $1}); $$ = new ExprConcatStrings(CUR_POS, false, $2); @@ -470,7 +466,7 @@ expr_simple $$ = new ExprCall(CUR_POS, new ExprVar(data->symbols.create("__findFile")), {new ExprVar(data->symbols.create("__nixPath")), - new ExprString(path)}); + new ExprString(std::move(path))}); } | URI { static bool noURLLiterals = settings.isExperimentalFeatureEnabled(Xp::NoUrlLiterals); @@ -538,7 +534,7 @@ ind_string_parts ; binds - : binds attrpath '=' expr ';' { $$ = $1; addAttr($$, *$2, $4, makeCurPos(@2, data), data->state); } + : binds attrpath '=' expr ';' { $$ = $1; addAttr($$, std::move(*$2), $4, makeCurPos(@2, data), data->state); delete $2; } | binds INHERIT attrs ';' { $$ = $1; for (auto & i : *$3) { @@ -547,6 +543,7 @@ binds auto pos = makeCurPos(@3, data); $$->attrs.emplace(i.symbol, ExprAttrs::AttrDef(new ExprVar(CUR_POS, i.symbol), pos, true)); } + delete $3; } | binds INHERIT '(' expr ')' attrs ';' { $$ = $1; @@ -556,6 +553,7 @@ binds dupAttr(data->state, i.symbol, makeCurPos(@6, data), $$->attrs[i.symbol].pos); $$->attrs.emplace(i.symbol, ExprAttrs::AttrDef(new ExprSelect(CUR_POS, $4, i.symbol), makeCurPos(@6, data))); } + delete $6; } | { $$ = new ExprAttrs(makeCurPos(@0, data)); } ; @@ -601,7 +599,7 @@ attrpath ; attr - : ID { $$ = $1; } + : ID | OR_KW { $$ = {"or", 2}; } ; @@ -617,9 +615,9 @@ expr_list formals : formal ',' formals - { $$ = $3; $$->formals.push_back(*$1); } + { $$ = $3; $$->formals.emplace_back(*$1); delete $1; } | formal - { $$ = new ParserFormals; $$->formals.push_back(*$1); $$->ellipsis = false; } + { $$ = new ParserFormals; $$->formals.emplace_back(*$1); $$->ellipsis = false; delete $1; } | { $$ = new ParserFormals; $$->ellipsis = false; } | ELLIPSIS @@ -643,29 +641,26 @@ formal #include "filetransfer.hh" #include "fetchers.hh" #include "store-api.hh" +#include "flake/flake.hh" namespace nix { -Expr * EvalState::parse(char * text, size_t length, FileOrigin origin, - const PathView path, const PathView basePath, std::shared_ptr<StaticEnv> & staticEnv) +Expr * EvalState::parse( + char * text, + size_t length, + Pos::Origin origin, + Path basePath, + std::shared_ptr<StaticEnv> & staticEnv) { yyscan_t scanner; - std::string file; - switch (origin) { - case foFile: - file = path; - break; - case foStdin: - case foString: - file = text; - break; - default: - assert(false); - } - ParseData data(*this, {file, origin}); - data.basePath = basePath; + ParseData data { + .state = *this, + .symbols = symbols, + .basePath = std::move(basePath), + .origin = {origin}, + }; yylex_init(&scanner); yy_scan_buffer(text, length, scanner); @@ -717,14 +712,15 @@ Expr * EvalState::parseExprFromFile(const Path & path, std::shared_ptr<StaticEnv auto buffer = readFile(path); // readFile should have left some extra space for terminators buffer.append("\0\0", 2); - return parse(buffer.data(), buffer.size(), foFile, path, dirOf(path), staticEnv); + return parse(buffer.data(), buffer.size(), path, dirOf(path), staticEnv); } -Expr * EvalState::parseExprFromString(std::string s, const Path & basePath, std::shared_ptr<StaticEnv> & staticEnv) +Expr * EvalState::parseExprFromString(std::string s_, const Path & basePath, std::shared_ptr<StaticEnv> & staticEnv) { - s.append("\0\0", 2); - return parse(s.data(), s.size(), foString, "", basePath, staticEnv); + auto s = make_ref<std::string>(std::move(s_)); + s->append("\0\0", 2); + return parse(s->data(), s->size(), Pos::String{.source = s}, basePath, staticEnv); } @@ -736,11 +732,12 @@ Expr * EvalState::parseExprFromString(std::string s, const Path & basePath) Expr * EvalState::parseStdin() { - //Activity act(*logger, lvlTalkative, format("parsing standard input")); + //Activity act(*logger, lvlTalkative, "parsing standard input"); auto buffer = drainFD(0); // drainFD should have left some extra space for terminators buffer.append("\0\0", 2); - return parse(buffer.data(), buffer.size(), foStdin, "", absPath("."), staticBaseEnv); + auto s = make_ref<std::string>(std::move(buffer)); + return parse(s->data(), s->size(), Pos::Stdin{.source = s}, absPath("."), staticBaseEnv); } @@ -788,13 +785,13 @@ Path EvalState::findFile(SearchPath & searchPath, const std::string_view path, c if (hasPrefix(path, "nix/")) return concatStrings(corepkgsPrefix, path.substr(4)); - debugThrowLastTrace(ThrownError({ + debugThrow(ThrownError({ .msg = hintfmt(evalSettings.pureEval ? "cannot look up '<%s>' in pure evaluation mode (use '--impure' to override)" : "file '%s' was not found in the Nix search path (add it using $NIX_PATH or -I)", path), .errPos = positions[pos] - })); + }), 0, 0); } @@ -805,17 +802,28 @@ std::pair<bool, std::string> EvalState::resolveSearchPathElem(const SearchPathEl std::pair<bool, std::string> res; - if (isUri(elem.second)) { + if (EvalSettings::isPseudoUrl(elem.second)) { try { - res = { true, store->toRealPath(fetchers::downloadTarball( - store, resolveUri(elem.second), "source", false).first.storePath) }; + auto storePath = fetchers::downloadTarball( + store, EvalSettings::resolvePseudoUrl(elem.second), "source", false).first.storePath; + res = { true, store->toRealPath(storePath) }; } catch (FileTransferError & e) { logWarning({ .msg = hintfmt("Nix search path entry '%1%' cannot be downloaded, ignoring", elem.second) }); res = { false, "" }; } - } else { + } + + else if (hasPrefix(elem.second, "flake:")) { + settings.requireExperimentalFeature(Xp::Flakes); + auto flakeRef = parseFlakeRef(elem.second.substr(6), {}, true, false); + debug("fetching flake search path element '%s''", elem.second); + auto storePath = flakeRef.resolve(store).fetchTree(store).first.storePath; + res = { true, store->toRealPath(storePath) }; + } + + else { auto path = absPath(elem.second); if (pathExists(path)) res = { true, path }; @@ -827,7 +835,7 @@ std::pair<bool, std::string> EvalState::resolveSearchPathElem(const SearchPathEl } } - debug(format("resolved search path element '%s' to '%s'") % elem.second % res.second); + debug("resolved search path element '%s' to '%s'", elem.second, res.second); searchPathResolved[elem.second] = res; return res; diff --git a/src/libexpr/primops.cc b/src/libexpr/primops.cc index bc253d0a3..fb7fc3ddb 100644 --- a/src/libexpr/primops.cc +++ b/src/libexpr/primops.cc @@ -5,14 +5,15 @@ #include "globals.hh" #include "json-to-value.hh" #include "names.hh" +#include "references.hh" #include "store-api.hh" #include "util.hh" -#include "json.hh" #include "value-to-json.hh" #include "value-to-xml.hh" #include "primops.hh" #include <boost/container/small_vector.hpp> +#include <nlohmann/json.hpp> #include <sys/types.h> #include <sys/stat.h> @@ -42,16 +43,32 @@ StringMap EvalState::realiseContext(const PathSet & context) std::vector<DerivedPath::Built> drvs; StringMap res; - for (auto & i : context) { - auto [ctx, outputName] = decodeContext(*store, i); - auto ctxS = store->printStorePath(ctx); - if (!store->isValidPath(ctx)) - debugThrowLastTrace(InvalidPathError(store->printStorePath(ctx))); - if (!outputName.empty() && ctx.isDerivation()) { - drvs.push_back({ctx, {outputName}}); - } else { - res.insert_or_assign(ctxS, ctxS); - } + for (auto & c_ : context) { + auto ensureValid = [&](const StorePath & p) { + if (!store->isValidPath(p)) + debugThrowLastTrace(InvalidPathError(store->printStorePath(p))); + }; + auto c = NixStringContextElem::parse(*store, c_); + std::visit(overloaded { + [&](const NixStringContextElem::Built & b) { + drvs.push_back(DerivedPath::Built { + .drvPath = b.drvPath, + .outputs = OutputsSpec::Names { b.output }, + }); + ensureValid(b.drvPath); + }, + [&](const NixStringContextElem::Opaque & o) { + auto ctxS = store->printStorePath(o.path); + res.insert_or_assign(ctxS, ctxS); + ensureValid(o.path); + }, + [&](const NixStringContextElem::DrvDeep & d) { + /* Treat same as Opaque */ + auto ctxS = store->printStorePath(d.drvPath); + res.insert_or_assign(ctxS, ctxS); + ensureValid(d.drvPath); + }, + }, c.raw()); } if (drvs.empty()) return {}; @@ -67,16 +84,12 @@ StringMap EvalState::realiseContext(const PathSet & context) store->buildPaths(buildReqs); /* Get all the output paths corresponding to the placeholders we had */ - for (auto & [drvPath, outputs] : drvs) { - const auto outputPaths = store->queryDerivationOutputMap(drvPath); - for (auto & outputName : outputs) { - auto outputPath = get(outputPaths, outputName); - if (!outputPath) - debugThrowLastTrace(Error("derivation '%s' does not have an output named '%s'", - store->printStorePath(drvPath), outputName)); + for (auto & drv : drvs) { + auto outputs = resolveDerivedPath(*store, drv); + for (auto & [outputName, outputPath] : outputs) { res.insert_or_assign( - downstreamPlaceholder(*store, drvPath, outputName), - store->printStorePath(*outputPath) + downstreamPlaceholder(*store, drv.drvPath, outputName), + store->printStorePath(outputPath) ); } } @@ -101,15 +114,7 @@ static Path realisePath(EvalState & state, const PosIdx pos, Value & v, const Re { PathSet context; - auto path = [&]() - { - try { - return state.coerceToPath(pos, v, context); - } catch (Error & e) { - e.addTrace(state.positions[pos], "while realising the context of a path"); - throw; - } - }(); + auto path = state.coerceToPath(noPos, v, context, "while realising the context of a path"); try { StringMap rewrites = state.realiseContext(context); @@ -196,9 +201,9 @@ static void import(EvalState & state, const PosIdx pos, Value & vPath, Value * v , "/"), **state.vImportedDrvToDerivation); } - state.forceFunction(**state.vImportedDrvToDerivation, pos); + state.forceFunction(**state.vImportedDrvToDerivation, pos, "while evaluating imported-drv-to-derivation.nix.gen.hh"); v.mkApp(*state.vImportedDrvToDerivation, w); - state.forceAttrs(v, pos); + state.forceAttrs(v, pos, "while calling imported-drv-to-derivation.nix.gen.hh"); } else if (path == corepkgsPrefix + "fetchurl.nix") { @@ -211,7 +216,7 @@ static void import(EvalState & state, const PosIdx pos, Value & vPath, Value * v if (!vScope) state.evalFile(path, v); else { - state.forceAttrs(*vScope, pos); + state.forceAttrs(*vScope, pos, "while evaluating the first argument passed to builtins.scopedImport"); Env * env = &state.allocEnv(vScope->attrs->size()); env->up = &state.baseEnv; @@ -247,6 +252,7 @@ static RegisterPrimOp primop_scopedImport(RegisterPrimOp::Info { static RegisterPrimOp primop_import({ .name = "import", .args = {"path"}, + // TODO turn "normal path values" into link below .doc = R"( Load, parse and return the Nix expression in the file *path*. If *path* is a directory, the file ` default.nix ` in that directory @@ -260,7 +266,7 @@ static RegisterPrimOp primop_import({ > > Unlike some languages, `import` is a regular function in Nix. > Paths using the angle bracket syntax (e.g., `import` *\<foo\>*) - > are [normal path values](language-values.md). + > are normal [path values](@docroot@/language/values.md#type-path). A Nix expression loaded by `import` must not contain any *free variables* (identifiers that are not defined in the Nix expression @@ -315,7 +321,7 @@ void prim_importNative(EvalState & state, const PosIdx pos, Value * * args, Valu { auto path = realisePath(state, pos, *args[0]); - std::string sym(state.forceStringNoCtx(*args[1], pos)); + std::string sym(state.forceStringNoCtx(*args[1], pos, "while evaluating the second argument passed to builtins.importNative")); void *handle = dlopen(path.c_str(), RTLD_LAZY | RTLD_LOCAL); if (!handle) @@ -340,48 +346,44 @@ void prim_importNative(EvalState & state, const PosIdx pos, Value * * args, Valu /* Execute a program and parse its output */ void prim_exec(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[0], pos); + state.forceList(*args[0], pos, "while evaluating the first argument passed to builtins.exec"); auto elems = args[0]->listElems(); auto count = args[0]->listSize(); if (count == 0) - state.debugThrowLastTrace(EvalError({ - .msg = hintfmt("at least one argument to 'exec' required"), - .errPos = state.positions[pos] - })); + state.error("at least one argument to 'exec' required").atPos(pos).debugThrow<EvalError>(); PathSet context; - auto program = state.coerceToString(pos, *elems[0], context, false, false).toOwned(); + auto program = state.coerceToString(pos, *elems[0], context, + "while evaluating the first element of the argument passed to builtins.exec", + false, false).toOwned(); Strings commandArgs; for (unsigned int i = 1; i < args[0]->listSize(); ++i) { - commandArgs.push_back(state.coerceToString(pos, *elems[i], context, false, false).toOwned()); + commandArgs.push_back( + state.coerceToString(pos, *elems[i], context, + "while evaluating an element of the argument passed to builtins.exec", + false, false).toOwned()); } try { auto _ = state.realiseContext(context); // FIXME: Handle CA derivations } catch (InvalidPathError & e) { - state.debugThrowLastTrace(EvalError({ - .msg = hintfmt("cannot execute '%1%', since path '%2%' is not valid", - program, e.path), - .errPos = state.positions[pos] - })); + state.error("cannot execute '%1%', since path '%2%' is not valid", program, e.path).atPos(pos).debugThrow<EvalError>(); } auto output = runProgram(program, true, commandArgs); Expr * parsed; try { - auto base = state.positions[pos]; - parsed = state.parseExprFromString(std::move(output), base.file); + parsed = state.parseExprFromString(std::move(output), "/"); } catch (Error & e) { - e.addTrace(state.positions[pos], "While parsing the output from '%1%'", program); + e.addTrace(state.positions[pos], "while parsing the output from '%1%'", program); throw; } try { state.eval(parsed, v); } catch (Error & e) { - e.addTrace(state.positions[pos], "While evaluating the output from '%1%'", program); + e.addTrace(state.positions[pos], "while evaluating the output from '%1%'", program); throw; } } - /* Return a string representing the type of the expression. */ static void prim_typeOf(EvalState & state, const PosIdx pos, Value * * args, Value & v) { @@ -532,42 +534,69 @@ static RegisterPrimOp primop_isPath({ .fun = prim_isPath, }); +template<typename Callable> + static inline void withExceptionContext(Trace trace, Callable&& func) +{ + try + { + func(); + } + catch(Error & e) + { + e.pushTrace(trace); + throw; + } +} + struct CompareValues { EvalState & state; + const PosIdx pos; + const std::string_view errorCtx; - CompareValues(EvalState & state) : state(state) { }; + CompareValues(EvalState & state, const PosIdx pos, const std::string_view && errorCtx) : state(state), pos(pos), errorCtx(errorCtx) { }; bool operator () (Value * v1, Value * v2) const { - if (v1->type() == nFloat && v2->type() == nInt) - return v1->fpoint < v2->integer; - if (v1->type() == nInt && v2->type() == nFloat) - return v1->integer < v2->fpoint; - if (v1->type() != v2->type()) - state.debugThrowLastTrace(EvalError("cannot compare %1% with %2%", showType(*v1), showType(*v2))); - switch (v1->type()) { - case nInt: - return v1->integer < v2->integer; - case nFloat: - return v1->fpoint < v2->fpoint; - case nString: - return strcmp(v1->string.s, v2->string.s) < 0; - case nPath: - return strcmp(v1->path, v2->path) < 0; - case nList: - // Lexicographic comparison - for (size_t i = 0;; i++) { - if (i == v2->listSize()) { - return false; - } else if (i == v1->listSize()) { - return true; - } else if (!state.eqValues(*v1->listElems()[i], *v2->listElems()[i])) { - return (*this)(v1->listElems()[i], v2->listElems()[i]); + return (*this)(v1, v2, errorCtx); + } + + bool operator () (Value * v1, Value * v2, std::string_view errorCtx) const + { + try { + if (v1->type() == nFloat && v2->type() == nInt) + return v1->fpoint < v2->integer; + if (v1->type() == nInt && v2->type() == nFloat) + return v1->integer < v2->fpoint; + if (v1->type() != v2->type()) + state.error("cannot compare %s with %s", showType(*v1), showType(*v2)).debugThrow<EvalError>(); + switch (v1->type()) { + case nInt: + return v1->integer < v2->integer; + case nFloat: + return v1->fpoint < v2->fpoint; + case nString: + return strcmp(v1->string.s, v2->string.s) < 0; + case nPath: + return strcmp(v1->path, v2->path) < 0; + case nList: + // Lexicographic comparison + for (size_t i = 0;; i++) { + if (i == v2->listSize()) { + return false; + } else if (i == v1->listSize()) { + return true; + } else if (!state.eqValues(*v1->listElems()[i], *v2->listElems()[i], pos, errorCtx)) { + return (*this)(v1->listElems()[i], v2->listElems()[i], "while comparing two list elements"); + } } - } - default: - state.debugThrowLastTrace(EvalError("cannot compare %1% with %2%", showType(*v1), showType(*v2))); + default: + state.error("cannot compare %s with %s; values of that type are incomparable", showType(*v1), showType(*v2)).debugThrow<EvalError>(); + } + } catch (Error & e) { + if (!errorCtx.empty()) + e.addTrace(nullptr, errorCtx); + throw; } } }; @@ -582,105 +611,67 @@ typedef std::list<Value *> ValueList; static Bindings::iterator getAttr( EvalState & state, - std::string_view funcName, Symbol attrSym, Bindings * attrSet, - const PosIdx pos) + std::string_view errorCtx) { Bindings::iterator value = attrSet->find(attrSym); if (value == attrSet->end()) { - hintformat errorMsg = hintfmt( - "attribute '%s' missing for call to '%s'", - state.symbols[attrSym], - funcName - ); - - auto aPos = attrSet->pos; - if (!aPos) { - state.debugThrowLastTrace(TypeError({ - .msg = errorMsg, - .errPos = state.positions[pos], - })); - } else { - auto e = TypeError({ - .msg = errorMsg, - .errPos = state.positions[aPos], - }); - - // Adding another trace for the function name to make it clear - // which call received wrong arguments. - e.addTrace(state.positions[pos], hintfmt("while invoking '%s'", funcName)); - state.debugThrowLastTrace(e); - } + state.error("attribute '%s' missing", state.symbols[attrSym]).withTrace(noPos, errorCtx).debugThrow<TypeError>(); } - return value; } static void prim_genericClosure(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); + state.forceAttrs(*args[0], noPos, "while evaluating the first argument passed to builtins.genericClosure"); /* Get the start set. */ - Bindings::iterator startSet = getAttr( - state, - "genericClosure", - state.sStartSet, - args[0]->attrs, - pos - ); + Bindings::iterator startSet = getAttr(state, state.sStartSet, args[0]->attrs, "in the attrset passed as argument to builtins.genericClosure"); - state.forceList(*startSet->value, pos); + state.forceList(*startSet->value, noPos, "while evaluating the 'startSet' attribute passed as argument to builtins.genericClosure"); ValueList workSet; for (auto elem : startSet->value->listItems()) workSet.push_back(elem); - /* Get the operator. */ - Bindings::iterator op = getAttr( - state, - "genericClosure", - state.sOperator, - args[0]->attrs, - pos - ); + if (startSet->value->listSize() == 0) { + v = *startSet->value; + return; + } - state.forceValue(*op->value, pos); + /* Get the operator. */ + Bindings::iterator op = getAttr(state, state.sOperator, args[0]->attrs, "in the attrset passed as argument to builtins.genericClosure"); + state.forceFunction(*op->value, noPos, "while evaluating the 'operator' attribute passed as argument to builtins.genericClosure"); - /* Construct the closure by applying the operator to element of + /* Construct the closure by applying the operator to elements of `workSet', adding the result to `workSet', continuing until no new elements are found. */ ValueList res; // `doneKeys' doesn't need to be a GC root, because its values are // reachable from res. - auto cmp = CompareValues(state); + auto cmp = CompareValues(state, noPos, "while comparing the `key` attributes of two genericClosure elements"); std::set<Value *, decltype(cmp)> doneKeys(cmp); while (!workSet.empty()) { Value * e = *(workSet.begin()); workSet.pop_front(); - state.forceAttrs(*e, pos); + state.forceAttrs(*e, noPos, "while evaluating one of the elements generated by (or initially passed to) builtins.genericClosure"); - Bindings::iterator key = - e->attrs->find(state.sKey); - if (key == e->attrs->end()) - state.debugThrowLastTrace(EvalError({ - .msg = hintfmt("attribute 'key' required"), - .errPos = state.positions[pos] - })); - state.forceValue(*key->value, pos); + Bindings::iterator key = getAttr(state, state.sKey, e->attrs, "in one of the attrsets generated by (or initially passed to) builtins.genericClosure"); + state.forceValue(*key->value, noPos); if (!doneKeys.insert(key->value).second) continue; res.push_back(e); /* Call the `operator' function with `e' as argument. */ - Value call; - call.mkApp(op->value, e); - state.forceList(call, pos); + Value newElements; + state.callFunction(*op->value, 1, &e, newElements, noPos); + state.forceList(newElements, noPos, "while evaluating the return value of the `operator` passed to builtins.genericClosure"); /* Add the values returned by the operator to the work set. */ - for (auto elem : call.listItems()) { - state.forceValue(*elem, pos); + for (auto elem : newElements.listItems()) { + state.forceValue(*elem, noPos); // "while evaluating one one of the elements returned by the `operator` passed to builtins.genericClosure"); workSet.push_back(elem); } } @@ -748,7 +739,7 @@ static RegisterPrimOp primop_break({ throw Error(ErrorInfo{ .level = lvlInfo, .msg = hintfmt("quit the debugger"), - .errPos = state.positions[noPos], + .errPos = nullptr, }); } } @@ -767,7 +758,8 @@ static RegisterPrimOp primop_abort({ .fun = [](EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto s = state.coerceToString(pos, *args[0], context).toOwned(); + auto s = state.coerceToString(pos, *args[0], context, + "while evaluating the error message passed to builtins.abort").toOwned(); state.debugThrowLastTrace(Abort("evaluation aborted with the following error message: '%1%'", s)); } }); @@ -785,7 +777,8 @@ static RegisterPrimOp primop_throw({ .fun = [](EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto s = state.coerceToString(pos, *args[0], context).toOwned(); + auto s = state.coerceToString(pos, *args[0], context, + "while evaluating the error message passed to builtin.throw").toOwned(); state.debugThrowLastTrace(ThrownError(s)); } }); @@ -797,7 +790,10 @@ static void prim_addErrorContext(EvalState & state, const PosIdx pos, Value * * v = *args[1]; } catch (Error & e) { PathSet context; - e.addTrace(std::nullopt, state.coerceToString(pos, *args[0], context).toOwned()); + auto message = state.coerceToString(pos, *args[0], context, + "while evaluating the error message passed to builtins.addErrorContext", + false, false).toOwned(); + e.addTrace(nullptr, message, true); throw; } } @@ -810,7 +806,8 @@ static RegisterPrimOp primop_addErrorContext(RegisterPrimOp::Info { static void prim_ceil(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto value = state.forceFloat(*args[0], args[0]->determinePos(pos)); + auto value = state.forceFloat(*args[0], args[0]->determinePos(pos), + "while evaluating the first argument passed to builtins.ceil"); v.mkInt(ceil(value)); } @@ -829,7 +826,7 @@ static RegisterPrimOp primop_ceil({ static void prim_floor(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto value = state.forceFloat(*args[0], args[0]->determinePos(pos)); + auto value = state.forceFloat(*args[0], args[0]->determinePos(pos), "while evaluating the first argument passed to builtins.floor"); v.mkInt(floor(value)); } @@ -903,7 +900,7 @@ static RegisterPrimOp primop_tryEval({ /* Return an environment variable. Use with care. */ static void prim_getEnv(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - std::string name(state.forceStringNoCtx(*args[0], pos)); + std::string name(state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.getEnv")); v.mkString(evalSettings.restrictEval || evalSettings.pureEval ? "" : getEnv(name).value_or("")); } @@ -1000,6 +997,7 @@ static void prim_second(EvalState & state, const PosIdx pos, Value * * args, Val * Derivations *************************************************************/ +static void derivationStrictInternal(EvalState & state, const std::string & name, Bindings * attrs, Value & v); /* Construct (as a unobservable side effect) a Nix derivation expression that performs the derivation described by the argument @@ -1010,38 +1008,68 @@ static void prim_second(EvalState & state, const PosIdx pos, Value * * args, Val derivation. */ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); + state.forceAttrs(*args[0], pos, "while evaluating the argument passed to builtins.derivationStrict"); + + Bindings * attrs = args[0]->attrs; /* Figure out the name first (for stack backtraces). */ - Bindings::iterator attr = getAttr( - state, - "derivationStrict", - state.sName, - args[0]->attrs, - pos - ); + Bindings::iterator nameAttr = getAttr(state, state.sName, attrs, "in the attrset passed as argument to builtins.derivationStrict"); std::string drvName; - const auto posDrvName = attr->pos; try { - drvName = state.forceStringNoCtx(*attr->value, pos); + drvName = state.forceStringNoCtx(*nameAttr->value, pos, "while evaluating the `name` attribute passed to builtins.derivationStrict"); + } catch (Error & e) { + e.addTrace(state.positions[nameAttr->pos], "while evaluating the derivation attribute 'name'"); + throw; + } + + try { + derivationStrictInternal(state, drvName, attrs, v); } catch (Error & e) { - e.addTrace(state.positions[posDrvName], "while evaluating the derivation attribute 'name'"); + Pos pos = state.positions[nameAttr->pos]; + /* + * Here we make two abuses of the error system + * + * 1. We print the location as a string to avoid a code snippet being + * printed. While the location of the name attribute is a good hint, the + * exact code there is irrelevant. + * + * 2. We mark this trace as a frame trace, meaning that we stop printing + * less important traces from now on. In particular, this prevents the + * display of the automatic "while calling builtins.derivationStrict" + * trace, which is of little use for the public we target here. + * + * Please keep in mind that error reporting is done on a best-effort + * basis in nix. There is no accurate location for a derivation, as it + * often results from the composition of several functions + * (derivationStrict, derivation, mkDerivation, mkPythonModule, etc.) + */ + e.addTrace(nullptr, hintfmt( + "while evaluating derivation '%s'\n" + " whose name attribute is located at %s", + drvName, pos), true); throw; } +} +static void derivationStrictInternal(EvalState & state, const std::string & +drvName, Bindings * attrs, Value & v) +{ /* Check whether attributes should be passed as a JSON file. */ - std::ostringstream jsonBuf; - std::unique_ptr<JSONObject> jsonObject; - attr = args[0]->attrs->find(state.sStructuredAttrs); - if (attr != args[0]->attrs->end() && state.forceBool(*attr->value, pos)) - jsonObject = std::make_unique<JSONObject>(jsonBuf); + using nlohmann::json; + std::optional<json> jsonObject; + auto attr = attrs->find(state.sStructuredAttrs); + if (attr != attrs->end() && + state.forceBool(*attr->value, noPos, + "while evaluating the `__structuredAttrs` " + "attribute passed to builtins.derivationStrict")) + jsonObject = json::object(); /* Check whether null attributes should be ignored. */ bool ignoreNulls = false; - attr = args[0]->attrs->find(state.sIgnoreNulls); - if (attr != args[0]->attrs->end()) - ignoreNulls = state.forceBool(*attr->value, pos); + attr = attrs->find(state.sIgnoreNulls); + if (attr != attrs->end()) + ignoreNulls = state.forceBool(*attr->value, noPos, "while evaluating the `__ignoreNulls` attribute " "passed to builtins.derivationStrict"); /* Build the derivation expression by processing the attributes. */ Derivation drv; @@ -1058,7 +1086,7 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * StringSet outputs; outputs.insert("out"); - for (auto & i : args[0]->attrs->lexicographicOrder(state.symbols)) { + for (auto & i : attrs->lexicographicOrder(state.symbols)) { if (i->name == state.sIgnoreNulls) continue; const std::string & key = state.symbols[i->name]; vomit("processing attribute '%1%'", key); @@ -1069,7 +1097,7 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * else state.debugThrowLastTrace(EvalError({ .msg = hintfmt("invalid value '%s' for 'outputHashMode' attribute", s), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); }; @@ -1079,7 +1107,7 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * if (outputs.find(j) != outputs.end()) state.debugThrowLastTrace(EvalError({ .msg = hintfmt("duplicate derivation output '%1%'", j), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); /* !!! Check whether j is a valid attribute name. */ @@ -1089,32 +1117,35 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * if (j == "drv") state.debugThrowLastTrace(EvalError({ .msg = hintfmt("invalid derivation output name 'drv'" ), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); outputs.insert(j); } if (outputs.empty()) state.debugThrowLastTrace(EvalError({ .msg = hintfmt("derivation cannot have an empty set of outputs"), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); }; try { + // This try-catch block adds context for most errors. + // Use this empty error context to signify that we defer to it. + const std::string_view context_below(""); if (ignoreNulls) { - state.forceValue(*i->value, pos); + state.forceValue(*i->value, noPos); if (i->value->type() == nNull) continue; } if (i->name == state.sContentAddressed) { - contentAddressed = state.forceBool(*i->value, pos); + contentAddressed = state.forceBool(*i->value, noPos, context_below); if (contentAddressed) settings.requireExperimentalFeature(Xp::CaDerivations); } else if (i->name == state.sImpure) { - isImpure = state.forceBool(*i->value, pos); + isImpure = state.forceBool(*i->value, noPos, context_below); if (isImpure) settings.requireExperimentalFeature(Xp::ImpureDerivations); } @@ -1122,9 +1153,11 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * /* The `args' attribute is special: it supplies the command-line arguments to the builder. */ else if (i->name == state.sArgs) { - state.forceList(*i->value, pos); + state.forceList(*i->value, noPos, context_below); for (auto elem : i->value->listItems()) { - auto s = state.coerceToString(posDrvName, *elem, context, true).toOwned(); + auto s = state.coerceToString(noPos, *elem, context, + "while evaluating an element of the argument list", + true).toOwned(); drv.args.push_back(s); } } @@ -1137,30 +1170,29 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * if (i->name == state.sStructuredAttrs) continue; - auto placeholder(jsonObject->placeholder(key)); - printValueAsJSON(state, true, *i->value, pos, placeholder, context); + (*jsonObject)[key] = printValueAsJSON(state, true, *i->value, noPos, context); if (i->name == state.sBuilder) - drv.builder = state.forceString(*i->value, context, posDrvName); + drv.builder = state.forceString(*i->value, context, noPos, context_below); else if (i->name == state.sSystem) - drv.platform = state.forceStringNoCtx(*i->value, posDrvName); + drv.platform = state.forceStringNoCtx(*i->value, noPos, context_below); else if (i->name == state.sOutputHash) - outputHash = state.forceStringNoCtx(*i->value, posDrvName); + outputHash = state.forceStringNoCtx(*i->value, noPos, context_below); else if (i->name == state.sOutputHashAlgo) - outputHashAlgo = state.forceStringNoCtx(*i->value, posDrvName); + outputHashAlgo = state.forceStringNoCtx(*i->value, noPos, context_below); else if (i->name == state.sOutputHashMode) - handleHashMode(state.forceStringNoCtx(*i->value, posDrvName)); + handleHashMode(state.forceStringNoCtx(*i->value, noPos, context_below)); else if (i->name == state.sOutputs) { /* Require ‘outputs’ to be a list of strings. */ - state.forceList(*i->value, posDrvName); + state.forceList(*i->value, noPos, context_below); Strings ss; for (auto elem : i->value->listItems()) - ss.emplace_back(state.forceStringNoCtx(*elem, posDrvName)); + ss.emplace_back(state.forceStringNoCtx(*elem, noPos, context_below)); handleOutputs(ss); } } else { - auto s = state.coerceToString(i->pos, *i->value, context, true).toOwned(); + auto s = state.coerceToString(noPos, *i->value, context, context_below, true).toOwned(); drv.env.emplace(key, s); if (i->name == state.sBuilder) drv.builder = std::move(s); else if (i->name == state.sSystem) drv.platform = std::move(s); @@ -1174,70 +1206,66 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * } } catch (Error & e) { - e.addTrace(state.positions[posDrvName], - "while evaluating the attribute '%1%' of the derivation '%2%'", - key, drvName); + e.addTrace(state.positions[i->pos], + hintfmt("while evaluating attribute '%1%' of derivation '%2%'", key, drvName), + true); throw; } } if (jsonObject) { + drv.env.emplace("__json", jsonObject->dump()); jsonObject.reset(); - drv.env.emplace("__json", jsonBuf.str()); } /* Everything in the context of the strings in the derivation attributes should be added as dependencies of the resulting derivation. */ - for (auto & path : context) { - - /* Paths marked with `=' denote that the path of a derivation - is explicitly passed to the builder. Since that allows the - builder to gain access to every path in the dependency - graph of the derivation (including all outputs), all paths - in the graph must be added to this derivation's list of - inputs to ensure that they are available when the builder - runs. */ - if (path.at(0) == '=') { - /* !!! This doesn't work if readOnlyMode is set. */ - StorePathSet refs; - state.store->computeFSClosure(state.store->parseStorePath(std::string_view(path).substr(1)), refs); - for (auto & j : refs) { - drv.inputSrcs.insert(j); - if (j.isDerivation()) - drv.inputDrvs[j] = state.store->readDerivation(j).outputNames(); - } - } - - /* Handle derivation outputs of the form ‘!<name>!<path>’. */ - else if (path.at(0) == '!') { - auto ctx = decodeContext(*state.store, path); - drv.inputDrvs[ctx.first].insert(ctx.second); - } - - /* Otherwise it's a source file. */ - else - drv.inputSrcs.insert(state.store->parseStorePath(path)); + for (auto & c_ : context) { + auto c = NixStringContextElem::parse(*state.store, c_); + std::visit(overloaded { + /* Since this allows the builder to gain access to every + path in the dependency graph of the derivation (including + all outputs), all paths in the graph must be added to + this derivation's list of inputs to ensure that they are + available when the builder runs. */ + [&](const NixStringContextElem::DrvDeep & d) { + /* !!! This doesn't work if readOnlyMode is set. */ + StorePathSet refs; + state.store->computeFSClosure(d.drvPath, refs); + for (auto & j : refs) { + drv.inputSrcs.insert(j); + if (j.isDerivation()) + drv.inputDrvs[j] = state.store->readDerivation(j).outputNames(); + } + }, + [&](const NixStringContextElem::Built & b) { + drv.inputDrvs[b.drvPath].insert(b.output); + }, + [&](const NixStringContextElem::Opaque & o) { + drv.inputSrcs.insert(o.path); + }, + }, c.raw()); } /* Do we have all required attributes? */ if (drv.builder == "") state.debugThrowLastTrace(EvalError({ .msg = hintfmt("required attribute 'builder' missing"), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); if (drv.platform == "") state.debugThrowLastTrace(EvalError({ .msg = hintfmt("required attribute 'system' missing"), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); /* Check whether the derivation name is valid. */ if (isDerivation(drvName)) state.debugThrowLastTrace(EvalError({ .msg = hintfmt("derivation names are not allowed to end in '%s'", drvExtension), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); if (outputHash) { @@ -1248,7 +1276,7 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * if (outputs.size() != 1 || *(outputs.begin()) != "out") state.debugThrowLastTrace(Error({ .msg = hintfmt("multiple outputs are not supported in fixed-output derivations"), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] })); auto h = newHashAllowEmpty(*outputHash, parseHashTypeOpt(outputHashAlgo)); @@ -1269,7 +1297,7 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * if (contentAddressed && isImpure) throw EvalError({ .msg = hintfmt("derivation cannot be both content-addressed and impure"), - .errPos = state.positions[posDrvName] + .errPos = state.positions[noPos] }); auto ht = parseHashTypeOpt(outputHashAlgo).value_or(htSHA256); @@ -1313,7 +1341,7 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * if (!h) throw AssertionError({ .msg = hintfmt("derivation produced no hash for output '%s'", i), - .errPos = state.positions[posDrvName], + .errPos = state.positions[noPos], }); auto outPath = state.store->makeOutputPath(i, *h, drvName); drv.env[i] = state.store->printStorePath(outPath); @@ -1346,11 +1374,12 @@ static void prim_derivationStrict(EvalState & state, const PosIdx pos, Value * * drvHashes.lock()->insert_or_assign(drvPath, h); } - auto attrs = state.buildBindings(1 + drv.outputs.size()); - attrs.alloc(state.sDrvPath).mkString(drvPathS, {"=" + drvPathS}); + auto result = state.buildBindings(1 + drv.outputs.size()); + result.alloc(state.sDrvPath).mkString(drvPathS, {"=" + drvPathS}); for (auto & i : drv.outputs) - mkOutputString(state, attrs, drvPath, drv, i); - v.mkAttrs(attrs); + mkOutputString(state, result, drvPath, drv, i); + + v.mkAttrs(result); } static RegisterPrimOp primop_derivationStrict(RegisterPrimOp::Info { @@ -1368,7 +1397,7 @@ static RegisterPrimOp primop_derivationStrict(RegisterPrimOp::Info { ‘out’. */ static void prim_placeholder(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - v.mkString(hashPlaceholder(state.forceStringNoCtx(*args[0], pos))); + v.mkString(hashPlaceholder(state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.placeholder"))); } static RegisterPrimOp primop_placeholder({ @@ -1392,7 +1421,7 @@ static RegisterPrimOp primop_placeholder({ static void prim_toPath(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - Path path = state.coerceToPath(pos, *args[0], context); + Path path = state.coerceToPath(pos, *args[0], context, "while evaluating the first argument passed to builtins.toPath"); v.mkString(canonPath(path), context); } @@ -1423,7 +1452,7 @@ static void prim_storePath(EvalState & state, const PosIdx pos, Value * * args, })); PathSet context; - Path path = state.checkSourcePath(state.coerceToPath(pos, *args[0], context)); + Path path = state.checkSourcePath(state.coerceToPath(pos, *args[0], context, "while evaluating the first argument passed to builtins.storePath")); /* Resolve symlinks in ‘path’, unless ‘path’ itself is a symlink directly in the store. The latter condition is necessary so e.g. nix-push does the right thing. */ @@ -1461,10 +1490,10 @@ static RegisterPrimOp primop_storePath({ static void prim_pathExists(EvalState & state, const PosIdx pos, Value * * args, Value & v) { /* We don’t check the path right now, because we don’t want to - throw if the path isn’t allowed, but just return false (and we - can’t just catch the exception here because we still want to - throw if something in the evaluation of `*args[0]` tries to - access an unauthorized path). */ + throw if the path isn’t allowed, but just return false (and we + can’t just catch the exception here because we still want to + throw if something in the evaluation of `*args[0]` tries to + access an unauthorized path). */ auto path = realisePath(state, pos, *args[0], { .checkForPureEval = false }); try { @@ -1493,7 +1522,9 @@ static RegisterPrimOp primop_pathExists({ static void prim_baseNameOf(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - v.mkString(baseNameOf(*state.coerceToString(pos, *args[0], context, false, false)), context); + v.mkString(baseNameOf(*state.coerceToString(pos, *args[0], context, + "while evaluating the first argument passed to builtins.baseNameOf", + false, false)), context); } static RegisterPrimOp primop_baseNameOf({ @@ -1513,7 +1544,9 @@ static RegisterPrimOp primop_baseNameOf({ static void prim_dirOf(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto path = state.coerceToString(pos, *args[0], context, false, false); + auto path = state.coerceToString(pos, *args[0], context, + "while evaluating the first argument passed to builtins.dirOf", + false, false); auto dir = dirOf(*path); if (args[0]->type() == nPath) v.mkPath(dir); else v.mkString(dir, context); } @@ -1542,6 +1575,10 @@ static void prim_readFile(EvalState & state, const PosIdx pos, Value * * args, V refs = state.store->queryPathInfo(state.store->toStorePath(path).first)->references; } catch (Error &) { // FIXME: should be InvalidPathError } + // Re-scan references to filter down to just the ones that actually occur in the file. + auto refsSink = PathRefScanSink::fromPaths(refs); + refsSink << s; + refs = refsSink.getResultPaths(); } auto context = state.store->printStorePathSet(refs); v.mkString(s, context); @@ -1560,28 +1597,24 @@ static RegisterPrimOp primop_readFile({ which are desugared to 'findFile __nixPath "x"'. */ static void prim_findFile(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[0], pos); + state.forceList(*args[0], pos, "while evaluating the first argument passed to builtins.findFile"); SearchPath searchPath; for (auto v2 : args[0]->listItems()) { - state.forceAttrs(*v2, pos); + state.forceAttrs(*v2, pos, "while evaluating an element of the list passed to builtins.findFile"); std::string prefix; Bindings::iterator i = v2->attrs->find(state.sPrefix); if (i != v2->attrs->end()) - prefix = state.forceStringNoCtx(*i->value, pos); + prefix = state.forceStringNoCtx(*i->value, pos, "while evaluating the `prefix` attribute of an element of the list passed to builtins.findFile"); - i = getAttr( - state, - "findFile", - state.sPath, - v2->attrs, - pos - ); + i = getAttr(state, state.sPath, v2->attrs, "in an element of the __nixPath"); PathSet context; - auto path = state.coerceToString(pos, *i->value, context, false, false).toOwned(); + auto path = state.coerceToString(pos, *i->value, context, + "while evaluating the `path` attribute of an element of the list passed to builtins.findFile", + false, false).toOwned(); try { auto rewrites = state.realiseContext(context); @@ -1596,7 +1629,7 @@ static void prim_findFile(EvalState & state, const PosIdx pos, Value * * args, V searchPath.emplace_back(prefix, path); } - auto path = state.forceStringNoCtx(*args[1], pos); + auto path = state.forceStringNoCtx(*args[1], pos, "while evaluating the second argument passed to builtins.findFile"); v.mkPath(state.checkSourcePath(state.findFile(searchPath, path, pos))); } @@ -1610,7 +1643,7 @@ static RegisterPrimOp primop_findFile(RegisterPrimOp::Info { /* Return the cryptographic hash of a file in base-16. */ static void prim_hashFile(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto type = state.forceStringNoCtx(*args[0], pos); + auto type = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.hashFile"); std::optional<HashType> ht = parseHashType(type); if (!ht) state.debugThrowLastTrace(Error({ @@ -1634,23 +1667,73 @@ static RegisterPrimOp primop_hashFile({ .fun = prim_hashFile, }); + +/* Stringize a directory entry enum. Used by `readFileType' and `readDir'. */ +static const char * dirEntTypeToString(unsigned char dtType) +{ + /* Enum DT_(DIR|LNK|REG|UNKNOWN) */ + switch(dtType) { + case DT_REG: return "regular"; break; + case DT_DIR: return "directory"; break; + case DT_LNK: return "symlink"; break; + default: return "unknown"; break; + } + return "unknown"; /* Unreachable */ +} + + +static void prim_readFileType(EvalState & state, const PosIdx pos, Value * * args, Value & v) +{ + auto path = realisePath(state, pos, *args[0]); + /* Retrieve the directory entry type and stringize it. */ + v.mkString(dirEntTypeToString(getFileType(path))); +} + +static RegisterPrimOp primop_readFileType({ + .name = "__readFileType", + .args = {"p"}, + .doc = R"( + Determine the directory entry type of a filesystem node, being + one of "directory", "regular", "symlink", or "unknown". + )", + .fun = prim_readFileType, +}); + /* Read a directory (without . or ..) */ static void prim_readDir(EvalState & state, const PosIdx pos, Value * * args, Value & v) { auto path = realisePath(state, pos, *args[0]); + // Retrieve directory entries for all nodes in a directory. + // This is similar to `getFileType` but is optimized to reduce system calls + // on many systems. DirEntries entries = readDirectory(path); auto attrs = state.buildBindings(entries.size()); + // If we hit unknown directory entry types we may need to fallback to + // using `getFileType` on some systems. + // In order to reduce system calls we make each lookup lazy by using + // `builtins.readFileType` application. + Value * readFileType = nullptr; + for (auto & ent : entries) { - if (ent.type == DT_UNKNOWN) - ent.type = getFileType(path + "/" + ent.name); - attrs.alloc(ent.name).mkString( - ent.type == DT_REG ? "regular" : - ent.type == DT_DIR ? "directory" : - ent.type == DT_LNK ? "symlink" : - "unknown"); + auto & attr = attrs.alloc(ent.name); + if (ent.type == DT_UNKNOWN) { + // Some filesystems or operating systems may not be able to return + // detailed node info quickly in this case we produce a thunk to + // query the file type lazily. + auto epath = state.allocValue(); + Path path2 = path + "/" + ent.name; + epath->mkString(path2); + if (!readFileType) + readFileType = &state.getBuiltin("readFileType"); + attr.mkApp(readFileType, epath); + } else { + // This branch of the conditional is much more likely. + // Here we just stringize the directory entry type. + attr.mkString(dirEntTypeToString(ent.type)); + } } v.mkAttrs(attrs); @@ -1817,7 +1900,7 @@ static RegisterPrimOp primop_toJSON({ /* Parse a JSON string to a value. */ static void prim_fromJSON(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto s = state.forceStringNoCtx(*args[0], pos); + auto s = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.fromJSON"); try { parseJSON(state, s, v); } catch (JSONParseError &e) { @@ -1846,8 +1929,8 @@ static RegisterPrimOp primop_fromJSON({ static void prim_toFile(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - std::string name(state.forceStringNoCtx(*args[0], pos)); - std::string contents(state.forceString(*args[1], context, pos)); + std::string name(state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.toFile")); + std::string contents(state.forceString(*args[1], context, pos, "while evaluating the second argument passed to builtins.toFile")); StorePathSet refs; @@ -1883,8 +1966,7 @@ static RegisterPrimOp primop_toFile({ path. The file has suffix *name*. This file can be used as an input to derivations. One application is to write builders “inline”. For instance, the following Nix expression combines the - [Nix expression for GNU Hello](expression-syntax.md) and its - [build script](build-script.md) into one file: + Nix expression for GNU Hello and its build script into one file: ```nix { stdenv, fetchurl, perl }: @@ -1927,8 +2009,8 @@ static RegisterPrimOp primop_toFile({ "; ``` - Note that `${configFile}` is an - [antiquotation](language-values.md), so the result of the + Note that `${configFile}` is a + [string interpolation](@docroot@/language/values.md#type-string), so the result of the expression `configFile` (i.e., a path like `/nix/store/m7p7jfny445k...-foo.conf`) will be spliced into the resulting string. @@ -2005,7 +2087,7 @@ static void addPath( Value res; state.callFunction(*filterFun, 2, args, res, pos); - return state.forceBool(res, pos); + return state.forceBool(res, pos, "while evaluating the return value of the path filter function"); }) : defaultPathFilter; std::optional<StorePath> expectedStorePath; @@ -2031,17 +2113,8 @@ static void addPath( static void prim_filterSource(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - Path path = state.coerceToPath(pos, *args[1], context); - - state.forceValue(*args[0], pos); - if (args[0]->type() != nFunction) - state.debugThrowLastTrace(TypeError({ - .msg = hintfmt( - "first argument in call to 'filterSource' is not a function but %1%", - showType(*args[0])), - .errPos = state.positions[pos] - })); - + Path path = state.coerceToPath(pos, *args[1], context, "while evaluating the second argument (the path to filter) passed to builtins.filterSource"); + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.filterSource"); addPath(state, pos, std::string(baseNameOf(path)), path, args[0], FileIngestionMethod::Recursive, std::nullopt, v, context); } @@ -2102,7 +2175,7 @@ static RegisterPrimOp primop_filterSource({ static void prim_path(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); + state.forceAttrs(*args[0], pos, "while evaluating the argument passed to builtins.path"); Path path; std::string name; Value * filterFun = nullptr; @@ -2113,16 +2186,15 @@ static void prim_path(EvalState & state, const PosIdx pos, Value * * args, Value for (auto & attr : *args[0]->attrs) { auto n = state.symbols[attr.name]; if (n == "path") - path = state.coerceToPath(attr.pos, *attr.value, context); + path = state.coerceToPath(attr.pos, *attr.value, context, "while evaluating the `path` attribute passed to builtins.path"); else if (attr.name == state.sName) - name = state.forceStringNoCtx(*attr.value, attr.pos); - else if (n == "filter") { - state.forceValue(*attr.value, pos); - filterFun = attr.value; - } else if (n == "recursive") - method = FileIngestionMethod { state.forceBool(*attr.value, attr.pos) }; + name = state.forceStringNoCtx(*attr.value, attr.pos, "while evaluating the `name` attribute passed to builtins.path"); + else if (n == "filter") + state.forceFunction(*(filterFun = attr.value), attr.pos, "while evaluating the `filter` parameter passed to builtins.path"); + else if (n == "recursive") + method = FileIngestionMethod { state.forceBool(*attr.value, attr.pos, "while evaluating the `recursive` attribute passed to builtins.path") }; else if (n == "sha256") - expectedHash = newHashAllowEmpty(state.forceStringNoCtx(*attr.value, attr.pos), htSHA256); + expectedHash = newHashAllowEmpty(state.forceStringNoCtx(*attr.value, attr.pos, "while evaluating the `sha256` attribute passed to builtins.path"), htSHA256); else state.debugThrowLastTrace(EvalError({ .msg = hintfmt("unsupported argument '%1%' to 'addPath'", state.symbols[attr.name]), @@ -2131,7 +2203,7 @@ static void prim_path(EvalState & state, const PosIdx pos, Value * * args, Value } if (path.empty()) state.debugThrowLastTrace(EvalError({ - .msg = hintfmt("'path' required"), + .msg = hintfmt("missing required 'path' attribute in the first argument to builtins.path"), .errPos = state.positions[pos] })); if (name.empty()) @@ -2185,7 +2257,7 @@ static RegisterPrimOp primop_path({ strings. */ static void prim_attrNames(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); + state.forceAttrs(*args[0], pos, "while evaluating the argument passed to builtins.attrNames"); state.mkList(v, args[0]->attrs->size()); @@ -2212,7 +2284,7 @@ static RegisterPrimOp primop_attrNames({ order as attrNames. */ static void prim_attrValues(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); + state.forceAttrs(*args[0], pos, "while evaluating the argument passed to builtins.attrValues"); state.mkList(v, args[0]->attrs->size()); @@ -2244,14 +2316,13 @@ static RegisterPrimOp primop_attrValues({ /* Dynamic version of the `.' operator. */ void prim_getAttr(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto attr = state.forceStringNoCtx(*args[0], pos); - state.forceAttrs(*args[1], pos); + auto attr = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.getAttr"); + state.forceAttrs(*args[1], pos, "while evaluating the second argument passed to builtins.getAttr"); Bindings::iterator i = getAttr( state, - "getAttr", state.symbols.create(attr), args[1]->attrs, - pos + "in the attribute set under consideration" ); // !!! add to stack trace? if (state.countCalls && i->pos) state.attrSelects[i->pos]++; @@ -2274,8 +2345,8 @@ static RegisterPrimOp primop_getAttr({ /* Return position information of the specified attribute. */ static void prim_unsafeGetAttrPos(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto attr = state.forceStringNoCtx(*args[0], pos); - state.forceAttrs(*args[1], pos); + auto attr = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.unsafeGetAttrPos"); + state.forceAttrs(*args[1], pos, "while evaluating the second argument passed to builtins.unsafeGetAttrPos"); Bindings::iterator i = args[1]->attrs->find(state.symbols.create(attr)); if (i == args[1]->attrs->end()) v.mkNull(); @@ -2292,8 +2363,8 @@ static RegisterPrimOp primop_unsafeGetAttrPos(RegisterPrimOp::Info { /* Dynamic version of the `?' operator. */ static void prim_hasAttr(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto attr = state.forceStringNoCtx(*args[0], pos); - state.forceAttrs(*args[1], pos); + auto attr = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.hasAttr"); + state.forceAttrs(*args[1], pos, "while evaluating the second argument passed to builtins.hasAttr"); v.mkBool(args[1]->attrs->find(state.symbols.create(attr)) != args[1]->attrs->end()); } @@ -2326,8 +2397,8 @@ static RegisterPrimOp primop_isAttrs({ static void prim_removeAttrs(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); - state.forceList(*args[1], pos); + state.forceAttrs(*args[0], pos, "while evaluating the first argument passed to builtins.removeAttrs"); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.removeAttrs"); /* Get the attribute names to be removed. We keep them as Attrs instead of Symbols so std::set_difference @@ -2335,7 +2406,7 @@ static void prim_removeAttrs(EvalState & state, const PosIdx pos, Value * * args boost::container::small_vector<Attr, 64> names; names.reserve(args[1]->listSize()); for (auto elem : args[1]->listItems()) { - state.forceStringNoCtx(*elem, pos); + state.forceStringNoCtx(*elem, pos, "while evaluating the values of the second argument passed to builtins.removeAttrs"); names.emplace_back(state.symbols.create(elem->string.s), nullptr); } std::sort(names.begin(), names.end()); @@ -2374,34 +2445,22 @@ static RegisterPrimOp primop_removeAttrs({ name, the first takes precedence. */ static void prim_listToAttrs(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[0], pos); + state.forceList(*args[0], pos, "while evaluating the argument passed to builtins.listToAttrs"); auto attrs = state.buildBindings(args[0]->listSize()); std::set<Symbol> seen; for (auto v2 : args[0]->listItems()) { - state.forceAttrs(*v2, pos); + state.forceAttrs(*v2, pos, "while evaluating an element of the list passed to builtins.listToAttrs"); - Bindings::iterator j = getAttr( - state, - "listToAttrs", - state.sName, - v2->attrs, - pos - ); + Bindings::iterator j = getAttr(state, state.sName, v2->attrs, "in a {name=...; value=...;} pair"); - auto name = state.forceStringNoCtx(*j->value, j->pos); + auto name = state.forceStringNoCtx(*j->value, j->pos, "while evaluating the `name` attribute of an element of the list passed to builtins.listToAttrs"); auto sym = state.symbols.create(name); if (seen.insert(sym).second) { - Bindings::iterator j2 = getAttr( - state, - "listToAttrs", - state.sValue, - v2->attrs, - pos - ); + Bindings::iterator j2 = getAttr(state, state.sValue, v2->attrs, "in a {name=...; value=...;} pair"); attrs.insert(sym, j2->value, j2->pos); } } @@ -2416,12 +2475,18 @@ static RegisterPrimOp primop_listToAttrs({ Construct a set from a list specifying the names and values of each attribute. Each element of the list should be a set consisting of a string-valued attribute `name` specifying the name of the attribute, - and an attribute `value` specifying its value. Example: + and an attribute `value` specifying its value. + + In case of duplicate occurrences of the same name, the first + takes precedence. + + Example: ```nix builtins.listToAttrs [ { name = "foo"; value = 123; } { name = "bar"; value = 456; } + { name = "bar"; value = 420; } ] ``` @@ -2436,15 +2501,65 @@ static RegisterPrimOp primop_listToAttrs({ static void prim_intersectAttrs(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); - state.forceAttrs(*args[1], pos); - - auto attrs = state.buildBindings(std::min(args[0]->attrs->size(), args[1]->attrs->size())); - - for (auto & i : *args[0]->attrs) { - Bindings::iterator j = args[1]->attrs->find(i.name); - if (j != args[1]->attrs->end()) - attrs.insert(*j); + state.forceAttrs(*args[0], pos, "while evaluating the first argument passed to builtins.intersectAttrs"); + state.forceAttrs(*args[1], pos, "while evaluating the second argument passed to builtins.intersectAttrs"); + + Bindings &left = *args[0]->attrs; + Bindings &right = *args[1]->attrs; + + auto attrs = state.buildBindings(std::min(left.size(), right.size())); + + // The current implementation has good asymptotic complexity and is reasonably + // simple. Further optimization may be possible, but does not seem productive, + // considering the state of eval performance in 2022. + // + // I have looked for reusable and/or standard solutions and these are my + // findings: + // + // STL + // === + // std::set_intersection is not suitable, as it only performs a simultaneous + // linear scan; not taking advantage of random access. This is O(n + m), so + // linear in the largest set, which is not acceptable for callPackage in Nixpkgs. + // + // Simultaneous scan, with alternating simple binary search + // === + // One alternative algorithm scans the attrsets simultaneously, jumping + // forward using `lower_bound` in case of inequality. This should perform + // well on very similar sets, having a local and predictable access pattern. + // On dissimilar sets, it seems to need more comparisons than the current + // algorithm, as few consecutive attrs match. `lower_bound` could take + // advantage of the decreasing remaining search space, but this causes + // the medians to move, which can mean that they don't stay in the cache + // like they would with the current naive `find`. + // + // Double binary search + // === + // The optimal algorithm may be "Double binary search", which doesn't + // scan at all, but rather divides both sets simultaneously. + // See "Fast Intersection Algorithms for Sorted Sequences" by Baeza-Yates et al. + // https://cs.uwaterloo.ca/~ajsaling/papers/intersection_alg_app10.pdf + // The only downsides I can think of are not having a linear access pattern + // for similar sets, and having to maintain a more intricate algorithm. + // + // Adaptive + // === + // Finally one could run try a simultaneous scan, count misses and fall back + // to double binary search when the counter hit some threshold and/or ratio. + + if (left.size() < right.size()) { + for (auto & l : left) { + Bindings::iterator r = right.find(l.name); + if (r != right.end()) + attrs.insert(*r); + } + } + else { + for (auto & r : right) { + Bindings::iterator l = left.find(r.name); + if (l != left.end()) + attrs.insert(r); + } } v.mkAttrs(attrs.alreadySorted()); @@ -2454,22 +2569,24 @@ static RegisterPrimOp primop_intersectAttrs({ .name = "__intersectAttrs", .args = {"e1", "e2"}, .doc = R"( - Return a set consisting of the attributes in the set *e2* that also - exist in the set *e1*. + Return a set consisting of the attributes in the set *e2* which have the + same name as some attribute in *e1*. + + Performs in O(*n* log *m*) where *n* is the size of the smaller set and *m* the larger set's size. )", .fun = prim_intersectAttrs, }); static void prim_catAttrs(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto attrName = state.symbols.create(state.forceStringNoCtx(*args[0], pos)); - state.forceList(*args[1], pos); + auto attrName = state.symbols.create(state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.catAttrs")); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.catAttrs"); Value * res[args[1]->listSize()]; unsigned int found = 0; for (auto v2 : args[1]->listItems()) { - state.forceAttrs(*v2, pos); + state.forceAttrs(*v2, pos, "while evaluating an element in the list passed as second argument to builtins.catAttrs"); Bindings::iterator i = v2->attrs->find(attrName); if (i != v2->attrs->end()) res[found++] = i->value; @@ -2542,7 +2659,7 @@ static RegisterPrimOp primop_functionArgs({ /* */ static void prim_mapAttrs(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[1], pos); + state.forceAttrs(*args[1], pos, "while evaluating the second argument passed to builtins.mapAttrs"); auto attrs = state.buildBindings(args[1]->attrs->size()); @@ -2583,21 +2700,16 @@ static void prim_zipAttrsWith(EvalState & state, const PosIdx pos, Value * * arg std::map<Symbol, std::pair<size_t, Value * *>> attrsSeen; - state.forceFunction(*args[0], pos); - state.forceList(*args[1], pos); + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.zipAttrsWith"); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.zipAttrsWith"); const auto listSize = args[1]->listSize(); const auto listElems = args[1]->listElems(); for (unsigned int n = 0; n < listSize; ++n) { Value * vElem = listElems[n]; - try { - state.forceAttrs(*vElem, noPos); - for (auto & attr : *vElem->attrs) - attrsSeen[attr.name].first++; - } catch (TypeError & e) { - e.addTrace(state.positions[pos], hintfmt("while invoking '%s'", "zipAttrsWith")); - state.debugThrowLastTrace(e); - } + state.forceAttrs(*vElem, noPos, "while evaluating a value of the list passed as second argument to builtins.zipAttrsWith"); + for (auto & attr : *vElem->attrs) + attrsSeen[attr.name].first++; } auto attrs = state.buildBindings(attrsSeen.size()); @@ -2681,7 +2793,7 @@ static RegisterPrimOp primop_isList({ static void elemAt(EvalState & state, const PosIdx pos, Value & list, int n, Value & v) { - state.forceList(list, pos); + state.forceList(list, pos, "while evaluating the first argument passed to builtins.elemAt"); if (n < 0 || (unsigned int) n >= list.listSize()) state.debugThrowLastTrace(Error({ .msg = hintfmt("list index %1% is out of bounds", n), @@ -2694,7 +2806,7 @@ static void elemAt(EvalState & state, const PosIdx pos, Value & list, int n, Val /* Return the n-1'th element of a list. */ static void prim_elemAt(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - elemAt(state, pos, *args[0], state.forceInt(*args[1], pos), v); + elemAt(state, pos, *args[0], state.forceInt(*args[1], pos, "while evaluating the second argument passed to builtins.elemAt"), v); } static RegisterPrimOp primop_elemAt({ @@ -2729,7 +2841,7 @@ static RegisterPrimOp primop_head({ don't want to use it! */ static void prim_tail(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[0], pos); + state.forceList(*args[0], pos, "while evaluating the first argument passed to builtins.tail"); if (args[0]->listSize() == 0) state.debugThrowLastTrace(Error({ .msg = hintfmt("'tail' called on an empty list"), @@ -2760,10 +2872,16 @@ static RegisterPrimOp primop_tail({ /* Apply a function to every element of a list. */ static void prim_map(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[1], pos); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.map"); - state.mkList(v, args[1]->listSize()); + if (args[1]->listSize() == 0) { + v = *args[1]; + return; + } + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.map"); + + state.mkList(v, args[1]->listSize()); for (unsigned int n = 0; n < v.listSize(); ++n) (v.listElems()[n] = state.allocValue())->mkApp( args[0], args[1]->listElems()[n]); @@ -2790,8 +2908,14 @@ static RegisterPrimOp primop_map({ returns true. */ static void prim_filter(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceFunction(*args[0], pos); - state.forceList(*args[1], pos); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.filter"); + + if (args[1]->listSize() == 0) { + v = *args[1]; + return; + } + + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.filter"); // FIXME: putting this on the stack is risky. Value * vs[args[1]->listSize()]; @@ -2801,7 +2925,7 @@ static void prim_filter(EvalState & state, const PosIdx pos, Value * * args, Val for (unsigned int n = 0; n < args[1]->listSize(); ++n) { Value res; state.callFunction(*args[0], *args[1]->listElems()[n], res, noPos); - if (state.forceBool(res, pos)) + if (state.forceBool(res, pos, "while evaluating the return value of the filtering function passed to builtins.filter")) vs[k++] = args[1]->listElems()[n]; else same = false; @@ -2829,9 +2953,9 @@ static RegisterPrimOp primop_filter({ static void prim_elem(EvalState & state, const PosIdx pos, Value * * args, Value & v) { bool res = false; - state.forceList(*args[1], pos); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.elem"); for (auto elem : args[1]->listItems()) - if (state.eqValues(*args[0], *elem)) { + if (state.eqValues(*args[0], *elem, pos, "while searching for the presence of the given element in the list")) { res = true; break; } @@ -2851,8 +2975,8 @@ static RegisterPrimOp primop_elem({ /* Concatenate a list of lists. */ static void prim_concatLists(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[0], pos); - state.concatLists(v, args[0]->listSize(), args[0]->listElems(), pos); + state.forceList(*args[0], pos, "while evaluating the first argument passed to builtins.concatLists"); + state.concatLists(v, args[0]->listSize(), args[0]->listElems(), pos, "while evaluating a value of the list passed to builtins.concatLists"); } static RegisterPrimOp primop_concatLists({ @@ -2867,7 +2991,7 @@ static RegisterPrimOp primop_concatLists({ /* Return the length of a list. This is an O(1) time operation. */ static void prim_length(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[0], pos); + state.forceList(*args[0], pos, "while evaluating the first argument passed to builtins.length"); v.mkInt(args[0]->listSize()); } @@ -2884,8 +3008,8 @@ static RegisterPrimOp primop_length({ right. The operator is applied strictly. */ static void prim_foldlStrict(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceFunction(*args[0], pos); - state.forceList(*args[2], pos); + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.foldlStrict"); + state.forceList(*args[2], pos, "while evaluating the third argument passed to builtins.foldlStrict"); if (args[2]->listSize()) { Value * vCur = args[1]; @@ -2908,22 +3032,22 @@ static RegisterPrimOp primop_foldlStrict({ .doc = R"( Reduce a list by applying a binary operator, from left to right, e.g. `foldl' op nul [x0 x1 x2 ...] = op (op (op nul x0) x1) x2) - ...`. The operator is applied strictly, i.e., its arguments are - evaluated first. For example, `foldl' (x: y: x + y) 0 [1 2 3]` - evaluates to 6. + ...`. For example, `foldl' (x: y: x + y) 0 [1 2 3]` evaluates to 6. + The return value of each application of `op` is evaluated immediately, + even for intermediate values. )", .fun = prim_foldlStrict, }); static void anyOrAll(bool any, EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceFunction(*args[0], pos); - state.forceList(*args[1], pos); + state.forceFunction(*args[0], pos, std::string("while evaluating the first argument passed to builtins.") + (any ? "any" : "all")); + state.forceList(*args[1], pos, std::string("while evaluating the second argument passed to builtins.") + (any ? "any" : "all")); Value vTmp; for (auto elem : args[1]->listItems()) { state.callFunction(*args[0], *elem, vTmp, pos); - bool res = state.forceBool(vTmp, pos); + bool res = state.forceBool(vTmp, pos, std::string("while evaluating the return value of the function passed to builtins.") + (any ? "any" : "all")); if (res == any) { v.mkBool(any); return; @@ -2966,16 +3090,16 @@ static RegisterPrimOp primop_all({ static void prim_genList(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto len = state.forceInt(*args[1], pos); + auto len = state.forceInt(*args[1], pos, "while evaluating the second argument passed to builtins.genList"); if (len < 0) - state.debugThrowLastTrace(EvalError({ - .msg = hintfmt("cannot create list of size %1%", len), - .errPos = state.positions[pos] - })); + state.error("cannot create list of size %1%", len).debugThrow<EvalError>(); - state.mkList(v, len); + // More strict than striclty (!) necessary, but acceptable + // as evaluating map without accessing any values makes little sense. + state.forceFunction(*args[0], noPos, "while evaluating the first argument passed to builtins.genList"); + state.mkList(v, len); for (unsigned int n = 0; n < (unsigned int) len; ++n) { auto arg = state.allocValue(); arg->mkInt(n); @@ -3004,10 +3128,16 @@ static void prim_lessThan(EvalState & state, const PosIdx pos, Value * * args, V static void prim_sort(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceFunction(*args[0], pos); - state.forceList(*args[1], pos); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.sort"); auto len = args[1]->listSize(); + if (len == 0) { + v = *args[1]; + return; + } + + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.sort"); + state.mkList(v, len); for (unsigned int n = 0; n < len; ++n) { state.forceValue(*args[1]->listElems()[n], pos); @@ -3017,13 +3147,15 @@ static void prim_sort(EvalState & state, const PosIdx pos, Value * * args, Value auto comparator = [&](Value * a, Value * b) { /* Optimization: if the comparator is lessThan, bypass callFunction. */ + /* TODO: (layus) this is absurd. An optimisation like this + should be outside the lambda creation */ if (args[0]->isPrimOp() && args[0]->primOp->fun == prim_lessThan) - return CompareValues(state)(a, b); + return CompareValues(state, noPos, "while evaluating the ordering function passed to builtins.sort")(a, b); Value * vs[] = {a, b}; Value vBool; - state.callFunction(*args[0], 2, vs, vBool, pos); - return state.forceBool(vBool, pos); + state.callFunction(*args[0], 2, vs, vBool, noPos); + return state.forceBool(vBool, pos, "while evaluating the return value of the sorting function passed to builtins.sort"); }; /* FIXME: std::sort can segfault if the comparator is not a strict @@ -3055,8 +3187,8 @@ static RegisterPrimOp primop_sort({ static void prim_partition(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceFunction(*args[0], pos); - state.forceList(*args[1], pos); + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.partition"); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.partition"); auto len = args[1]->listSize(); @@ -3067,7 +3199,7 @@ static void prim_partition(EvalState & state, const PosIdx pos, Value * * args, state.forceValue(*vElem, pos); Value res; state.callFunction(*args[0], *vElem, res, pos); - if (state.forceBool(res, pos)) + if (state.forceBool(res, pos, "while evaluating the return value of the partition function passed to builtins.partition")) right.push_back(vElem); else wrong.push_back(vElem); @@ -3115,15 +3247,15 @@ static RegisterPrimOp primop_partition({ static void prim_groupBy(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceFunction(*args[0], pos); - state.forceList(*args[1], pos); + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.groupBy"); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.groupBy"); ValueVectorMap attrs; for (auto vElem : args[1]->listItems()) { Value res; state.callFunction(*args[0], *vElem, res, pos); - auto name = state.forceStringNoCtx(res, pos); + auto name = state.forceStringNoCtx(res, pos, "while evaluating the return value of the grouping function passed to builtins.groupBy"); auto sym = state.symbols.create(name); auto vector = attrs.try_emplace(sym, ValueVector()).first; vector->second.push_back(vElem); @@ -3167,8 +3299,8 @@ static RegisterPrimOp primop_groupBy({ static void prim_concatMap(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceFunction(*args[0], pos); - state.forceList(*args[1], pos); + state.forceFunction(*args[0], pos, "while evaluating the first argument passed to builtins.concatMap"); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.concatMap"); auto nrLists = args[1]->listSize(); Value lists[nrLists]; @@ -3177,12 +3309,7 @@ static void prim_concatMap(EvalState & state, const PosIdx pos, Value * * args, for (unsigned int n = 0; n < nrLists; ++n) { Value * vElem = args[1]->listElems()[n]; state.callFunction(*args[0], *vElem, lists[n], pos); - try { - state.forceList(lists[n], lists[n].determinePos(args[0]->determinePos(pos))); - } catch (TypeError &e) { - e.addTrace(state.positions[pos], hintfmt("while invoking '%s'", "concatMap")); - state.debugThrowLastTrace(e); - } + state.forceList(lists[n], lists[n].determinePos(args[0]->determinePos(pos)), "while evaluating the return value of the function passed to buitlins.concatMap"); len += lists[n].listSize(); } @@ -3217,9 +3344,11 @@ static void prim_add(EvalState & state, const PosIdx pos, Value * * args, Value state.forceValue(*args[0], pos); state.forceValue(*args[1], pos); if (args[0]->type() == nFloat || args[1]->type() == nFloat) - v.mkFloat(state.forceFloat(*args[0], pos) + state.forceFloat(*args[1], pos)); + v.mkFloat(state.forceFloat(*args[0], pos, "while evaluating the first argument of the addition") + + state.forceFloat(*args[1], pos, "while evaluating the second argument of the addition")); else - v.mkInt(state.forceInt(*args[0], pos) + state.forceInt(*args[1], pos)); + v.mkInt( state.forceInt(*args[0], pos, "while evaluating the first argument of the addition") + + state.forceInt(*args[1], pos, "while evaluating the second argument of the addition")); } static RegisterPrimOp primop_add({ @@ -3236,9 +3365,11 @@ static void prim_sub(EvalState & state, const PosIdx pos, Value * * args, Value state.forceValue(*args[0], pos); state.forceValue(*args[1], pos); if (args[0]->type() == nFloat || args[1]->type() == nFloat) - v.mkFloat(state.forceFloat(*args[0], pos) - state.forceFloat(*args[1], pos)); + v.mkFloat(state.forceFloat(*args[0], pos, "while evaluating the first argument of the subtraction") + - state.forceFloat(*args[1], pos, "while evaluating the second argument of the subtraction")); else - v.mkInt(state.forceInt(*args[0], pos) - state.forceInt(*args[1], pos)); + v.mkInt( state.forceInt(*args[0], pos, "while evaluating the first argument of the subtraction") + - state.forceInt(*args[1], pos, "while evaluating the second argument of the subtraction")); } static RegisterPrimOp primop_sub({ @@ -3255,9 +3386,11 @@ static void prim_mul(EvalState & state, const PosIdx pos, Value * * args, Value state.forceValue(*args[0], pos); state.forceValue(*args[1], pos); if (args[0]->type() == nFloat || args[1]->type() == nFloat) - v.mkFloat(state.forceFloat(*args[0], pos) * state.forceFloat(*args[1], pos)); + v.mkFloat(state.forceFloat(*args[0], pos, "while evaluating the first of the multiplication") + * state.forceFloat(*args[1], pos, "while evaluating the second argument of the multiplication")); else - v.mkInt(state.forceInt(*args[0], pos) * state.forceInt(*args[1], pos)); + v.mkInt( state.forceInt(*args[0], pos, "while evaluating the first argument of the multiplication") + * state.forceInt(*args[1], pos, "while evaluating the second argument of the multiplication")); } static RegisterPrimOp primop_mul({ @@ -3274,7 +3407,7 @@ static void prim_div(EvalState & state, const PosIdx pos, Value * * args, Value state.forceValue(*args[0], pos); state.forceValue(*args[1], pos); - NixFloat f2 = state.forceFloat(*args[1], pos); + NixFloat f2 = state.forceFloat(*args[1], pos, "while evaluating the second operand of the division"); if (f2 == 0) state.debugThrowLastTrace(EvalError({ .msg = hintfmt("division by zero"), @@ -3282,10 +3415,10 @@ static void prim_div(EvalState & state, const PosIdx pos, Value * * args, Value })); if (args[0]->type() == nFloat || args[1]->type() == nFloat) { - v.mkFloat(state.forceFloat(*args[0], pos) / state.forceFloat(*args[1], pos)); + v.mkFloat(state.forceFloat(*args[0], pos, "while evaluating the first operand of the division") / f2); } else { - NixInt i1 = state.forceInt(*args[0], pos); - NixInt i2 = state.forceInt(*args[1], pos); + NixInt i1 = state.forceInt(*args[0], pos, "while evaluating the first operand of the division"); + NixInt i2 = state.forceInt(*args[1], pos, "while evaluating the second operand of the division"); /* Avoid division overflow as it might raise SIGFPE. */ if (i1 == std::numeric_limits<NixInt>::min() && i2 == -1) state.debugThrowLastTrace(EvalError({ @@ -3308,7 +3441,8 @@ static RegisterPrimOp primop_div({ static void prim_bitAnd(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - v.mkInt(state.forceInt(*args[0], pos) & state.forceInt(*args[1], pos)); + v.mkInt(state.forceInt(*args[0], pos, "while evaluating the first argument passed to builtins.bitAnd") + & state.forceInt(*args[1], pos, "while evaluating the second argument passed to builtins.bitAnd")); } static RegisterPrimOp primop_bitAnd({ @@ -3322,7 +3456,8 @@ static RegisterPrimOp primop_bitAnd({ static void prim_bitOr(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - v.mkInt(state.forceInt(*args[0], pos) | state.forceInt(*args[1], pos)); + v.mkInt(state.forceInt(*args[0], pos, "while evaluating the first argument passed to builtins.bitOr") + | state.forceInt(*args[1], pos, "while evaluating the second argument passed to builtins.bitOr")); } static RegisterPrimOp primop_bitOr({ @@ -3336,7 +3471,8 @@ static RegisterPrimOp primop_bitOr({ static void prim_bitXor(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - v.mkInt(state.forceInt(*args[0], pos) ^ state.forceInt(*args[1], pos)); + v.mkInt(state.forceInt(*args[0], pos, "while evaluating the first argument passed to builtins.bitXor") + ^ state.forceInt(*args[1], pos, "while evaluating the second argument passed to builtins.bitXor")); } static RegisterPrimOp primop_bitXor({ @@ -3352,7 +3488,8 @@ static void prim_lessThan(EvalState & state, const PosIdx pos, Value * * args, V { state.forceValue(*args[0], pos); state.forceValue(*args[1], pos); - CompareValues comp{state}; + // pos is exact here, no need for a message. + CompareValues comp(state, noPos, ""); v.mkBool(comp(args[0], args[1])); } @@ -3379,7 +3516,9 @@ static RegisterPrimOp primop_lessThan({ static void prim_toString(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto s = state.coerceToString(pos, *args[0], context, true, false); + auto s = state.coerceToString(pos, *args[0], context, + "while evaluating the first argument passed to builtins.toString", + true, false); v.mkString(*s, context); } @@ -3413,10 +3552,10 @@ static RegisterPrimOp primop_toString({ non-negative. */ static void prim_substring(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - int start = state.forceInt(*args[0], pos); - int len = state.forceInt(*args[1], pos); + int start = state.forceInt(*args[0], pos, "while evaluating the first argument (the start offset) passed to builtins.substring"); + int len = state.forceInt(*args[1], pos, "while evaluating the second argument (the substring length) passed to builtins.substring"); PathSet context; - auto s = state.coerceToString(pos, *args[2], context); + auto s = state.coerceToString(pos, *args[2], context, "while evaluating the third argument (the string) passed to builtins.substring"); if (start < 0) state.debugThrowLastTrace(EvalError({ @@ -3450,7 +3589,7 @@ static RegisterPrimOp primop_substring({ static void prim_stringLength(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto s = state.coerceToString(pos, *args[0], context); + auto s = state.coerceToString(pos, *args[0], context, "while evaluating the argument passed to builtins.stringLength"); v.mkInt(s->size()); } @@ -3467,7 +3606,7 @@ static RegisterPrimOp primop_stringLength({ /* Return the cryptographic hash of a string in base-16. */ static void prim_hashString(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto type = state.forceStringNoCtx(*args[0], pos); + auto type = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.hashString"); std::optional<HashType> ht = parseHashType(type); if (!ht) state.debugThrowLastTrace(Error({ @@ -3476,7 +3615,7 @@ static void prim_hashString(EvalState & state, const PosIdx pos, Value * * args, })); PathSet context; // discarded - auto s = state.forceString(*args[1], context, pos); + auto s = state.forceString(*args[1], context, pos, "while evaluating the second argument passed to builtins.hashString"); v.mkString(hashString(*ht, s).to_string(Base16, false)); } @@ -3515,14 +3654,14 @@ std::shared_ptr<RegexCache> makeRegexCache() void prim_match(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto re = state.forceStringNoCtx(*args[0], pos); + auto re = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.match"); try { auto regex = state.regexCache->get(re); PathSet context; - const auto str = state.forceString(*args[1], context, pos); + const auto str = state.forceString(*args[1], context, pos, "while evaluating the second argument passed to builtins.match"); std::cmatch match; if (!std::regex_match(str.begin(), str.end(), match, regex)) { @@ -3595,14 +3734,14 @@ static RegisterPrimOp primop_match({ non-matching parts interleaved by the lists of the matching groups. */ void prim_split(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto re = state.forceStringNoCtx(*args[0], pos); + auto re = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.split"); try { auto regex = state.regexCache->get(re); PathSet context; - const auto str = state.forceString(*args[1], context, pos); + const auto str = state.forceString(*args[1], context, pos, "while evaluating the second argument passed to builtins.split"); auto begin = std::cregex_iterator(str.begin(), str.end(), regex); auto end = std::cregex_iterator(); @@ -3700,8 +3839,8 @@ static void prim_concatStringsSep(EvalState & state, const PosIdx pos, Value * * { PathSet context; - auto sep = state.forceString(*args[0], context, pos); - state.forceList(*args[1], pos); + auto sep = state.forceString(*args[0], context, pos, "while evaluating the first argument (the separator string) passed to builtins.concatStringsSep"); + state.forceList(*args[1], pos, "while evaluating the second argument (the list of strings to concat) passed to builtins.concatStringsSep"); std::string res; res.reserve((args[1]->listSize() + 32) * sep.size()); @@ -3709,7 +3848,7 @@ static void prim_concatStringsSep(EvalState & state, const PosIdx pos, Value * * for (auto elem : args[1]->listItems()) { if (first) first = false; else res += sep; - res += *state.coerceToString(pos, *elem, context); + res += *state.coerceToString(pos, *elem, context, "while evaluating one element of the list of strings to concat passed to builtins.concatStringsSep"); } v.mkString(res, context); @@ -3728,29 +3867,26 @@ static RegisterPrimOp primop_concatStringsSep({ static void prim_replaceStrings(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceList(*args[0], pos); - state.forceList(*args[1], pos); + state.forceList(*args[0], pos, "while evaluating the first argument passed to builtins.replaceStrings"); + state.forceList(*args[1], pos, "while evaluating the second argument passed to builtins.replaceStrings"); if (args[0]->listSize() != args[1]->listSize()) - state.debugThrowLastTrace(EvalError({ - .msg = hintfmt("'from' and 'to' arguments to 'replaceStrings' have different lengths"), - .errPos = state.positions[pos] - })); + state.error("'from' and 'to' arguments passed to builtins.replaceStrings have different lengths").atPos(pos).debugThrow<EvalError>(); std::vector<std::string> from; from.reserve(args[0]->listSize()); for (auto elem : args[0]->listItems()) - from.emplace_back(state.forceString(*elem, pos)); + from.emplace_back(state.forceString(*elem, pos, "while evaluating one of the strings to replace passed to builtins.replaceStrings")); std::vector<std::pair<std::string, PathSet>> to; to.reserve(args[1]->listSize()); for (auto elem : args[1]->listItems()) { PathSet ctx; - auto s = state.forceString(*elem, ctx, pos); + auto s = state.forceString(*elem, ctx, pos, "while evaluating one of the replacement strings passed to builtins.replaceStrings"); to.emplace_back(s, std::move(ctx)); } PathSet context; - auto s = state.forceString(*args[2], context, pos); + auto s = state.forceString(*args[2], context, pos, "while evaluating the third argument passed to builtins.replaceStrings"); std::string res; // Loops one past last character to handle the case where 'from' contains an empty string. @@ -3808,7 +3944,7 @@ static RegisterPrimOp primop_replaceStrings({ static void prim_parseDrvName(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto name = state.forceStringNoCtx(*args[0], pos); + auto name = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.parseDrvName"); DrvName parsed(name); auto attrs = state.buildBindings(2); attrs.alloc(state.sName).mkString(parsed.name); @@ -3821,8 +3957,8 @@ static RegisterPrimOp primop_parseDrvName({ .args = {"s"}, .doc = R"( Split the string *s* into a package name and version. The package - name is everything up to but not including the first dash followed - by a digit, and the version is everything following that dash. The + name is everything up to but not including the first dash not followed + by a letter, and the version is everything following that dash. The result is returned in a set `{ name, version }`. Thus, `builtins.parseDrvName "nix-0.12pre12876"` returns `{ name = "nix"; version = "0.12pre12876"; }`. @@ -3832,8 +3968,8 @@ static RegisterPrimOp primop_parseDrvName({ static void prim_compareVersions(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto version1 = state.forceStringNoCtx(*args[0], pos); - auto version2 = state.forceStringNoCtx(*args[1], pos); + auto version1 = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.compareVersions"); + auto version2 = state.forceStringNoCtx(*args[1], pos, "while evaluating the second argument passed to builtins.compareVersions"); v.mkInt(compareVersions(version1, version2)); } @@ -3852,7 +3988,7 @@ static RegisterPrimOp primop_compareVersions({ static void prim_splitVersion(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - auto version = state.forceStringNoCtx(*args[0], pos); + auto version = state.forceStringNoCtx(*args[0], pos, "while evaluating the first argument passed to builtins.splitVersion"); auto iter = version.cbegin(); Strings components; while (iter != version.cend()) { @@ -4008,7 +4144,7 @@ void EvalState::createBaseEnv() // the parser needs two NUL bytes as terminators; one of them // is implied by being a C string. "\0"; - eval(parse(code, sizeof(code), foFile, derivationNixPath, "/", staticBaseEnv), *vDerivation); + eval(parse(code, sizeof(code), derivationNixPath, "/", staticBaseEnv), *vDerivation); } diff --git a/src/libexpr/primops/context.cc b/src/libexpr/primops/context.cc index 979136984..db43e5771 100644 --- a/src/libexpr/primops/context.cc +++ b/src/libexpr/primops/context.cc @@ -8,7 +8,7 @@ namespace nix { static void prim_unsafeDiscardStringContext(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto s = state.coerceToString(pos, *args[0], context); + auto s = state.coerceToString(pos, *args[0], context, "while evaluating the argument passed to builtins.unsafeDiscardStringContext"); v.mkString(*s); } @@ -18,7 +18,7 @@ static RegisterPrimOp primop_unsafeDiscardStringContext("__unsafeDiscardStringCo static void prim_hasContext(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - state.forceString(*args[0], context, pos); + state.forceString(*args[0], context, pos, "while evaluating the argument passed to builtins.hasContext"); v.mkBool(!context.empty()); } @@ -34,11 +34,18 @@ static RegisterPrimOp primop_hasContext("__hasContext", 1, prim_hasContext); static void prim_unsafeDiscardOutputDependency(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto s = state.coerceToString(pos, *args[0], context); + auto s = state.coerceToString(pos, *args[0], context, "while evaluating the argument passed to builtins.unsafeDiscardOutputDependency"); PathSet context2; - for (auto & p : context) - context2.insert(p.at(0) == '=' ? std::string(p, 1) : p); + for (auto && p : context) { + auto c = NixStringContextElem::parse(*state.store, p); + if (auto * ptr = std::get_if<NixStringContextElem::DrvDeep>(&c)) { + context2.emplace(state.store->printStorePath(ptr->drvPath)); + } else { + /* Can reuse original item */ + context2.emplace(std::move(p)); + } + } v.mkString(*s, context2); } @@ -73,35 +80,21 @@ static void prim_getContext(EvalState & state, const PosIdx pos, Value * * args, Strings outputs; }; PathSet context; - state.forceString(*args[0], context, pos); - auto contextInfos = std::map<Path, ContextInfo>(); + state.forceString(*args[0], context, pos, "while evaluating the argument passed to builtins.getContext"); + auto contextInfos = std::map<StorePath, ContextInfo>(); for (const auto & p : context) { - Path drv; - std::string output; - const Path * path = &p; - if (p.at(0) == '=') { - drv = std::string(p, 1); - path = &drv; - } else if (p.at(0) == '!') { - NixStringContextElem ctx = decodeContext(*state.store, p); - drv = state.store->printStorePath(ctx.first); - output = ctx.second; - path = &drv; - } - auto isPath = drv.empty(); - auto isAllOutputs = (!drv.empty()) && output.empty(); - - auto iter = contextInfos.find(*path); - if (iter == contextInfos.end()) { - contextInfos.emplace(*path, ContextInfo{isPath, isAllOutputs, output.empty() ? Strings{} : Strings{std::move(output)}}); - } else { - if (isPath) - iter->second.path = true; - else if (isAllOutputs) - iter->second.allOutputs = true; - else - iter->second.outputs.emplace_back(std::move(output)); - } + NixStringContextElem ctx = NixStringContextElem::parse(*state.store, p); + std::visit(overloaded { + [&](NixStringContextElem::DrvDeep & d) { + contextInfos[d.drvPath].allOutputs = true; + }, + [&](NixStringContextElem::Built & b) { + contextInfos[b.drvPath].outputs.emplace_back(std::move(b.output)); + }, + [&](NixStringContextElem::Opaque & o) { + contextInfos[o.path].path = true; + }, + }, ctx.raw()); } auto attrs = state.buildBindings(contextInfos.size()); @@ -120,7 +113,7 @@ static void prim_getContext(EvalState & state, const PosIdx pos, Value * * args, for (const auto & [i, output] : enumerate(info.second.outputs)) (outputsVal.listElems()[i] = state.allocValue())->mkString(output); } - attrs.alloc(info.first).mkAttrs(infoAttrs); + attrs.alloc(state.store->printStorePath(info.first)).mkAttrs(infoAttrs); } v.mkAttrs(attrs); @@ -137,9 +130,9 @@ static RegisterPrimOp primop_getContext("__getContext", 1, prim_getContext); static void prim_appendContext(EvalState & state, const PosIdx pos, Value * * args, Value & v) { PathSet context; - auto orig = state.forceString(*args[0], context, pos); + auto orig = state.forceString(*args[0], context, noPos, "while evaluating the first argument passed to builtins.appendContext"); - state.forceAttrs(*args[1], pos); + state.forceAttrs(*args[1], pos, "while evaluating the second argument passed to builtins.appendContext"); auto sPath = state.symbols.create("path"); auto sAllOutputs = state.symbols.create("allOutputs"); @@ -147,24 +140,24 @@ static void prim_appendContext(EvalState & state, const PosIdx pos, Value * * ar const auto & name = state.symbols[i.name]; if (!state.store->isStorePath(name)) throw EvalError({ - .msg = hintfmt("Context key '%s' is not a store path", name), + .msg = hintfmt("context key '%s' is not a store path", name), .errPos = state.positions[i.pos] }); if (!settings.readOnlyMode) state.store->ensurePath(state.store->parseStorePath(name)); - state.forceAttrs(*i.value, i.pos); + state.forceAttrs(*i.value, i.pos, "while evaluating the value of a string context"); auto iter = i.value->attrs->find(sPath); if (iter != i.value->attrs->end()) { - if (state.forceBool(*iter->value, iter->pos)) + if (state.forceBool(*iter->value, iter->pos, "while evaluating the `path` attribute of a string context")) context.emplace(name); } iter = i.value->attrs->find(sAllOutputs); if (iter != i.value->attrs->end()) { - if (state.forceBool(*iter->value, iter->pos)) { + if (state.forceBool(*iter->value, iter->pos, "while evaluating the `allOutputs` attribute of a string context")) { if (!isDerivation(name)) { throw EvalError({ - .msg = hintfmt("Tried to add all-outputs context of %s, which is not a derivation, to a string", name), + .msg = hintfmt("tried to add all-outputs context of %s, which is not a derivation, to a string", name), .errPos = state.positions[i.pos] }); } @@ -174,15 +167,15 @@ static void prim_appendContext(EvalState & state, const PosIdx pos, Value * * ar iter = i.value->attrs->find(state.sOutputs); if (iter != i.value->attrs->end()) { - state.forceList(*iter->value, iter->pos); + state.forceList(*iter->value, iter->pos, "while evaluating the `outputs` attribute of a string context"); if (iter->value->listSize() && !isDerivation(name)) { throw EvalError({ - .msg = hintfmt("Tried to add derivation output context of %s, which is not a derivation, to a string", name), + .msg = hintfmt("tried to add derivation output context of %s, which is not a derivation, to a string", name), .errPos = state.positions[i.pos] }); } for (auto elem : iter->value->listItems()) { - auto outputName = state.forceStringNoCtx(*elem, iter->pos); + auto outputName = state.forceStringNoCtx(*elem, iter->pos, "while evaluating an output name within a string context"); context.insert(concatStrings("!", outputName, "!", name)); } } diff --git a/src/libexpr/primops/fetchClosure.cc b/src/libexpr/primops/fetchClosure.cc index 662c9652e..0dfa97fa3 100644 --- a/src/libexpr/primops/fetchClosure.cc +++ b/src/libexpr/primops/fetchClosure.cc @@ -7,7 +7,7 @@ namespace nix { static void prim_fetchClosure(EvalState & state, const PosIdx pos, Value * * args, Value & v) { - state.forceAttrs(*args[0], pos); + state.forceAttrs(*args[0], pos, "while evaluating the argument passed to builtins.fetchClosure"); std::optional<std::string> fromStoreUrl; std::optional<StorePath> fromPath; @@ -19,7 +19,8 @@ static void prim_fetchClosure(EvalState & state, const PosIdx pos, Value * * arg if (attrName == "fromPath") { PathSet context; - fromPath = state.coerceToStorePath(attr.pos, *attr.value, context); + fromPath = state.coerceToStorePath(attr.pos, *attr.value, context, + "while evaluating the 'fromPath' attribute passed to builtins.fetchClosure"); } else if (attrName == "toPath") { @@ -27,12 +28,14 @@ static void prim_fetchClosure(EvalState & state, const PosIdx pos, Value * * arg toCA = true; if (attr.value->type() != nString || attr.value->string.s != std::string("")) { PathSet context; - toPath = state.coerceToStorePath(attr.pos, *attr.value, context); + toPath = state.coerceToStorePath(attr.pos, *attr.value, context, + "while evaluating the 'toPath' attribute passed to builtins.fetchClosure"); } } else if (attrName == "fromStore") - fromStoreUrl = state.forceStringNoCtx(*attr.value, attr.pos); + fromStoreUrl = state.forceStringNoCtx(*attr.value, attr.pos, + "while evaluating the 'fromStore' attribute passed to builtins.fetchClosure"); else throw Error({ diff --git a/src/libexpr/primops/fetchMercurial.cc b/src/libexpr/primops/fetchMercurial.cc index 249c0934e..c41bd60b6 100644 --- a/src/libexpr/primops/fetchMercurial.cc +++ b/src/libexpr/primops/fetchMercurial.cc @@ -19,23 +19,23 @@ static void prim_fetchMercurial(EvalState & state, const PosIdx pos, Value * * a if (args[0]->type() == nAttrs) { - state.forceAttrs(*args[0], pos); - for (auto & attr : *args[0]->attrs) { std::string_view n(state.symbols[attr.name]); if (n == "url") - url = state.coerceToString(attr.pos, *attr.value, context, false, false).toOwned(); + url = state.coerceToString(attr.pos, *attr.value, context, + "while evaluating the `url` attribute passed to builtins.fetchMercurial", + false, false).toOwned(); else if (n == "rev") { // Ugly: unlike fetchGit, here the "rev" attribute can // be both a revision or a branch/tag name. - auto value = state.forceStringNoCtx(*attr.value, attr.pos); + auto value = state.forceStringNoCtx(*attr.value, attr.pos, "while evaluating the `rev` attribute passed to builtins.fetchMercurial"); if (std::regex_match(value.begin(), value.end(), revRegex)) rev = Hash::parseAny(value, htSHA1); else ref = value; } else if (n == "name") - name = state.forceStringNoCtx(*attr.value, attr.pos); + name = state.forceStringNoCtx(*attr.value, attr.pos, "while evaluating the `name` attribute passed to builtins.fetchMercurial"); else throw EvalError({ .msg = hintfmt("unsupported argument '%s' to 'fetchMercurial'", state.symbols[attr.name]), @@ -50,7 +50,9 @@ static void prim_fetchMercurial(EvalState & state, const PosIdx pos, Value * * a }); } else - url = state.coerceToString(pos, *args[0], context, false, false).toOwned(); + url = state.coerceToString(pos, *args[0], context, + "while evaluating the first argument passed to builtins.fetchMercurial", + false, false).toOwned(); // FIXME: git externals probably can be used to bypass the URI // whitelist. Ah well. diff --git a/src/libexpr/primops/fetchTree.cc b/src/libexpr/primops/fetchTree.cc index 84e7f5c02..0f54fedde 100644 --- a/src/libexpr/primops/fetchTree.cc +++ b/src/libexpr/primops/fetchTree.cc @@ -4,6 +4,7 @@ #include "fetchers.hh" #include "filetransfer.hh" #include "registry.hh" +#include "url.hh" #include <ctime> #include <iomanip> @@ -68,7 +69,16 @@ void emitTreeAttrs( std::string fixURI(std::string uri, EvalState & state, const std::string & defaultScheme = "file") { state.checkURI(uri); - return uri.find("://") != std::string::npos ? uri : defaultScheme + "://" + uri; + if (uri.find("://") == std::string::npos) { + const auto p = ParsedURL { + .scheme = defaultScheme, + .authority = "", + .path = uri + }; + return p.to_string(); + } else { + return uri; + } } std::string fixURIForGit(std::string uri, EvalState & state) @@ -102,7 +112,7 @@ static void fetchTree( state.forceValue(*args[0], pos); if (args[0]->type() == nAttrs) { - state.forceAttrs(*args[0], pos); + state.forceAttrs(*args[0], pos, "while evaluating the argument passed to builtins.fetchTree"); fetchers::Attrs attrs; @@ -112,7 +122,7 @@ static void fetchTree( .msg = hintfmt("unexpected attribute 'type'"), .errPos = state.positions[pos] })); - type = state.forceStringNoCtx(*aType->value, aType->pos); + type = state.forceStringNoCtx(*aType->value, aType->pos, "while evaluating the `type` attribute passed to builtins.fetchTree"); } else if (!type) state.debugThrowLastTrace(EvalError({ .msg = hintfmt("attribute 'type' is missing in call to 'fetchTree'"), @@ -125,7 +135,7 @@ static void fetchTree( if (attr.name == state.sType) continue; state.forceValue(*attr.value, attr.pos); if (attr.value->type() == nPath || attr.value->type() == nString) { - auto s = state.coerceToString(attr.pos, *attr.value, context, false, false).toOwned(); + auto s = state.coerceToString(attr.pos, *attr.value, context, "", false, false).toOwned(); attrs.emplace(state.symbols[attr.name], state.symbols[attr.name] == "url" ? type == "git" @@ -151,7 +161,9 @@ static void fetchTree( input = fetchers::Input::fromAttrs(std::move(attrs)); } else { - auto url = state.coerceToString(pos, *args[0], context, false, false).toOwned(); + auto url = state.coerceToString(pos, *args[0], context, + "while evaluating the first argument passed to the fetcher", + false, false).toOwned(); if (type == "git") { fetchers::Attrs attrs; @@ -195,16 +207,14 @@ static void fetch(EvalState & state, const PosIdx pos, Value * * args, Value & v if (args[0]->type() == nAttrs) { - state.forceAttrs(*args[0], pos); - for (auto & attr : *args[0]->attrs) { std::string_view n(state.symbols[attr.name]); if (n == "url") - url = state.forceStringNoCtx(*attr.value, attr.pos); + url = state.forceStringNoCtx(*attr.value, attr.pos, "while evaluating the url we should fetch"); else if (n == "sha256") - expectedHash = newHashAllowEmpty(state.forceStringNoCtx(*attr.value, attr.pos), htSHA256); + expectedHash = newHashAllowEmpty(state.forceStringNoCtx(*attr.value, attr.pos, "while evaluating the sha256 of the content we should fetch"), htSHA256); else if (n == "name") - name = state.forceStringNoCtx(*attr.value, attr.pos); + name = state.forceStringNoCtx(*attr.value, attr.pos, "while evaluating the name of the content we should fetch"); else state.debugThrowLastTrace(EvalError({ .msg = hintfmt("unsupported argument '%s' to '%s'", n, who), @@ -218,9 +228,10 @@ static void fetch(EvalState & state, const PosIdx pos, Value * * args, Value & v .errPos = state.positions[pos] })); } else - url = state.forceStringNoCtx(*args[0], pos); + url = state.forceStringNoCtx(*args[0], pos, "while evaluating the url we should fetch"); - url = resolveUri(*url); + if (who == "fetchTarball") + url = evalSettings.resolvePseudoUrl(*url); state.checkURI(*url); @@ -342,36 +353,44 @@ static RegisterPrimOp primop_fetchGit({ of the repo at that URL is fetched. Otherwise, it can be an attribute with the following attributes (all except `url` optional): - - url\ - The URL of the repo. + - `url` + + The URL of the repo. + + - `name` (default: *basename of the URL*) + + The name of the directory the repo should be exported to in the store. + + - `rev` (default: *the tip of `ref`*) + + The [Git revision] to fetch. + This is typically a commit hash. + + [Git revision]: https://git-scm.com/docs/git-rev-parse#_specifying_revisions - - name\ - The name of the directory the repo should be exported to in the - store. Defaults to the basename of the URL. + - `ref` (default: `HEAD`) - - rev\ - The git revision to fetch. Defaults to the tip of `ref`. + The [Git reference] under which to look for the requested revision. + This is often a branch or tag name. - - ref\ - The git ref to look for the requested revision under. This is - often a branch or tag name. Defaults to `HEAD`. + [Git reference]: https://git-scm.com/book/en/v2/Git-Internals-Git-References - By default, the `ref` value is prefixed with `refs/heads/`. As - of Nix 2.3.0 Nix will not prefix `refs/heads/` if `ref` starts - with `refs/`. + By default, the `ref` value is prefixed with `refs/heads/`. + As of 2.3.0, Nix will not prefix `refs/heads/` if `ref` starts with `refs/`. - - submodules\ - A Boolean parameter that specifies whether submodules should be - checked out. Defaults to `false`. + - `submodules` (default: `false`) - - shallow\ - A Boolean parameter that specifies whether fetching a shallow clone - is allowed. Defaults to `false`. + A Boolean parameter that specifies whether submodules should be checked out. - - allRefs\ - Whether to fetch all refs of the repository. With this argument being - true, it's possible to load a `rev` from *any* `ref` (by default only - `rev`s from the specified `ref` are supported). + - `shallow` (default: `false`) + + A Boolean parameter that specifies whether fetching a shallow clone is allowed. + + - `allRefs` + + Whether to fetch all references of the repository. + With this argument being true, it's possible to load a `rev` from *any* `ref` + (by default only `rev`s from the specified `ref` are supported). Here are some examples of how to use `fetchGit`. @@ -455,6 +474,17 @@ static RegisterPrimOp primop_fetchGit({ > **Note** > > This behavior is disabled in *Pure evaluation mode*. + + - To fetch the content of a checked-out work directory: + + ```nix + builtins.fetchGit ./work-dir + ``` + + If the URL points to a local directory, and no `ref` or `rev` is + given, `fetchGit` will use the current content of the checked-out + files, even if they are not committed or added to Git's index. It will + only consider files added to the Git repository, as listed by `git ls-files`. )", .fun = prim_fetchGit, }); diff --git a/src/libexpr/primops/fromTOML.cc b/src/libexpr/primops/fromTOML.cc index 9753e2ac9..8a5231781 100644 --- a/src/libexpr/primops/fromTOML.cc +++ b/src/libexpr/primops/fromTOML.cc @@ -7,7 +7,7 @@ namespace nix { static void prim_fromTOML(EvalState & state, const PosIdx pos, Value * * args, Value & val) { - auto toml = state.forceStringNoCtx(*args[0], pos); + auto toml = state.forceStringNoCtx(*args[0], pos, "while evaluating the argument passed to builtins.fromTOML"); std::istringstream tomlStream(std::string{toml}); diff --git a/src/libexpr/tests/error_traces.cc b/src/libexpr/tests/error_traces.cc new file mode 100644 index 000000000..24e95ac39 --- /dev/null +++ b/src/libexpr/tests/error_traces.cc @@ -0,0 +1,1298 @@ +#include <gmock/gmock.h> +#include <gtest/gtest.h> + +#include "tests/libexpr.hh" + +namespace nix { + + using namespace testing; + + // Testing eval of PrimOp's + class ErrorTraceTest : public LibExprTest { }; + + TEST_F(ErrorTraceTest, TraceBuilder) { + ASSERT_THROW( + state.error("Not much").debugThrow<EvalError>(), + EvalError + ); + + ASSERT_THROW( + state.error("Not much").withTrace(noPos, "No more").debugThrow<EvalError>(), + EvalError + ); + + ASSERT_THROW( + try { + try { + state.error("Not much").withTrace(noPos, "No more").debugThrow<EvalError>(); + } catch (Error & e) { + e.addTrace(state.positions[noPos], "Something", ""); + throw; + } + } catch (BaseError & e) { + ASSERT_EQ(PrintToString(e.info().msg), + PrintToString(hintfmt("Not much"))); + auto trace = e.info().traces.rbegin(); + ASSERT_EQ(e.info().traces.size(), 2); + ASSERT_EQ(PrintToString(trace->hint), + PrintToString(hintfmt("No more"))); + trace++; + ASSERT_EQ(PrintToString(trace->hint), + PrintToString(hintfmt("Something"))); + throw; + } + , EvalError + ); + } + + TEST_F(ErrorTraceTest, NestedThrows) { + try { + state.error("Not much").withTrace(noPos, "No more").debugThrow<EvalError>(); + } catch (BaseError & e) { + try { + state.error("Not much more").debugThrow<EvalError>(); + } catch (Error & e2) { + e.addTrace(state.positions[noPos], "Something", ""); + //e2.addTrace(state.positions[noPos], "Something", ""); + ASSERT_TRUE(e.info().traces.size() == 2); + ASSERT_TRUE(e2.info().traces.size() == 0); + ASSERT_FALSE(&e.info() == &e2.info()); + } + } + } + +#define ASSERT_TRACE1(args, type, message) \ + ASSERT_THROW( \ + std::string expr(args); \ + std::string name = expr.substr(0, expr.find(" ")); \ + try { \ + Value v = eval("builtins." args); \ + state.forceValueDeep(v); \ + } catch (BaseError & e) { \ + ASSERT_EQ(PrintToString(e.info().msg), \ + PrintToString(message)); \ + ASSERT_EQ(e.info().traces.size(), 1) << "while testing " args << std::endl << e.what(); \ + auto trace = e.info().traces.rbegin(); \ + ASSERT_EQ(PrintToString(trace->hint), \ + PrintToString(hintfmt("while calling the '%s' builtin", name))); \ + throw; \ + } \ + , type \ + ) + +#define ASSERT_TRACE2(args, type, message, context) \ + ASSERT_THROW( \ + std::string expr(args); \ + std::string name = expr.substr(0, expr.find(" ")); \ + try { \ + Value v = eval("builtins." args); \ + state.forceValueDeep(v); \ + } catch (BaseError & e) { \ + ASSERT_EQ(PrintToString(e.info().msg), \ + PrintToString(message)); \ + ASSERT_EQ(e.info().traces.size(), 2) << "while testing " args << std::endl << e.what(); \ + auto trace = e.info().traces.rbegin(); \ + ASSERT_EQ(PrintToString(trace->hint), \ + PrintToString(context)); \ + ++trace; \ + ASSERT_EQ(PrintToString(trace->hint), \ + PrintToString(hintfmt("while calling the '%s' builtin", name))); \ + throw; \ + } \ + , type \ + ) + + TEST_F(ErrorTraceTest, genericClosure) { + ASSERT_TRACE2("genericClosure 1", + TypeError, + hintfmt("value is %s while a set was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.genericClosure")); + + ASSERT_TRACE2("genericClosure {}", + TypeError, + hintfmt("attribute '%s' missing", "startSet"), + hintfmt("in the attrset passed as argument to builtins.genericClosure")); + + ASSERT_TRACE2("genericClosure { startSet = 1; }", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the 'startSet' attribute passed as argument to builtins.genericClosure")); + + ASSERT_TRACE2("genericClosure { startSet = [{ key = 1;}]; operator = true; }", + TypeError, + hintfmt("value is %s while a function was expected", "a Boolean"), + hintfmt("while evaluating the 'operator' attribute passed as argument to builtins.genericClosure")); + + ASSERT_TRACE2("genericClosure { startSet = [{ key = 1;}]; operator = item: true; }", + TypeError, + hintfmt("value is %s while a list was expected", "a Boolean"), + hintfmt("while evaluating the return value of the `operator` passed to builtins.genericClosure")); + + ASSERT_TRACE2("genericClosure { startSet = [{ key = 1;}]; operator = item: [ true ]; }", + TypeError, + hintfmt("value is %s while a set was expected", "a Boolean"), + hintfmt("while evaluating one of the elements generated by (or initially passed to) builtins.genericClosure")); + + ASSERT_TRACE2("genericClosure { startSet = [{ key = 1;}]; operator = item: [ {} ]; }", + TypeError, + hintfmt("attribute '%s' missing", "key"), + hintfmt("in one of the attrsets generated by (or initially passed to) builtins.genericClosure")); + + ASSERT_TRACE2("genericClosure { startSet = [{ key = 1;}]; operator = item: [{ key = ''a''; }]; }", + EvalError, + hintfmt("cannot compare %s with %s", "a string", "an integer"), + hintfmt("while comparing the `key` attributes of two genericClosure elements")); + + ASSERT_TRACE2("genericClosure { startSet = [ true ]; operator = item: [{ key = ''a''; }]; }", + TypeError, + hintfmt("value is %s while a set was expected", "a Boolean"), + hintfmt("while evaluating one of the elements generated by (or initially passed to) builtins.genericClosure")); + + } + + + TEST_F(ErrorTraceTest, replaceStrings) { + ASSERT_TRACE2("replaceStrings 0 0 {}", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.replaceStrings")); + + ASSERT_TRACE2("replaceStrings [] 0 {}", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the second argument passed to builtins.replaceStrings")); + + ASSERT_TRACE1("replaceStrings [ 0 ] [] {}", + EvalError, + hintfmt("'from' and 'to' arguments passed to builtins.replaceStrings have different lengths")); + + ASSERT_TRACE2("replaceStrings [ 1 ] [ \"new\" ] {}", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating one of the strings to replace passed to builtins.replaceStrings")); + + ASSERT_TRACE2("replaceStrings [ \"old\" ] [ true ] {}", + TypeError, + hintfmt("value is %s while a string was expected", "a Boolean"), + hintfmt("while evaluating one of the replacement strings passed to builtins.replaceStrings")); + + ASSERT_TRACE2("replaceStrings [ \"old\" ] [ \"new\" ] {}", + TypeError, + hintfmt("value is %s while a string was expected", "a set"), + hintfmt("while evaluating the third argument passed to builtins.replaceStrings")); + + } + + + TEST_F(ErrorTraceTest, scopedImport) { + } + + + TEST_F(ErrorTraceTest, import) { + } + + + TEST_F(ErrorTraceTest, typeOf) { + } + + + TEST_F(ErrorTraceTest, isNull) { + } + + + TEST_F(ErrorTraceTest, isFunction) { + } + + + TEST_F(ErrorTraceTest, isInt) { + } + + + TEST_F(ErrorTraceTest, isFloat) { + } + + + TEST_F(ErrorTraceTest, isString) { + } + + + TEST_F(ErrorTraceTest, isBool) { + } + + + TEST_F(ErrorTraceTest, isPath) { + } + + + TEST_F(ErrorTraceTest, break) { + } + + + TEST_F(ErrorTraceTest, abort) { + } + + + TEST_F(ErrorTraceTest, throw) { + } + + + TEST_F(ErrorTraceTest, addErrorContext) { + } + + + TEST_F(ErrorTraceTest, ceil) { + ASSERT_TRACE2("ceil \"foo\"", + TypeError, + hintfmt("value is %s while a float was expected", "a string"), + hintfmt("while evaluating the first argument passed to builtins.ceil")); + + } + + + TEST_F(ErrorTraceTest, floor) { + ASSERT_TRACE2("floor \"foo\"", + TypeError, + hintfmt("value is %s while a float was expected", "a string"), + hintfmt("while evaluating the first argument passed to builtins.floor")); + + } + + + TEST_F(ErrorTraceTest, tryEval) { + } + + + TEST_F(ErrorTraceTest, getEnv) { + ASSERT_TRACE2("getEnv [ ]", + TypeError, + hintfmt("value is %s while a string was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.getEnv")); + + } + + + TEST_F(ErrorTraceTest, seq) { + } + + + TEST_F(ErrorTraceTest, deepSeq) { + } + + + TEST_F(ErrorTraceTest, trace) { + } + + + TEST_F(ErrorTraceTest, placeholder) { + ASSERT_TRACE2("placeholder []", + TypeError, + hintfmt("value is %s while a string was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.placeholder")); + + } + + + TEST_F(ErrorTraceTest, toPath) { + ASSERT_TRACE2("toPath []", + TypeError, + hintfmt("cannot coerce %s to a string", "a list"), + hintfmt("while evaluating the first argument passed to builtins.toPath")); + + ASSERT_TRACE2("toPath \"foo\"", + EvalError, + hintfmt("string '%s' doesn't represent an absolute path", "foo"), + hintfmt("while evaluating the first argument passed to builtins.toPath")); + + } + + + TEST_F(ErrorTraceTest, storePath) { + ASSERT_TRACE2("storePath true", + TypeError, + hintfmt("cannot coerce %s to a string", "a Boolean"), + hintfmt("while evaluating the first argument passed to builtins.storePath")); + + } + + + TEST_F(ErrorTraceTest, pathExists) { + ASSERT_TRACE2("pathExists []", + TypeError, + hintfmt("cannot coerce %s to a string", "a list"), + hintfmt("while realising the context of a path")); + + ASSERT_TRACE2("pathExists \"zorglub\"", + EvalError, + hintfmt("string '%s' doesn't represent an absolute path", "zorglub"), + hintfmt("while realising the context of a path")); + + } + + + TEST_F(ErrorTraceTest, baseNameOf) { + ASSERT_TRACE2("baseNameOf []", + TypeError, + hintfmt("cannot coerce %s to a string", "a list"), + hintfmt("while evaluating the first argument passed to builtins.baseNameOf")); + + } + + + TEST_F(ErrorTraceTest, dirOf) { + } + + + TEST_F(ErrorTraceTest, readFile) { + } + + + TEST_F(ErrorTraceTest, findFile) { + } + + + TEST_F(ErrorTraceTest, hashFile) { + } + + + TEST_F(ErrorTraceTest, readDir) { + } + + + TEST_F(ErrorTraceTest, toXML) { + } + + + TEST_F(ErrorTraceTest, toJSON) { + } + + + TEST_F(ErrorTraceTest, fromJSON) { + } + + + TEST_F(ErrorTraceTest, toFile) { + } + + + TEST_F(ErrorTraceTest, filterSource) { + ASSERT_TRACE2("filterSource [] []", + TypeError, + hintfmt("cannot coerce %s to a string", "a list"), + hintfmt("while evaluating the second argument (the path to filter) passed to builtins.filterSource")); + + ASSERT_TRACE2("filterSource [] \"foo\"", + EvalError, + hintfmt("string '%s' doesn't represent an absolute path", "foo"), + hintfmt("while evaluating the second argument (the path to filter) passed to builtins.filterSource")); + + ASSERT_TRACE2("filterSource [] ./.", + TypeError, + hintfmt("value is %s while a function was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.filterSource")); + + // Usupported by store "dummy" + + // ASSERT_TRACE2("filterSource (_: 1) ./.", + // TypeError, + // hintfmt("attempt to call something which is not a function but %s", "an integer"), + // hintfmt("while adding path '/home/layus/projects/nix'")); + + // ASSERT_TRACE2("filterSource (_: _: 1) ./.", + // TypeError, + // hintfmt("value is %s while a Boolean was expected", "an integer"), + // hintfmt("while evaluating the return value of the path filter function")); + + } + + + TEST_F(ErrorTraceTest, path) { + } + + + TEST_F(ErrorTraceTest, attrNames) { + ASSERT_TRACE2("attrNames []", + TypeError, + hintfmt("value is %s while a set was expected", "a list"), + hintfmt("while evaluating the argument passed to builtins.attrNames")); + + } + + + TEST_F(ErrorTraceTest, attrValues) { + ASSERT_TRACE2("attrValues []", + TypeError, + hintfmt("value is %s while a set was expected", "a list"), + hintfmt("while evaluating the argument passed to builtins.attrValues")); + + } + + + TEST_F(ErrorTraceTest, getAttr) { + ASSERT_TRACE2("getAttr [] []", + TypeError, + hintfmt("value is %s while a string was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.getAttr")); + + ASSERT_TRACE2("getAttr \"foo\" []", + TypeError, + hintfmt("value is %s while a set was expected", "a list"), + hintfmt("while evaluating the second argument passed to builtins.getAttr")); + + ASSERT_TRACE2("getAttr \"foo\" {}", + TypeError, + hintfmt("attribute '%s' missing", "foo"), + hintfmt("in the attribute set under consideration")); + + } + + + TEST_F(ErrorTraceTest, unsafeGetAttrPos) { + } + + + TEST_F(ErrorTraceTest, hasAttr) { + ASSERT_TRACE2("hasAttr [] []", + TypeError, + hintfmt("value is %s while a string was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.hasAttr")); + + ASSERT_TRACE2("hasAttr \"foo\" []", + TypeError, + hintfmt("value is %s while a set was expected", "a list"), + hintfmt("while evaluating the second argument passed to builtins.hasAttr")); + + } + + + TEST_F(ErrorTraceTest, isAttrs) { + } + + + TEST_F(ErrorTraceTest, removeAttrs) { + ASSERT_TRACE2("removeAttrs \"\" \"\"", + TypeError, + hintfmt("value is %s while a set was expected", "a string"), + hintfmt("while evaluating the first argument passed to builtins.removeAttrs")); + + ASSERT_TRACE2("removeAttrs \"\" [ 1 ]", + TypeError, + hintfmt("value is %s while a set was expected", "a string"), + hintfmt("while evaluating the first argument passed to builtins.removeAttrs")); + + ASSERT_TRACE2("removeAttrs \"\" [ \"1\" ]", + TypeError, + hintfmt("value is %s while a set was expected", "a string"), + hintfmt("while evaluating the first argument passed to builtins.removeAttrs")); + + } + + + TEST_F(ErrorTraceTest, listToAttrs) { + ASSERT_TRACE2("listToAttrs 1", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the argument passed to builtins.listToAttrs")); + + ASSERT_TRACE2("listToAttrs [ 1 ]", + TypeError, + hintfmt("value is %s while a set was expected", "an integer"), + hintfmt("while evaluating an element of the list passed to builtins.listToAttrs")); + + ASSERT_TRACE2("listToAttrs [ {} ]", + TypeError, + hintfmt("attribute '%s' missing", "name"), + hintfmt("in a {name=...; value=...;} pair")); + + ASSERT_TRACE2("listToAttrs [ { name = 1; } ]", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the `name` attribute of an element of the list passed to builtins.listToAttrs")); + + ASSERT_TRACE2("listToAttrs [ { name = \"foo\"; } ]", + TypeError, + hintfmt("attribute '%s' missing", "value"), + hintfmt("in a {name=...; value=...;} pair")); + + } + + + TEST_F(ErrorTraceTest, intersectAttrs) { + ASSERT_TRACE2("intersectAttrs [] []", + TypeError, + hintfmt("value is %s while a set was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.intersectAttrs")); + + ASSERT_TRACE2("intersectAttrs {} []", + TypeError, + hintfmt("value is %s while a set was expected", "a list"), + hintfmt("while evaluating the second argument passed to builtins.intersectAttrs")); + + } + + + TEST_F(ErrorTraceTest, catAttrs) { + ASSERT_TRACE2("catAttrs [] {}", + TypeError, + hintfmt("value is %s while a string was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.catAttrs")); + + ASSERT_TRACE2("catAttrs \"foo\" {}", + TypeError, + hintfmt("value is %s while a list was expected", "a set"), + hintfmt("while evaluating the second argument passed to builtins.catAttrs")); + + ASSERT_TRACE2("catAttrs \"foo\" [ 1 ]", + TypeError, + hintfmt("value is %s while a set was expected", "an integer"), + hintfmt("while evaluating an element in the list passed as second argument to builtins.catAttrs")); + + ASSERT_TRACE2("catAttrs \"foo\" [ { foo = 1; } 1 { bar = 5;} ]", + TypeError, + hintfmt("value is %s while a set was expected", "an integer"), + hintfmt("while evaluating an element in the list passed as second argument to builtins.catAttrs")); + + } + + + TEST_F(ErrorTraceTest, functionArgs) { + ASSERT_TRACE1("functionArgs {}", + TypeError, + hintfmt("'functionArgs' requires a function")); + + } + + + TEST_F(ErrorTraceTest, mapAttrs) { + ASSERT_TRACE2("mapAttrs [] []", + TypeError, + hintfmt("value is %s while a set was expected", "a list"), + hintfmt("while evaluating the second argument passed to builtins.mapAttrs")); + + // XXX: defered + // ASSERT_TRACE2("mapAttrs \"\" { foo.bar = 1; }", + // TypeError, + // hintfmt("attempt to call something which is not a function but %s", "a string"), + // hintfmt("while evaluating the attribute 'foo'")); + + // ASSERT_TRACE2("mapAttrs (x: x + \"1\") { foo.bar = 1; }", + // TypeError, + // hintfmt("attempt to call something which is not a function but %s", "a string"), + // hintfmt("while evaluating the attribute 'foo'")); + + // ASSERT_TRACE2("mapAttrs (x: y: x + 1) { foo.bar = 1; }", + // TypeError, + // hintfmt("cannot coerce %s to a string", "an integer"), + // hintfmt("while evaluating a path segment")); + + } + + + TEST_F(ErrorTraceTest, zipAttrsWith) { + ASSERT_TRACE2("zipAttrsWith [] [ 1 ]", + TypeError, + hintfmt("value is %s while a function was expected", "a list"), + hintfmt("while evaluating the first argument passed to builtins.zipAttrsWith")); + + ASSERT_TRACE2("zipAttrsWith (_: 1) [ 1 ]", + TypeError, + hintfmt("value is %s while a set was expected", "an integer"), + hintfmt("while evaluating a value of the list passed as second argument to builtins.zipAttrsWith")); + + // XXX: How to properly tell that the fucntion takes two arguments ? + // The same question also applies to sort, and maybe others. + // Due to lazyness, we only create a thunk, and it fails later on. + // ASSERT_TRACE2("zipAttrsWith (_: 1) [ { foo = 1; } ]", + // TypeError, + // hintfmt("attempt to call something which is not a function but %s", "an integer"), + // hintfmt("while evaluating the attribute 'foo'")); + + // XXX: Also deferred deeply + // ASSERT_TRACE2("zipAttrsWith (a: b: a + b) [ { foo = 1; } { foo = 2; } ]", + // TypeError, + // hintfmt("cannot coerce %s to a string", "a list"), + // hintfmt("while evaluating a path segment")); + + } + + + TEST_F(ErrorTraceTest, isList) { + } + + + TEST_F(ErrorTraceTest, elemAt) { + ASSERT_TRACE2("elemAt \"foo\" (-1)", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the first argument passed to builtins.elemAt")); + + ASSERT_TRACE1("elemAt [] (-1)", + Error, + hintfmt("list index %d is out of bounds", -1)); + + ASSERT_TRACE1("elemAt [\"foo\"] 3", + Error, + hintfmt("list index %d is out of bounds", 3)); + + } + + + TEST_F(ErrorTraceTest, head) { + ASSERT_TRACE2("head 1", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.elemAt")); + + ASSERT_TRACE1("head []", + Error, + hintfmt("list index %d is out of bounds", 0)); + + } + + + TEST_F(ErrorTraceTest, tail) { + ASSERT_TRACE2("tail 1", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.tail")); + + ASSERT_TRACE1("tail []", + Error, + hintfmt("'tail' called on an empty list")); + + } + + + TEST_F(ErrorTraceTest, map) { + ASSERT_TRACE2("map 1 \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.map")); + + ASSERT_TRACE2("map 1 [ 1 ]", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.map")); + + } + + + TEST_F(ErrorTraceTest, filter) { + ASSERT_TRACE2("filter 1 \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.filter")); + + ASSERT_TRACE2("filter 1 [ \"foo\" ]", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.filter")); + + ASSERT_TRACE2("filter (_: 5) [ \"foo\" ]", + TypeError, + hintfmt("value is %s while a Boolean was expected", "an integer"), + hintfmt("while evaluating the return value of the filtering function passed to builtins.filter")); + + } + + + TEST_F(ErrorTraceTest, elem) { + ASSERT_TRACE2("elem 1 \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.elem")); + + } + + + TEST_F(ErrorTraceTest, concatLists) { + ASSERT_TRACE2("concatLists 1", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.concatLists")); + + ASSERT_TRACE2("concatLists [ 1 ]", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating a value of the list passed to builtins.concatLists")); + + ASSERT_TRACE2("concatLists [ [1] \"foo\" ]", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating a value of the list passed to builtins.concatLists")); + + } + + + TEST_F(ErrorTraceTest, length) { + ASSERT_TRACE2("length 1", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.length")); + + ASSERT_TRACE2("length \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the first argument passed to builtins.length")); + + } + + + TEST_F(ErrorTraceTest, foldlPrime) { + ASSERT_TRACE2("foldl' 1 \"foo\" true", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.foldlStrict")); + + ASSERT_TRACE2("foldl' (_: 1) \"foo\" true", + TypeError, + hintfmt("value is %s while a list was expected", "a Boolean"), + hintfmt("while evaluating the third argument passed to builtins.foldlStrict")); + + ASSERT_TRACE1("foldl' (_: 1) \"foo\" [ true ]", + TypeError, + hintfmt("attempt to call something which is not a function but %s", "an integer")); + + ASSERT_TRACE2("foldl' (a: b: a && b) \"foo\" [ true ]", + TypeError, + hintfmt("value is %s while a Boolean was expected", "a string"), + hintfmt("in the left operand of the AND (&&) operator")); + + } + + + TEST_F(ErrorTraceTest, any) { + ASSERT_TRACE2("any 1 \"foo\"", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.any")); + + ASSERT_TRACE2("any (_: 1) \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.any")); + + ASSERT_TRACE2("any (_: 1) [ \"foo\" ]", + TypeError, + hintfmt("value is %s while a Boolean was expected", "an integer"), + hintfmt("while evaluating the return value of the function passed to builtins.any")); + + } + + + TEST_F(ErrorTraceTest, all) { + ASSERT_TRACE2("all 1 \"foo\"", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.all")); + + ASSERT_TRACE2("all (_: 1) \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.all")); + + ASSERT_TRACE2("all (_: 1) [ \"foo\" ]", + TypeError, + hintfmt("value is %s while a Boolean was expected", "an integer"), + hintfmt("while evaluating the return value of the function passed to builtins.all")); + + } + + + TEST_F(ErrorTraceTest, genList) { + ASSERT_TRACE2("genList 1 \"foo\"", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.genList")); + + ASSERT_TRACE2("genList 1 2", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.genList", "an integer")); + + // XXX: defered + // ASSERT_TRACE2("genList (x: x + \"foo\") 2 #TODO", + // TypeError, + // hintfmt("cannot add %s to an integer", "a string"), + // hintfmt("while evaluating anonymous lambda")); + + ASSERT_TRACE1("genList false (-3)", + EvalError, + hintfmt("cannot create list of size %d", -3)); + + } + + + TEST_F(ErrorTraceTest, sort) { + ASSERT_TRACE2("sort 1 \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.sort")); + + ASSERT_TRACE2("sort 1 [ \"foo\" ]", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.sort")); + + ASSERT_TRACE1("sort (_: 1) [ \"foo\" \"bar\" ]", + TypeError, + hintfmt("attempt to call something which is not a function but %s", "an integer")); + + ASSERT_TRACE2("sort (_: _: 1) [ \"foo\" \"bar\" ]", + TypeError, + hintfmt("value is %s while a Boolean was expected", "an integer"), + hintfmt("while evaluating the return value of the sorting function passed to builtins.sort")); + + // XXX: Trace too deep, need better asserts + // ASSERT_TRACE1("sort (a: b: a <= b) [ \"foo\" {} ] # TODO", + // TypeError, + // hintfmt("cannot compare %s with %s", "a string", "a set")); + + // ASSERT_TRACE1("sort (a: b: a <= b) [ {} {} ] # TODO", + // TypeError, + // hintfmt("cannot compare %s with %s; values of that type are incomparable", "a set", "a set")); + + } + + + TEST_F(ErrorTraceTest, partition) { + ASSERT_TRACE2("partition 1 \"foo\"", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.partition")); + + ASSERT_TRACE2("partition (_: 1) \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.partition")); + + ASSERT_TRACE2("partition (_: 1) [ \"foo\" ]", + TypeError, + hintfmt("value is %s while a Boolean was expected", "an integer"), + hintfmt("while evaluating the return value of the partition function passed to builtins.partition")); + + } + + + TEST_F(ErrorTraceTest, groupBy) { + ASSERT_TRACE2("groupBy 1 \"foo\"", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.groupBy")); + + ASSERT_TRACE2("groupBy (_: 1) \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.groupBy")); + + ASSERT_TRACE2("groupBy (x: x) [ \"foo\" \"bar\" 1 ]", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the return value of the grouping function passed to builtins.groupBy")); + + } + + + TEST_F(ErrorTraceTest, concatMap) { + ASSERT_TRACE2("concatMap 1 \"foo\"", + TypeError, + hintfmt("value is %s while a function was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.concatMap")); + + ASSERT_TRACE2("concatMap (x: 1) \"foo\"", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the second argument passed to builtins.concatMap")); + + ASSERT_TRACE2("concatMap (x: 1) [ \"foo\" ] # TODO", + TypeError, + hintfmt("value is %s while a list was expected", "an integer"), + hintfmt("while evaluating the return value of the function passed to buitlins.concatMap")); + + ASSERT_TRACE2("concatMap (x: \"foo\") [ 1 2 ] # TODO", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the return value of the function passed to buitlins.concatMap")); + + } + + + TEST_F(ErrorTraceTest, add) { + ASSERT_TRACE2("add \"foo\" 1", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the first argument of the addition")); + + ASSERT_TRACE2("add 1 \"foo\"", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the second argument of the addition")); + + } + + + TEST_F(ErrorTraceTest, sub) { + ASSERT_TRACE2("sub \"foo\" 1", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the first argument of the subtraction")); + + ASSERT_TRACE2("sub 1 \"foo\"", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the second argument of the subtraction")); + + } + + + TEST_F(ErrorTraceTest, mul) { + ASSERT_TRACE2("mul \"foo\" 1", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the first argument of the multiplication")); + + ASSERT_TRACE2("mul 1 \"foo\"", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the second argument of the multiplication")); + + } + + + TEST_F(ErrorTraceTest, div) { + ASSERT_TRACE2("div \"foo\" 1 # TODO: an integer was expected -> a number", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the first operand of the division")); + + ASSERT_TRACE2("div 1 \"foo\"", + TypeError, + hintfmt("value is %s while a float was expected", "a string"), + hintfmt("while evaluating the second operand of the division")); + + ASSERT_TRACE1("div \"foo\" 0", + EvalError, + hintfmt("division by zero")); + + } + + + TEST_F(ErrorTraceTest, bitAnd) { + ASSERT_TRACE2("bitAnd 1.1 2", + TypeError, + hintfmt("value is %s while an integer was expected", "a float"), + hintfmt("while evaluating the first argument passed to builtins.bitAnd")); + + ASSERT_TRACE2("bitAnd 1 2.2", + TypeError, + hintfmt("value is %s while an integer was expected", "a float"), + hintfmt("while evaluating the second argument passed to builtins.bitAnd")); + + } + + + TEST_F(ErrorTraceTest, bitOr) { + ASSERT_TRACE2("bitOr 1.1 2", + TypeError, + hintfmt("value is %s while an integer was expected", "a float"), + hintfmt("while evaluating the first argument passed to builtins.bitOr")); + + ASSERT_TRACE2("bitOr 1 2.2", + TypeError, + hintfmt("value is %s while an integer was expected", "a float"), + hintfmt("while evaluating the second argument passed to builtins.bitOr")); + + } + + + TEST_F(ErrorTraceTest, bitXor) { + ASSERT_TRACE2("bitXor 1.1 2", + TypeError, + hintfmt("value is %s while an integer was expected", "a float"), + hintfmt("while evaluating the first argument passed to builtins.bitXor")); + + ASSERT_TRACE2("bitXor 1 2.2", + TypeError, + hintfmt("value is %s while an integer was expected", "a float"), + hintfmt("while evaluating the second argument passed to builtins.bitXor")); + + } + + + TEST_F(ErrorTraceTest, lessThan) { + ASSERT_TRACE1("lessThan 1 \"foo\"", + EvalError, + hintfmt("cannot compare %s with %s", "an integer", "a string")); + + ASSERT_TRACE1("lessThan {} {}", + EvalError, + hintfmt("cannot compare %s with %s; values of that type are incomparable", "a set", "a set")); + + ASSERT_TRACE2("lessThan [ 1 2 ] [ \"foo\" ]", + EvalError, + hintfmt("cannot compare %s with %s", "an integer", "a string"), + hintfmt("while comparing two list elements")); + + } + + + TEST_F(ErrorTraceTest, toString) { + ASSERT_TRACE2("toString { a = 1; }", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating the first argument passed to builtins.toString")); + + } + + + TEST_F(ErrorTraceTest, substring) { + ASSERT_TRACE2("substring {} \"foo\" true", + TypeError, + hintfmt("value is %s while an integer was expected", "a set"), + hintfmt("while evaluating the first argument (the start offset) passed to builtins.substring")); + + ASSERT_TRACE2("substring 3 \"foo\" true", + TypeError, + hintfmt("value is %s while an integer was expected", "a string"), + hintfmt("while evaluating the second argument (the substring length) passed to builtins.substring")); + + ASSERT_TRACE2("substring 0 3 {}", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating the third argument (the string) passed to builtins.substring")); + + ASSERT_TRACE1("substring (-3) 3 \"sometext\"", + EvalError, + hintfmt("negative start position in 'substring'")); + + } + + + TEST_F(ErrorTraceTest, stringLength) { + ASSERT_TRACE2("stringLength {} # TODO: context is missing ???", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating the argument passed to builtins.stringLength")); + + } + + + TEST_F(ErrorTraceTest, hashString) { + ASSERT_TRACE2("hashString 1 {}", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.hashString")); + + ASSERT_TRACE1("hashString \"foo\" \"content\"", + UsageError, + hintfmt("unknown hash algorithm '%s'", "foo")); + + ASSERT_TRACE2("hashString \"sha256\" {}", + TypeError, + hintfmt("value is %s while a string was expected", "a set"), + hintfmt("while evaluating the second argument passed to builtins.hashString")); + + } + + + TEST_F(ErrorTraceTest, match) { + ASSERT_TRACE2("match 1 {}", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.match")); + + ASSERT_TRACE2("match \"foo\" {}", + TypeError, + hintfmt("value is %s while a string was expected", "a set"), + hintfmt("while evaluating the second argument passed to builtins.match")); + + ASSERT_TRACE1("match \"(.*\" \"\"", + EvalError, + hintfmt("invalid regular expression '%s'", "(.*")); + + } + + + TEST_F(ErrorTraceTest, split) { + ASSERT_TRACE2("split 1 {}", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.split")); + + ASSERT_TRACE2("split \"foo\" {}", + TypeError, + hintfmt("value is %s while a string was expected", "a set"), + hintfmt("while evaluating the second argument passed to builtins.split")); + + ASSERT_TRACE1("split \"f(o*o\" \"1foo2\"", + EvalError, + hintfmt("invalid regular expression '%s'", "f(o*o")); + + } + + + TEST_F(ErrorTraceTest, concatStringsSep) { + ASSERT_TRACE2("concatStringsSep 1 {}", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the first argument (the separator string) passed to builtins.concatStringsSep")); + + ASSERT_TRACE2("concatStringsSep \"foo\" {}", + TypeError, + hintfmt("value is %s while a list was expected", "a set"), + hintfmt("while evaluating the second argument (the list of strings to concat) passed to builtins.concatStringsSep")); + + ASSERT_TRACE2("concatStringsSep \"foo\" [ 1 2 {} ] # TODO: coerce to string is buggy", + TypeError, + hintfmt("cannot coerce %s to a string", "an integer"), + hintfmt("while evaluating one element of the list of strings to concat passed to builtins.concatStringsSep")); + + } + + + TEST_F(ErrorTraceTest, parseDrvName) { + ASSERT_TRACE2("parseDrvName 1", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.parseDrvName")); + + } + + + TEST_F(ErrorTraceTest, compareVersions) { + ASSERT_TRACE2("compareVersions 1 {}", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.compareVersions")); + + ASSERT_TRACE2("compareVersions \"abd\" {}", + TypeError, + hintfmt("value is %s while a string was expected", "a set"), + hintfmt("while evaluating the second argument passed to builtins.compareVersions")); + + } + + + TEST_F(ErrorTraceTest, splitVersion) { + ASSERT_TRACE2("splitVersion 1", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the first argument passed to builtins.splitVersion")); + + } + + + TEST_F(ErrorTraceTest, traceVerbose) { + } + + + /* // Needs different ASSERTs + TEST_F(ErrorTraceTest, derivationStrict) { + ASSERT_TRACE2("derivationStrict \"\"", + TypeError, + hintfmt("value is %s while a set was expected", "a string"), + hintfmt("while evaluating the argument passed to builtins.derivationStrict")); + + ASSERT_TRACE2("derivationStrict {}", + TypeError, + hintfmt("attribute '%s' missing", "name"), + hintfmt("in the attrset passed as argument to builtins.derivationStrict")); + + ASSERT_TRACE2("derivationStrict { name = 1; }", + TypeError, + hintfmt("value is %s while a string was expected", "an integer"), + hintfmt("while evaluating the `name` attribute passed to builtins.derivationStrict")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; }", + TypeError, + hintfmt("required attribute 'builder' missing"), + hintfmt("while evaluating derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; __structuredAttrs = 15; }", + TypeError, + hintfmt("value is %s while a Boolean was expected", "an integer"), + hintfmt("while evaluating the `__structuredAttrs` attribute passed to builtins.derivationStrict")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; __ignoreNulls = 15; }", + TypeError, + hintfmt("value is %s while a Boolean was expected", "an integer"), + hintfmt("while evaluating the `__ignoreNulls` attribute passed to builtins.derivationStrict")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; outputHashMode = 15; }", + TypeError, + hintfmt("invalid value '15' for 'outputHashMode' attribute"), + hintfmt("while evaluating the attribute 'outputHashMode' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; outputHashMode = \"custom\"; }", + TypeError, + hintfmt("invalid value 'custom' for 'outputHashMode' attribute"), + hintfmt("while evaluating the attribute 'outputHashMode' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = {}; }", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating the attribute 'system' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = {}; }", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating the attribute 'outputs' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"drv\"; }", + TypeError, + hintfmt("invalid derivation output name 'drv'"), + hintfmt("while evaluating the attribute 'outputs' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = []; }", + TypeError, + hintfmt("derivation cannot have an empty set of outputs"), + hintfmt("while evaluating the attribute 'outputs' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = [ \"drv\" ]; }", + TypeError, + hintfmt("invalid derivation output name 'drv'"), + hintfmt("while evaluating the attribute 'outputs' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = [ \"out\" \"out\" ]; }", + TypeError, + hintfmt("duplicate derivation output 'out'"), + hintfmt("while evaluating the attribute 'outputs' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"out\"; __contentAddressed = \"true\"; }", + TypeError, + hintfmt("value is %s while a Boolean was expected", "a string"), + hintfmt("while evaluating the attribute '__contentAddressed' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"out\"; __impure = \"true\"; }", + TypeError, + hintfmt("value is %s while a Boolean was expected", "a string"), + hintfmt("while evaluating the attribute '__impure' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"out\"; __impure = \"true\"; }", + TypeError, + hintfmt("value is %s while a Boolean was expected", "a string"), + hintfmt("while evaluating the attribute '__impure' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"out\"; args = \"foo\"; }", + TypeError, + hintfmt("value is %s while a list was expected", "a string"), + hintfmt("while evaluating the attribute 'args' of derivation 'foo'")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"out\"; args = [ {} ]; }", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating an element of the argument list")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"out\"; args = [ \"a\" {} ]; }", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating an element of the argument list")); + + ASSERT_TRACE2("derivationStrict { name = \"foo\"; builder = 1; system = 1; outputs = \"out\"; FOO = {}; }", + TypeError, + hintfmt("cannot coerce %s to a string", "a set"), + hintfmt("while evaluating the attribute 'FOO' of derivation 'foo'")); + + } + */ + +} /* namespace nix */ diff --git a/src/libexpr/tests/json.cc b/src/libexpr/tests/json.cc index f1ea1b197..411bc0ac3 100644 --- a/src/libexpr/tests/json.cc +++ b/src/libexpr/tests/json.cc @@ -1,4 +1,4 @@ -#include "libexprtests.hh" +#include "tests/libexpr.hh" #include "value-to-json.hh" namespace nix { diff --git a/src/libexpr/tests/libexprtests.hh b/src/libexpr/tests/libexpr.hh index 4f6915882..8534d9567 100644 --- a/src/libexpr/tests/libexprtests.hh +++ b/src/libexpr/tests/libexpr.hh @@ -7,17 +7,19 @@ #include "eval-inline.hh" #include "store-api.hh" +#include "tests/libstore.hh" namespace nix { - class LibExprTest : public ::testing::Test { + class LibExprTest : public LibStoreTest { public: static void SetUpTestSuite() { + LibStoreTest::SetUpTestSuite(); initGC(); } protected: LibExprTest() - : store(openStore("dummy://")) + : LibStoreTest() , state({}, store) { } @@ -35,7 +37,6 @@ namespace nix { return state.symbols.create(value); } - ref<Store> store; EvalState state; }; @@ -123,7 +124,7 @@ namespace nix { MATCHER_P(IsAttrsOfSize, n, fmt("Is a set of size [%1%]", n)) { if (arg.type() != nAttrs) { - *result_listener << "Expexted set got " << arg.type(); + *result_listener << "Expected set got " << arg.type(); return false; } else if (arg.attrs->size() != (size_t)n) { *result_listener << "Expected a set with " << n << " attributes but got " << arg.attrs->size(); diff --git a/src/libexpr/tests/local.mk b/src/libexpr/tests/local.mk index b95980cab..3e5504f71 100644 --- a/src/libexpr/tests/local.mk +++ b/src/libexpr/tests/local.mk @@ -2,14 +2,18 @@ check: libexpr-tests_RUN programs += libexpr-tests +libexpr-tests_NAME := libnixexpr-tests + libexpr-tests_DIR := $(d) libexpr-tests_INSTALL_DIR := -libexpr-tests_SOURCES := $(wildcard $(d)/*.cc) +libexpr-tests_SOURCES := \ + $(wildcard $(d)/*.cc) \ + $(wildcard $(d)/value/*.cc) libexpr-tests_CXXFLAGS += -I src/libexpr -I src/libutil -I src/libstore -I src/libexpr/tests -libexpr-tests_LIBS = libexpr libutil libstore libfetchers +libexpr-tests_LIBS = libstore-tests libutils-tests libexpr libutil libstore libfetchers libexpr-tests_LDFLAGS := $(GTEST_LIBS) -lgmock diff --git a/src/libexpr/tests/primops.cc b/src/libexpr/tests/primops.cc index 16cf66d2c..ce3b5d11f 100644 --- a/src/libexpr/tests/primops.cc +++ b/src/libexpr/tests/primops.cc @@ -1,7 +1,7 @@ #include <gmock/gmock.h> #include <gtest/gtest.h> -#include "libexprtests.hh" +#include "tests/libexpr.hh" namespace nix { class CaptureLogger : public Logger @@ -15,8 +15,8 @@ namespace nix { return oss.str(); } - void log(Verbosity lvl, const FormatOrString & fs) override { - oss << fs.s << std::endl; + void log(Verbosity lvl, std::string_view s) override { + oss << s << std::endl; } void logEI(const ErrorInfo & ei) override { @@ -151,20 +151,7 @@ namespace nix { // The `y` attribute is at position const char* expr = "builtins.unsafeGetAttrPos \"y\" { y = \"x\"; }"; auto v = eval(expr); - ASSERT_THAT(v, IsAttrsOfSize(3)); - - auto file = v.attrs->find(createSymbol("file")); - ASSERT_NE(file, nullptr); - // FIXME: The file when running these tests is the input string?!? - ASSERT_THAT(*file->value, IsStringEq(expr)); - - auto line = v.attrs->find(createSymbol("line")); - ASSERT_NE(line, nullptr); - ASSERT_THAT(*line->value, IsIntEq(1)); - - auto column = v.attrs->find(createSymbol("column")); - ASSERT_NE(column, nullptr); - ASSERT_THAT(*column->value, IsIntEq(33)); + ASSERT_THAT(v, IsNull()); } TEST_F(PrimOpTest, hasAttr) { @@ -617,7 +604,7 @@ namespace nix { TEST_F(PrimOpTest, storeDir) { auto v = eval("builtins.storeDir"); - ASSERT_THAT(v, IsStringEq("/nix/store")); + ASSERT_THAT(v, IsStringEq(settings.nixStore)); } TEST_F(PrimOpTest, nixVersion) { @@ -836,4 +823,10 @@ namespace nix { for (const auto [n, elem] : enumerate(v.listItems())) ASSERT_THAT(*elem, IsStringEq(expected[n])); } + + TEST_F(PrimOpTest, genericClosure_not_strict) { + // Operator should not be used when startSet is empty + auto v = eval("builtins.genericClosure { startSet = []; }"); + ASSERT_THAT(v, IsListOfSize(0)); + } } /* namespace nix */ diff --git a/src/libexpr/tests/trivial.cc b/src/libexpr/tests/trivial.cc index 8ce276e52..171727ac7 100644 --- a/src/libexpr/tests/trivial.cc +++ b/src/libexpr/tests/trivial.cc @@ -1,4 +1,4 @@ -#include "libexprtests.hh" +#include "tests/libexpr.hh" namespace nix { // Testing of trivial expressions diff --git a/src/libexpr/tests/value/context.cc b/src/libexpr/tests/value/context.cc new file mode 100644 index 000000000..083359b7a --- /dev/null +++ b/src/libexpr/tests/value/context.cc @@ -0,0 +1,127 @@ +#include <nlohmann/json.hpp> +#include <gtest/gtest.h> +#include <rapidcheck/gtest.h> + +#include "tests/path.hh" +#include "tests/libexpr.hh" +#include "tests/value/context.hh" + +namespace nix { + +// Testing of trivial expressions +struct NixStringContextElemTest : public LibExprTest { + const Store & store() const { + return *LibExprTest::store; + } +}; + +TEST_F(NixStringContextElemTest, empty_invalid) { + EXPECT_THROW( + NixStringContextElem::parse(store(), ""), + BadNixStringContextElem); +} + +TEST_F(NixStringContextElemTest, single_bang_invalid) { + EXPECT_THROW( + NixStringContextElem::parse(store(), "!"), + BadNixStringContextElem); +} + +TEST_F(NixStringContextElemTest, double_bang_invalid) { + EXPECT_THROW( + NixStringContextElem::parse(store(), "!!/"), + BadStorePath); +} + +TEST_F(NixStringContextElemTest, eq_slash_invalid) { + EXPECT_THROW( + NixStringContextElem::parse(store(), "=/"), + BadStorePath); +} + +TEST_F(NixStringContextElemTest, slash_invalid) { + EXPECT_THROW( + NixStringContextElem::parse(store(), "/"), + BadStorePath); +} + +TEST_F(NixStringContextElemTest, opaque) { + std::string_view opaque = "/nix/store/g1w7hy3qg1w7hy3qg1w7hy3qg1w7hy3q-x"; + auto elem = NixStringContextElem::parse(store(), opaque); + auto * p = std::get_if<NixStringContextElem::Opaque>(&elem); + ASSERT_TRUE(p); + ASSERT_EQ(p->path, store().parseStorePath(opaque)); + ASSERT_EQ(elem.to_string(store()), opaque); +} + +TEST_F(NixStringContextElemTest, drvDeep) { + std::string_view drvDeep = "=/nix/store/g1w7hy3qg1w7hy3qg1w7hy3qg1w7hy3q-x.drv"; + auto elem = NixStringContextElem::parse(store(), drvDeep); + auto * p = std::get_if<NixStringContextElem::DrvDeep>(&elem); + ASSERT_TRUE(p); + ASSERT_EQ(p->drvPath, store().parseStorePath(drvDeep.substr(1))); + ASSERT_EQ(elem.to_string(store()), drvDeep); +} + +TEST_F(NixStringContextElemTest, built) { + std::string_view built = "!foo!/nix/store/g1w7hy3qg1w7hy3qg1w7hy3qg1w7hy3q-x.drv"; + auto elem = NixStringContextElem::parse(store(), built); + auto * p = std::get_if<NixStringContextElem::Built>(&elem); + ASSERT_TRUE(p); + ASSERT_EQ(p->output, "foo"); + ASSERT_EQ(p->drvPath, store().parseStorePath(built.substr(5))); + ASSERT_EQ(elem.to_string(store()), built); +} + +} + +namespace rc { +using namespace nix; + +Gen<NixStringContextElem::Opaque> Arbitrary<NixStringContextElem::Opaque>::arbitrary() +{ + return gen::just(NixStringContextElem::Opaque { + .path = *gen::arbitrary<StorePath>(), + }); +} + +Gen<NixStringContextElem::DrvDeep> Arbitrary<NixStringContextElem::DrvDeep>::arbitrary() +{ + return gen::just(NixStringContextElem::DrvDeep { + .drvPath = *gen::arbitrary<StorePath>(), + }); +} + +Gen<NixStringContextElem::Built> Arbitrary<NixStringContextElem::Built>::arbitrary() +{ + return gen::just(NixStringContextElem::Built { + .drvPath = *gen::arbitrary<StorePath>(), + .output = (*gen::arbitrary<StorePathName>()).name, + }); +} + +Gen<NixStringContextElem> Arbitrary<NixStringContextElem>::arbitrary() +{ + switch (*gen::inRange<uint8_t>(0, 2)) { + case 0: + return gen::just<NixStringContextElem>(*gen::arbitrary<NixStringContextElem::Opaque>()); + case 1: + return gen::just<NixStringContextElem>(*gen::arbitrary<NixStringContextElem::DrvDeep>()); + default: + return gen::just<NixStringContextElem>(*gen::arbitrary<NixStringContextElem::Built>()); + } +} + +} + +namespace nix { + +RC_GTEST_FIXTURE_PROP( + NixStringContextElemTest, + prop_round_rip, + (const NixStringContextElem & o)) +{ + RC_ASSERT(o == NixStringContextElem::parse(store(), o.to_string(store()))); +} + +} diff --git a/src/libexpr/tests/value/context.hh b/src/libexpr/tests/value/context.hh new file mode 100644 index 000000000..54d21760e --- /dev/null +++ b/src/libexpr/tests/value/context.hh @@ -0,0 +1,30 @@ +#pragma once + +#include <rapidcheck/gen/Arbitrary.h> + +#include <value/context.hh> + +namespace rc { +using namespace nix; + +template<> +struct Arbitrary<NixStringContextElem::Opaque> { + static Gen<NixStringContextElem::Opaque> arbitrary(); +}; + +template<> +struct Arbitrary<NixStringContextElem::Built> { + static Gen<NixStringContextElem::Built> arbitrary(); +}; + +template<> +struct Arbitrary<NixStringContextElem::DrvDeep> { + static Gen<NixStringContextElem::DrvDeep> arbitrary(); +}; + +template<> +struct Arbitrary<NixStringContextElem> { + static Gen<NixStringContextElem> arbitrary(); +}; + +} diff --git a/src/libexpr/value-to-json.cc b/src/libexpr/value-to-json.cc index 03504db61..c35c876e3 100644 --- a/src/libexpr/value-to-json.cc +++ b/src/libexpr/value-to-json.cc @@ -1,81 +1,83 @@ #include "value-to-json.hh" -#include "json.hh" #include "eval-inline.hh" #include "util.hh" +#include "store-api.hh" #include <cstdlib> #include <iomanip> +#include <nlohmann/json.hpp> namespace nix { - -void printValueAsJSON(EvalState & state, bool strict, - Value & v, const PosIdx pos, JSONPlaceholder & out, PathSet & context) +using json = nlohmann::json; +json printValueAsJSON(EvalState & state, bool strict, + Value & v, const PosIdx pos, PathSet & context, bool copyToStore) { checkInterrupt(); if (strict) state.forceValue(v, pos); + json out; + switch (v.type()) { case nInt: - out.write(v.integer); + out = v.integer; break; case nBool: - out.write(v.boolean); + out = v.boolean; break; case nString: copyContext(v, context); - out.write(v.string.s); + out = v.string.s; break; case nPath: - out.write(state.copyPathToStore(context, v.path)); + if (copyToStore) + out = state.store->printStorePath(state.copyPathToStore(context, v.path)); + else + out = v.path; break; case nNull: - out.write(nullptr); break; case nAttrs: { auto maybeString = state.tryAttrsToString(pos, v, context, false, false); if (maybeString) { - out.write(*maybeString); + out = *maybeString; break; } auto i = v.attrs->find(state.sOutPath); if (i == v.attrs->end()) { - auto obj(out.object()); + out = json::object(); StringSet names; for (auto & j : *v.attrs) names.emplace(state.symbols[j.name]); for (auto & j : names) { Attr & a(*v.attrs->find(state.symbols.create(j))); - auto placeholder(obj.placeholder(j)); - printValueAsJSON(state, strict, *a.value, a.pos, placeholder, context); + out[j] = printValueAsJSON(state, strict, *a.value, a.pos, context, copyToStore); } } else - printValueAsJSON(state, strict, *i->value, i->pos, out, context); + return printValueAsJSON(state, strict, *i->value, i->pos, context, copyToStore); break; } case nList: { - auto list(out.list()); - for (auto elem : v.listItems()) { - auto placeholder(list.placeholder()); - printValueAsJSON(state, strict, *elem, pos, placeholder, context); - } + out = json::array(); + for (auto elem : v.listItems()) + out.push_back(printValueAsJSON(state, strict, *elem, pos, context, copyToStore)); break; } case nExternal: - v.external->printValueAsJSON(state, strict, out, context); + return v.external->printValueAsJSON(state, strict, context, copyToStore); break; case nFloat: - out.write(v.fpoint); + out = v.fpoint; break; case nThunk: @@ -88,17 +90,17 @@ void printValueAsJSON(EvalState & state, bool strict, state.debugThrowLastTrace(e); throw e; } + return out; } void printValueAsJSON(EvalState & state, bool strict, - Value & v, const PosIdx pos, std::ostream & str, PathSet & context) + Value & v, const PosIdx pos, std::ostream & str, PathSet & context, bool copyToStore) { - JSONPlaceholder out(str); - printValueAsJSON(state, strict, v, pos, out, context); + str << printValueAsJSON(state, strict, v, pos, context, copyToStore); } -void ExternalValueBase::printValueAsJSON(EvalState & state, bool strict, - JSONPlaceholder & out, PathSet & context) const +json ExternalValueBase::printValueAsJSON(EvalState & state, bool strict, + PathSet & context, bool copyToStore) const { state.debugThrowLastTrace(TypeError("cannot convert %1% to JSON", showType())); } diff --git a/src/libexpr/value-to-json.hh b/src/libexpr/value-to-json.hh index c020a817a..22f26b790 100644 --- a/src/libexpr/value-to-json.hh +++ b/src/libexpr/value-to-json.hh @@ -5,15 +5,14 @@ #include <string> #include <map> +#include <nlohmann/json_fwd.hpp> namespace nix { -class JSONPlaceholder; +nlohmann::json printValueAsJSON(EvalState & state, bool strict, + Value & v, const PosIdx pos, PathSet & context, bool copyToStore = true); void printValueAsJSON(EvalState & state, bool strict, - Value & v, const PosIdx pos, JSONPlaceholder & out, PathSet & context); - -void printValueAsJSON(EvalState & state, bool strict, - Value & v, const PosIdx pos, std::ostream & str, PathSet & context); + Value & v, const PosIdx pos, std::ostream & str, PathSet & context, bool copyToStore = true); } diff --git a/src/libexpr/value-to-xml.cc b/src/libexpr/value-to-xml.cc index 7c3bf9492..341c8922f 100644 --- a/src/libexpr/value-to-xml.cc +++ b/src/libexpr/value-to-xml.cc @@ -24,9 +24,10 @@ static void printValueAsXML(EvalState & state, bool strict, bool location, static void posToXML(EvalState & state, XMLAttrs & xmlAttrs, const Pos & pos) { - xmlAttrs["path"] = pos.file; - xmlAttrs["line"] = (format("%1%") % pos.line).str(); - xmlAttrs["column"] = (format("%1%") % pos.column).str(); + if (auto path = std::get_if<Path>(&pos.origin)) + xmlAttrs["path"] = *path; + xmlAttrs["line"] = fmt("%1%", pos.line); + xmlAttrs["column"] = fmt("%1%", pos.column); } @@ -63,7 +64,7 @@ static void printValueAsXML(EvalState & state, bool strict, bool location, switch (v.type()) { case nInt: - doc.writeEmptyElement("int", singletonAttrs("value", (format("%1%") % v.integer).str())); + doc.writeEmptyElement("int", singletonAttrs("value", fmt("%1%", v.integer))); break; case nBool: @@ -155,7 +156,7 @@ static void printValueAsXML(EvalState & state, bool strict, bool location, break; case nFloat: - doc.writeEmptyElement("float", singletonAttrs("value", (format("%1%") % v.fpoint).str())); + doc.writeEmptyElement("float", singletonAttrs("value", fmt("%1%", v.fpoint))); break; case nThunk: diff --git a/src/libexpr/value.hh b/src/libexpr/value.hh index 2008df74d..508dbe218 100644 --- a/src/libexpr/value.hh +++ b/src/libexpr/value.hh @@ -3,10 +3,12 @@ #include <cassert> #include "symbol-table.hh" +#include "value/context.hh" #if HAVE_BOEHMGC #include <gc/gc_allocator.h> #endif +#include <nlohmann/json_fwd.hpp> namespace nix { @@ -62,13 +64,10 @@ class StorePath; class Store; class EvalState; class XMLWriter; -class JSONPlaceholder; typedef int64_t NixInt; typedef double NixFloat; -typedef std::pair<StorePath, std::string> NixStringContextElem; -typedef std::vector<NixStringContextElem> NixStringContext; /* External values must descend from ExternalValueBase, so that * type-agnostic nix functions (e.g. showType) can be implemented @@ -98,8 +97,8 @@ class ExternalValueBase virtual bool operator ==(const ExternalValueBase & b) const; /* Print the value as JSON. Defaults to unconvertable, i.e. throws an error */ - virtual void printValueAsJSON(EvalState & state, bool strict, - JSONPlaceholder & out, PathSet & context) const; + virtual nlohmann::json printValueAsJSON(EvalState & state, bool strict, + PathSet & context, bool copyToStore = true) const; /* Print the value as XML. Defaults to unevaluated */ virtual void printValueAsXML(EvalState & state, bool strict, bool location, diff --git a/src/libexpr/value/context.cc b/src/libexpr/value/context.cc new file mode 100644 index 000000000..61d9c53df --- /dev/null +++ b/src/libexpr/value/context.cc @@ -0,0 +1,67 @@ +#include "value/context.hh" +#include "store-api.hh" + +#include <optional> + +namespace nix { + +NixStringContextElem NixStringContextElem::parse(const Store & store, std::string_view s0) +{ + std::string_view s = s0; + + if (s.size() == 0) { + throw BadNixStringContextElem(s0, + "String context element should never be an empty string"); + } + switch (s.at(0)) { + case '!': { + s = s.substr(1); // advance string to parse after first ! + size_t index = s.find("!"); + // This makes index + 1 safe. Index can be the length (one after index + // of last character), so given any valid character index --- a + // successful find --- we can add one. + if (index == std::string_view::npos) { + throw BadNixStringContextElem(s0, + "String content element beginning with '!' should have a second '!'"); + } + return NixStringContextElem::Built { + .drvPath = store.parseStorePath(s.substr(index + 1)), + .output = std::string(s.substr(0, index)), + }; + } + case '=': { + return NixStringContextElem::DrvDeep { + .drvPath = store.parseStorePath(s.substr(1)), + }; + } + default: { + return NixStringContextElem::Opaque { + .path = store.parseStorePath(s), + }; + } + } +} + +std::string NixStringContextElem::to_string(const Store & store) const { + return std::visit(overloaded { + [&](const NixStringContextElem::Built & b) { + std::string res; + res += '!'; + res += b.output; + res += '!'; + res += store.printStorePath(b.drvPath); + return res; + }, + [&](const NixStringContextElem::DrvDeep & d) { + std::string res; + res += '='; + res += store.printStorePath(d.drvPath); + return res; + }, + [&](const NixStringContextElem::Opaque & o) { + return store.printStorePath(o.path); + }, + }, raw()); +} + +} diff --git a/src/libexpr/value/context.hh b/src/libexpr/value/context.hh new file mode 100644 index 000000000..721563cba --- /dev/null +++ b/src/libexpr/value/context.hh @@ -0,0 +1,97 @@ +#pragma once + +#include "util.hh" +#include "comparator.hh" +#include "path.hh" + +#include <variant> + +#include <nlohmann/json_fwd.hpp> + +namespace nix { + +class BadNixStringContextElem : public Error +{ +public: + std::string_view raw; + + template<typename... Args> + BadNixStringContextElem(std::string_view raw_, const Args & ... args) + : Error("") + { + raw = raw_; + auto hf = hintfmt(args...); + err.msg = hintfmt("Bad String Context element: %1%: %2%", normaltxt(hf.str()), raw); + } +}; + +class Store; + +/* Plain opaque path to some store object. + + Encoded as just the path: ‘<path>’. +*/ +struct NixStringContextElem_Opaque { + StorePath path; + + GENERATE_CMP(NixStringContextElem_Opaque, me->path); +}; + +/* Path to a derivation and its entire build closure. + + The path doesn't just refer to derivation itself and its closure, but + also all outputs of all derivations in that closure (including the + root derivation). + + Encoded in the form ‘=<drvPath>’. +*/ +struct NixStringContextElem_DrvDeep { + StorePath drvPath; + + GENERATE_CMP(NixStringContextElem_DrvDeep, me->drvPath); +}; + +/* Derivation output. + + Encoded in the form ‘!<output>!<drvPath>’. +*/ +struct NixStringContextElem_Built { + StorePath drvPath; + std::string output; + + GENERATE_CMP(NixStringContextElem_Built, me->drvPath, me->output); +}; + +using _NixStringContextElem_Raw = std::variant< + NixStringContextElem_Opaque, + NixStringContextElem_DrvDeep, + NixStringContextElem_Built +>; + +struct NixStringContextElem : _NixStringContextElem_Raw { + using Raw = _NixStringContextElem_Raw; + using Raw::Raw; + + using Opaque = NixStringContextElem_Opaque; + using DrvDeep = NixStringContextElem_DrvDeep; + using Built = NixStringContextElem_Built; + + inline const Raw & raw() const { + return static_cast<const Raw &>(*this); + } + inline Raw & raw() { + return static_cast<Raw &>(*this); + } + + /* Decode a context string, one of: + - ‘<path>’ + - ‘=<path>’ + - ‘!<name>!<path>’ + */ + static NixStringContextElem parse(const Store & store, std::string_view s); + std::string to_string(const Store & store) const; +}; + +typedef std::vector<NixStringContextElem> NixStringContext; + +} diff --git a/src/libfetchers/fetch-settings.hh b/src/libfetchers/fetch-settings.hh index 6452143a1..7049dea30 100644 --- a/src/libfetchers/fetch-settings.hh +++ b/src/libfetchers/fetch-settings.hh @@ -57,7 +57,7 @@ struct FetchSettings : public Config ``` This example specifies three tokens, one each for accessing - github.com, gitlab.mycompany.com, and sourceforge.net. + github.com, gitlab.mycompany.com, and gitlab.com. The `input.foo` uses the "gitlab" fetcher, which might requires specifying the token type along with the token @@ -71,7 +71,12 @@ struct FetchSettings : public Config "Whether to warn about dirty Git/Mercurial trees."}; Setting<std::string> flakeRegistry{this, "https://channels.nixos.org/flake-registry.json", "flake-registry", - "Path or URI of the global flake registry."}; + R"( + Path or URI of the global flake registry. + + When empty, disables the global flake registry. + )"}; + Setting<bool> useRegistries{this, true, "use-registries", "Whether to use flake registries to resolve flake references."}; diff --git a/src/libfetchers/fetchers.cc b/src/libfetchers/fetchers.cc index 6957d2da4..c767e72e5 100644 --- a/src/libfetchers/fetchers.cc +++ b/src/libfetchers/fetchers.cc @@ -266,7 +266,7 @@ std::optional<time_t> Input::getLastModified() const return {}; } -ParsedURL InputScheme::toURL(const Input & input) +ParsedURL InputScheme::toURL(const Input & input) const { throw Error("don't know how to convert input '%s' to a URL", attrsToJSON(input.attrs)); } @@ -274,7 +274,7 @@ ParsedURL InputScheme::toURL(const Input & input) Input InputScheme::applyOverrides( const Input & input, std::optional<std::string> ref, - std::optional<Hash> rev) + std::optional<Hash> rev) const { if (ref) throw Error("don't know how to set branch/tag name of input '%s' to '%s'", input.to_string(), *ref); @@ -293,7 +293,7 @@ void InputScheme::markChangedFile(const Input & input, std::string_view file, st assert(false); } -void InputScheme::clone(const Input & input, const Path & destDir) +void InputScheme::clone(const Input & input, const Path & destDir) const { throw Error("do not know how to clone input '%s'", input.to_string()); } diff --git a/src/libfetchers/fetchers.hh b/src/libfetchers/fetchers.hh index bc9a76b0b..95c0f5974 100644 --- a/src/libfetchers/fetchers.hh +++ b/src/libfetchers/fetchers.hh @@ -63,6 +63,11 @@ public: one that contains a commit hash or content hash. */ bool isLocked() const { return locked; } + /* Check whether the input carries all necessary info required + for cache insertion and substitution. + These fields are used to uniquely identify cached trees + within the "tarball TTL" window without necessarily + indicating that the input's origin is unchanged. */ bool hasAllInfo() const; bool operator ==(const Input & other) const; @@ -107,26 +112,25 @@ public: * recognized. The Input object contains the information the fetcher * needs to actually perform the "fetch()" when called. */ - struct InputScheme { virtual ~InputScheme() { } - virtual std::optional<Input> inputFromURL(const ParsedURL & url) = 0; + virtual std::optional<Input> inputFromURL(const ParsedURL & url) const = 0; - virtual std::optional<Input> inputFromAttrs(const Attrs & attrs) = 0; + virtual std::optional<Input> inputFromAttrs(const Attrs & attrs) const = 0; - virtual ParsedURL toURL(const Input & input); + virtual ParsedURL toURL(const Input & input) const; - virtual bool hasAllInfo(const Input & input) = 0; + virtual bool hasAllInfo(const Input & input) const = 0; virtual Input applyOverrides( const Input & input, std::optional<std::string> ref, - std::optional<Hash> rev); + std::optional<Hash> rev) const; - virtual void clone(const Input & input, const Path & destDir); + virtual void clone(const Input & input, const Path & destDir) const; virtual std::optional<Path> getSourcePath(const Input & input); diff --git a/src/libfetchers/git.cc b/src/libfetchers/git.cc index 7d01aaa7a..1da8c9609 100644 --- a/src/libfetchers/git.cc +++ b/src/libfetchers/git.cc @@ -18,6 +18,7 @@ using namespace std::string_literals; namespace nix::fetchers { + namespace { // Explicit initial branch of our bare repo to suppress warnings from new version of git. @@ -26,23 +27,23 @@ namespace { // old version of git, which will ignore unrecognized `-c` options. const std::string gitInitialBranch = "__nix_dummy_branch"; -bool isCacheFileWithinTtl(const time_t now, const struct stat & st) +bool isCacheFileWithinTtl(time_t now, const struct stat & st) { return st.st_mtime + settings.tarballTtl > now; } -bool touchCacheFile(const Path& path, const time_t& touch_time) +bool touchCacheFile(const Path & path, time_t touch_time) { - struct timeval times[2]; - times[0].tv_sec = touch_time; - times[0].tv_usec = 0; - times[1].tv_sec = touch_time; - times[1].tv_usec = 0; + struct timeval times[2]; + times[0].tv_sec = touch_time; + times[0].tv_usec = 0; + times[1].tv_sec = touch_time; + times[1].tv_usec = 0; - return lutimes(path.c_str(), times) == 0; + return lutimes(path.c_str(), times) == 0; } -Path getCachePath(std::string key) +Path getCachePath(std::string_view key) { return getCacheDir() + "/nix/gitv3/" + hashString(htSHA256, key).to_string(Base32, false); @@ -57,13 +58,12 @@ Path getCachePath(std::string key) // ... std::optional<std::string> readHead(const Path & path) { - auto [exit_code, output] = runProgram(RunOptions { + auto [status, output] = runProgram(RunOptions { .program = "git", + // FIXME: use 'HEAD' to avoid returning all refs .args = {"ls-remote", "--symref", path}, }); - if (exit_code != 0) { - return std::nullopt; - } + if (status != 0) return std::nullopt; std::string_view line = output; line = line.substr(0, line.find("\n")); @@ -82,12 +82,11 @@ std::optional<std::string> readHead(const Path & path) } // Persist the HEAD ref from the remote repo in the local cached repo. -bool storeCachedHead(const std::string& actualUrl, const std::string& headRef) +bool storeCachedHead(const std::string & actualUrl, const std::string & headRef) { Path cacheDir = getCachePath(actualUrl); - auto gitDir = "."; try { - runProgram("git", true, { "-C", cacheDir, "--git-dir", gitDir, "symbolic-ref", "--", "HEAD", headRef }); + runProgram("git", true, { "-C", cacheDir, "--git-dir", ".", "symbolic-ref", "--", "HEAD", headRef }); } catch (ExecError &e) { if (!WIFEXITED(e.status)) throw; return false; @@ -96,7 +95,7 @@ bool storeCachedHead(const std::string& actualUrl, const std::string& headRef) return true; } -std::optional<std::string> readHeadCached(const std::string& actualUrl) +std::optional<std::string> readHeadCached(const std::string & actualUrl) { // Create a cache path to store the branch of the HEAD ref. Append something // in front of the URL to prevent collision with the repository itself. @@ -110,16 +109,15 @@ std::optional<std::string> readHeadCached(const std::string& actualUrl) cachedRef = readHead(cacheDir); if (cachedRef != std::nullopt && *cachedRef != gitInitialBranch && - isCacheFileWithinTtl(now, st)) { + isCacheFileWithinTtl(now, st)) + { debug("using cached HEAD ref '%s' for repo '%s'", *cachedRef, actualUrl); return cachedRef; } } auto ref = readHead(actualUrl); - if (ref) { - return ref; - } + if (ref) return ref; if (cachedRef) { // If the cached git ref is expired in fetch() below, and the 'git fetch' @@ -250,7 +248,7 @@ std::pair<StorePath, Input> fetchFromWorkdir(ref<Store> store, Input & input, co struct GitInputScheme : InputScheme { - std::optional<Input> inputFromURL(const ParsedURL & url) override + std::optional<Input> inputFromURL(const ParsedURL & url) const override { if (url.scheme != "git" && url.scheme != "git+http" && @@ -265,10 +263,10 @@ struct GitInputScheme : InputScheme Attrs attrs; attrs.emplace("type", "git"); - for (auto &[name, value] : url.query) { + for (auto & [name, value] : url.query) { if (name == "rev" || name == "ref") attrs.emplace(name, value); - else if (name == "shallow" || name == "submodules") + else if (name == "shallow" || name == "submodules" || name == "allRefs") attrs.emplace(name, Explicit<bool> { value == "1" }); else url2.query.emplace(name, value); @@ -279,7 +277,7 @@ struct GitInputScheme : InputScheme return inputFromAttrs(attrs); } - std::optional<Input> inputFromAttrs(const Attrs & attrs) override + std::optional<Input> inputFromAttrs(const Attrs & attrs) const override { if (maybeGetStrAttr(attrs, "type") != "git") return {}; @@ -302,7 +300,7 @@ struct GitInputScheme : InputScheme return input; } - ParsedURL toURL(const Input & input) override + ParsedURL toURL(const Input & input) const override { auto url = parseURL(getStrAttr(input.attrs, "url")); if (url.scheme != "git") url.scheme = "git+" + url.scheme; @@ -313,7 +311,7 @@ struct GitInputScheme : InputScheme return url; } - bool hasAllInfo(const Input & input) override + bool hasAllInfo(const Input & input) const override { bool maybeDirty = !input.getRef(); bool shallow = maybeGetBoolAttr(input.attrs, "shallow").value_or(false); @@ -325,7 +323,7 @@ struct GitInputScheme : InputScheme Input applyOverrides( const Input & input, std::optional<std::string> ref, - std::optional<Hash> rev) override + std::optional<Hash> rev) const override { auto res(input); if (rev) res.attrs.insert_or_assign("rev", rev->gitRev()); @@ -335,7 +333,7 @@ struct GitInputScheme : InputScheme return res; } - void clone(const Input & input, const Path & destDir) override + void clone(const Input & input, const Path & destDir) const override { auto [isLocal, actualUrl] = getActualUrl(input); @@ -370,7 +368,7 @@ struct GitInputScheme : InputScheme auto gitDir = ".git"; runProgram("git", true, - { "-C", *sourcePath, "--git-dir", gitDir, "add", "--force", "--intent-to-add", "--", std::string(file) }); + { "-C", *sourcePath, "--git-dir", gitDir, "add", "--intent-to-add", "--", std::string(file) }); if (commitMsg) runProgram("git", true, @@ -485,6 +483,10 @@ struct GitInputScheme : InputScheme } input.attrs.insert_or_assign("ref", *head); unlockedAttrs.insert_or_assign("ref", *head); + } else { + if (!input.getRev()) { + unlockedAttrs.insert_or_assign("ref", input.getRef().value()); + } } if (auto res = getCache()->lookup(store, unlockedAttrs)) { @@ -599,9 +601,9 @@ struct GitInputScheme : InputScheme { throw Error( "Cannot find Git revision '%s' in ref '%s' of repository '%s'! " - "Please make sure that the " ANSI_BOLD "rev" ANSI_NORMAL " exists on the " - ANSI_BOLD "ref" ANSI_NORMAL " you've specified or add " ANSI_BOLD - "allRefs = true;" ANSI_NORMAL " to " ANSI_BOLD "fetchGit" ANSI_NORMAL ".", + "Please make sure that the " ANSI_BOLD "rev" ANSI_NORMAL " exists on the " + ANSI_BOLD "ref" ANSI_NORMAL " you've specified or add " ANSI_BOLD + "allRefs = true;" ANSI_NORMAL " to " ANSI_BOLD "fetchGit" ANSI_NORMAL ".", input.getRev()->gitRev(), *input.getRef(), actualUrl @@ -613,15 +615,42 @@ struct GitInputScheme : InputScheme AutoDelete delTmpGitDir(tmpGitDir, true); runProgram("git", true, { "-c", "init.defaultBranch=" + gitInitialBranch, "init", tmpDir, "--separate-git-dir", tmpGitDir }); - // TODO: repoDir might lack the ref (it only checks if rev - // exists, see FIXME above) so use a big hammer and fetch - // everything to ensure we get the rev. - runProgram("git", true, { "-C", tmpDir, "fetch", "--quiet", "--force", - "--update-head-ok", "--", repoDir, "refs/*:refs/*" }); + + { + // TODO: repoDir might lack the ref (it only checks if rev + // exists, see FIXME above) so use a big hammer and fetch + // everything to ensure we get the rev. + Activity act(*logger, lvlTalkative, actUnknown, fmt("making temporary clone of '%s'", repoDir)); + runProgram("git", true, { "-C", tmpDir, "fetch", "--quiet", "--force", + "--update-head-ok", "--", repoDir, "refs/*:refs/*" }); + } runProgram("git", true, { "-C", tmpDir, "checkout", "--quiet", input.getRev()->gitRev() }); - runProgram("git", true, { "-C", tmpDir, "remote", "add", "origin", actualUrl }); - runProgram("git", true, { "-C", tmpDir, "submodule", "--quiet", "update", "--init", "--recursive" }); + + /* Ensure that we use the correct origin for fetching + submodules. This matters for submodules with relative + URLs. */ + if (isLocal) { + writeFile(tmpGitDir + "/config", readFile(repoDir + "/" + gitDir + "/config")); + + /* Restore the config.bare setting we may have just + copied erroneously from the user's repo. */ + runProgram("git", true, { "-C", tmpDir, "config", "core.bare", "false" }); + } else + runProgram("git", true, { "-C", tmpDir, "config", "remote.origin.url", actualUrl }); + + /* As an optimisation, copy the modules directory of the + source repo if it exists. */ + auto modulesPath = repoDir + "/" + gitDir + "/modules"; + if (pathExists(modulesPath)) { + Activity act(*logger, lvlTalkative, actUnknown, fmt("copying submodules of '%s'", actualUrl)); + runProgram("cp", true, { "-R", "--", modulesPath, tmpGitDir + "/modules" }); + } + + { + Activity act(*logger, lvlTalkative, actUnknown, fmt("fetching submodules of '%s'", actualUrl)); + runProgram("git", true, { "-C", tmpDir, "submodule", "--quiet", "update", "--init", "--recursive" }); + } filter = isNotDotGitDirectory; } else { diff --git a/src/libfetchers/github.cc b/src/libfetchers/github.cc index a491d82a6..1ed09d30d 100644 --- a/src/libfetchers/github.cc +++ b/src/libfetchers/github.cc @@ -26,11 +26,11 @@ std::regex hostRegex(hostRegexS, std::regex::ECMAScript); struct GitArchiveInputScheme : InputScheme { - virtual std::string type() = 0; + virtual std::string type() const = 0; virtual std::optional<std::pair<std::string, std::string>> accessHeaderFromToken(const std::string & token) const = 0; - std::optional<Input> inputFromURL(const ParsedURL & url) override + std::optional<Input> inputFromURL(const ParsedURL & url) const override { if (url.scheme != type()) return {}; @@ -100,7 +100,7 @@ struct GitArchiveInputScheme : InputScheme return input; } - std::optional<Input> inputFromAttrs(const Attrs & attrs) override + std::optional<Input> inputFromAttrs(const Attrs & attrs) const override { if (maybeGetStrAttr(attrs, "type") != type()) return {}; @@ -116,7 +116,7 @@ struct GitArchiveInputScheme : InputScheme return input; } - ParsedURL toURL(const Input & input) override + ParsedURL toURL(const Input & input) const override { auto owner = getStrAttr(input.attrs, "owner"); auto repo = getStrAttr(input.attrs, "repo"); @@ -132,7 +132,7 @@ struct GitArchiveInputScheme : InputScheme }; } - bool hasAllInfo(const Input & input) override + bool hasAllInfo(const Input & input) const override { return input.getRev() && maybeGetIntAttr(input.attrs, "lastModified"); } @@ -140,7 +140,7 @@ struct GitArchiveInputScheme : InputScheme Input applyOverrides( const Input & _input, std::optional<std::string> ref, - std::optional<Hash> rev) override + std::optional<Hash> rev) const override { auto input(_input); if (rev && ref) @@ -227,7 +227,7 @@ struct GitArchiveInputScheme : InputScheme struct GitHubInputScheme : GitArchiveInputScheme { - std::string type() override { return "github"; } + std::string type() const override { return "github"; } std::optional<std::pair<std::string, std::string>> accessHeaderFromToken(const std::string & token) const override { @@ -240,14 +240,29 @@ struct GitHubInputScheme : GitArchiveInputScheme return std::pair<std::string, std::string>("Authorization", fmt("token %s", token)); } + std::string getHost(const Input & input) const + { + return maybeGetStrAttr(input.attrs, "host").value_or("github.com"); + } + + std::string getOwner(const Input & input) const + { + return getStrAttr(input.attrs, "owner"); + } + + std::string getRepo(const Input & input) const + { + return getStrAttr(input.attrs, "repo"); + } + Hash getRevFromRef(nix::ref<Store> store, const Input & input) const override { - auto host = maybeGetStrAttr(input.attrs, "host").value_or("github.com"); + auto host = getHost(input); auto url = fmt( host == "github.com" ? "https://api.%s/repos/%s/%s/commits/%s" : "https://%s/api/v3/repos/%s/%s/commits/%s", - host, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"), *input.getRef()); + host, getOwner(input), getRepo(input), *input.getRef()); Headers headers = makeHeadersWithAuthTokens(host); @@ -262,25 +277,30 @@ struct GitHubInputScheme : GitArchiveInputScheme DownloadUrl getDownloadUrl(const Input & input) const override { - // FIXME: use regular /archive URLs instead? api.github.com - // might have stricter rate limits. - auto host = maybeGetStrAttr(input.attrs, "host").value_or("github.com"); - auto url = fmt( - host == "github.com" - ? "https://api.%s/repos/%s/%s/tarball/%s" - : "https://%s/api/v3/repos/%s/%s/tarball/%s", - host, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"), - input.getRev()->to_string(Base16, false)); + auto host = getHost(input); Headers headers = makeHeadersWithAuthTokens(host); + + // If we have no auth headers then we default to the public archive + // urls so we do not run into rate limits. + const auto urlFmt = + host != "github.com" + ? "https://%s/api/v3/repos/%s/%s/tarball/%s" + : headers.empty() + ? "https://%s/%s/%s/archive/%s.tar.gz" + : "https://api.%s/repos/%s/%s/tarball/%s"; + + const auto url = fmt(urlFmt, host, getOwner(input), getRepo(input), + input.getRev()->to_string(Base16, false)); + return DownloadUrl { url, headers }; } - void clone(const Input & input, const Path & destDir) override + void clone(const Input & input, const Path & destDir) const override { - auto host = maybeGetStrAttr(input.attrs, "host").value_or("github.com"); + auto host = getHost(input); Input::fromURL(fmt("git+https://%s/%s/%s.git", - host, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"))) + host, getOwner(input), getRepo(input))) .applyOverrides(input.getRef(), input.getRev()) .clone(destDir); } @@ -288,7 +308,7 @@ struct GitHubInputScheme : GitArchiveInputScheme struct GitLabInputScheme : GitArchiveInputScheme { - std::string type() override { return "gitlab"; } + std::string type() const override { return "gitlab"; } std::optional<std::pair<std::string, std::string>> accessHeaderFromToken(const std::string & token) const override { @@ -343,7 +363,7 @@ struct GitLabInputScheme : GitArchiveInputScheme return DownloadUrl { url, headers }; } - void clone(const Input & input, const Path & destDir) override + void clone(const Input & input, const Path & destDir) const override { auto host = maybeGetStrAttr(input.attrs, "host").value_or("gitlab.com"); // FIXME: get username somewhere @@ -356,7 +376,7 @@ struct GitLabInputScheme : GitArchiveInputScheme struct SourceHutInputScheme : GitArchiveInputScheme { - std::string type() override { return "sourcehut"; } + std::string type() const override { return "sourcehut"; } std::optional<std::pair<std::string, std::string>> accessHeaderFromToken(const std::string & token) const override { @@ -430,7 +450,7 @@ struct SourceHutInputScheme : GitArchiveInputScheme return DownloadUrl { url, headers }; } - void clone(const Input & input, const Path & destDir) override + void clone(const Input & input, const Path & destDir) const override { auto host = maybeGetStrAttr(input.attrs, "host").value_or("git.sr.ht"); Input::fromURL(fmt("git+https://%s/%s/%s", diff --git a/src/libfetchers/indirect.cc b/src/libfetchers/indirect.cc index 9288fc6cf..b99504a16 100644 --- a/src/libfetchers/indirect.cc +++ b/src/libfetchers/indirect.cc @@ -7,7 +7,7 @@ std::regex flakeRegex("[a-zA-Z][a-zA-Z0-9_-]*", std::regex::ECMAScript); struct IndirectInputScheme : InputScheme { - std::optional<Input> inputFromURL(const ParsedURL & url) override + std::optional<Input> inputFromURL(const ParsedURL & url) const override { if (url.scheme != "flake") return {}; @@ -50,7 +50,7 @@ struct IndirectInputScheme : InputScheme return input; } - std::optional<Input> inputFromAttrs(const Attrs & attrs) override + std::optional<Input> inputFromAttrs(const Attrs & attrs) const override { if (maybeGetStrAttr(attrs, "type") != "indirect") return {}; @@ -68,7 +68,7 @@ struct IndirectInputScheme : InputScheme return input; } - ParsedURL toURL(const Input & input) override + ParsedURL toURL(const Input & input) const override { ParsedURL url; url.scheme = "flake"; @@ -78,7 +78,7 @@ struct IndirectInputScheme : InputScheme return url; } - bool hasAllInfo(const Input & input) override + bool hasAllInfo(const Input & input) const override { return false; } @@ -86,7 +86,7 @@ struct IndirectInputScheme : InputScheme Input applyOverrides( const Input & _input, std::optional<std::string> ref, - std::optional<Hash> rev) override + std::optional<Hash> rev) const override { auto input(_input); if (rev) input.attrs.insert_or_assign("rev", rev->gitRev()); diff --git a/src/libfetchers/mercurial.cc b/src/libfetchers/mercurial.cc index 5c5671681..86e8f81f4 100644 --- a/src/libfetchers/mercurial.cc +++ b/src/libfetchers/mercurial.cc @@ -43,7 +43,7 @@ static std::string runHg(const Strings & args, const std::optional<std::string> struct MercurialInputScheme : InputScheme { - std::optional<Input> inputFromURL(const ParsedURL & url) override + std::optional<Input> inputFromURL(const ParsedURL & url) const override { if (url.scheme != "hg+http" && url.scheme != "hg+https" && @@ -69,7 +69,7 @@ struct MercurialInputScheme : InputScheme return inputFromAttrs(attrs); } - std::optional<Input> inputFromAttrs(const Attrs & attrs) override + std::optional<Input> inputFromAttrs(const Attrs & attrs) const override { if (maybeGetStrAttr(attrs, "type") != "hg") return {}; @@ -89,7 +89,7 @@ struct MercurialInputScheme : InputScheme return input; } - ParsedURL toURL(const Input & input) override + ParsedURL toURL(const Input & input) const override { auto url = parseURL(getStrAttr(input.attrs, "url")); url.scheme = "hg+" + url.scheme; @@ -98,7 +98,7 @@ struct MercurialInputScheme : InputScheme return url; } - bool hasAllInfo(const Input & input) override + bool hasAllInfo(const Input & input) const override { // FIXME: ugly, need to distinguish between dirty and clean // default trees. @@ -108,7 +108,7 @@ struct MercurialInputScheme : InputScheme Input applyOverrides( const Input & input, std::optional<std::string> ref, - std::optional<Hash> rev) override + std::optional<Hash> rev) const override { auto res(input); if (rev) res.attrs.insert_or_assign("rev", rev->gitRev()); diff --git a/src/libfetchers/path.cc b/src/libfetchers/path.cc index f0ef97da5..61541e69d 100644 --- a/src/libfetchers/path.cc +++ b/src/libfetchers/path.cc @@ -6,7 +6,7 @@ namespace nix::fetchers { struct PathInputScheme : InputScheme { - std::optional<Input> inputFromURL(const ParsedURL & url) override + std::optional<Input> inputFromURL(const ParsedURL & url) const override { if (url.scheme != "path") return {}; @@ -32,7 +32,7 @@ struct PathInputScheme : InputScheme return input; } - std::optional<Input> inputFromAttrs(const Attrs & attrs) override + std::optional<Input> inputFromAttrs(const Attrs & attrs) const override { if (maybeGetStrAttr(attrs, "type") != "path") return {}; @@ -54,7 +54,7 @@ struct PathInputScheme : InputScheme return input; } - ParsedURL toURL(const Input & input) override + ParsedURL toURL(const Input & input) const override { auto query = attrsToQuery(input.attrs); query.erase("path"); @@ -66,7 +66,7 @@ struct PathInputScheme : InputScheme }; } - bool hasAllInfo(const Input & input) override + bool hasAllInfo(const Input & input) const override { return true; } diff --git a/src/libfetchers/registry.cc b/src/libfetchers/registry.cc index acd1ff866..43c03beec 100644 --- a/src/libfetchers/registry.cc +++ b/src/libfetchers/registry.cc @@ -153,6 +153,9 @@ static std::shared_ptr<Registry> getGlobalRegistry(ref<Store> store) { static auto reg = [&]() { auto path = fetchSettings.flakeRegistry.get(); + if (path == "") { + return std::make_shared<Registry>(Registry::Global); // empty registry + } if (!hasPrefix(path, "/")) { auto storePath = downloadFile(store, path, "flake-registry.json", false).storePath; diff --git a/src/libfetchers/tarball.cc b/src/libfetchers/tarball.cc index 6c551bd93..e9686262a 100644 --- a/src/libfetchers/tarball.cc +++ b/src/libfetchers/tarball.cc @@ -185,7 +185,7 @@ struct CurlInputScheme : InputScheme virtual bool isValidURL(const ParsedURL & url) const = 0; - std::optional<Input> inputFromURL(const ParsedURL & url) override + std::optional<Input> inputFromURL(const ParsedURL & url) const override { if (!isValidURL(url)) return std::nullopt; @@ -203,7 +203,7 @@ struct CurlInputScheme : InputScheme return input; } - std::optional<Input> inputFromAttrs(const Attrs & attrs) override + std::optional<Input> inputFromAttrs(const Attrs & attrs) const override { auto type = maybeGetStrAttr(attrs, "type"); if (type != inputType()) return {}; @@ -220,16 +220,17 @@ struct CurlInputScheme : InputScheme return input; } - ParsedURL toURL(const Input & input) override + ParsedURL toURL(const Input & input) const override { auto url = parseURL(getStrAttr(input.attrs, "url")); - // NAR hashes are preferred over file hashes since tar/zip files // don't have a canonical representation. + // NAR hashes are preferred over file hashes since tar/zip + // files don't have a canonical representation. if (auto narHash = input.getNarHash()) url.query.insert_or_assign("narHash", narHash->to_string(SRI, true)); return url; } - bool hasAllInfo(const Input & input) override + bool hasAllInfo(const Input & input) const override { return true; } diff --git a/src/libmain/common-args.cc b/src/libmain/common-args.cc index 12f5403ea..f92920d18 100644 --- a/src/libmain/common-args.cc +++ b/src/libmain/common-args.cc @@ -32,6 +32,7 @@ MixCommonArgs::MixCommonArgs(const std::string & programName) addFlag({ .longName = "option", .description = "Set the Nix configuration setting *name* to *value* (overriding `nix.conf`).", + .category = miscCategory, .labels = {"name", "value"}, .handler = {[](std::string name, std::string value) { try { diff --git a/src/libmain/common-args.hh b/src/libmain/common-args.hh index 25453b8c6..f180d83ce 100644 --- a/src/libmain/common-args.hh +++ b/src/libmain/common-args.hh @@ -6,6 +6,7 @@ namespace nix { //static constexpr auto commonArgsCategory = "Miscellaneous common options"; static constexpr auto loggingCategory = "Logging-related options"; +static constexpr auto miscCategory = "Miscellaneous global options"; class MixCommonArgs : public virtual Args { diff --git a/src/libmain/loggers.cc b/src/libmain/loggers.cc index cdf23859b..cda5cb939 100644 --- a/src/libmain/loggers.cc +++ b/src/libmain/loggers.cc @@ -30,8 +30,11 @@ Logger * makeDefaultLogger() { return makeJSONLogger(*makeSimpleLogger(true)); case LogFormat::bar: return makeProgressBar(); - case LogFormat::barWithLogs: - return makeProgressBar(true); + case LogFormat::barWithLogs: { + auto logger = makeProgressBar(); + logger->setPrintBuildLogs(true); + return logger; + } default: abort(); } diff --git a/src/libmain/nix-main.pc.in b/src/libmain/nix-main.pc.in index 37b03dcd4..fb3ead6fa 100644 --- a/src/libmain/nix-main.pc.in +++ b/src/libmain/nix-main.pc.in @@ -6,4 +6,4 @@ Name: Nix Description: Nix Package Manager Version: @PACKAGE_VERSION@ Libs: -L${libdir} -lnixmain -Cflags: -I${includedir}/nix -std=c++17 +Cflags: -I${includedir}/nix -std=c++2a diff --git a/src/libmain/progress-bar.cc b/src/libmain/progress-bar.cc index f4306ab91..024259584 100644 --- a/src/libmain/progress-bar.cc +++ b/src/libmain/progress-bar.cc @@ -8,6 +8,7 @@ #include <map> #include <thread> #include <iostream> +#include <chrono> namespace nix { @@ -48,6 +49,7 @@ private: bool visible = true; ActivityId parent; std::optional<std::string> name; + std::chrono::time_point<std::chrono::steady_clock> startTime; }; struct ActivitiesByType @@ -79,22 +81,22 @@ private: std::condition_variable quitCV, updateCV; - bool printBuildLogs; + bool printBuildLogs = false; bool isTTY; public: - ProgressBar(bool printBuildLogs, bool isTTY) - : printBuildLogs(printBuildLogs) - , isTTY(isTTY) + ProgressBar(bool isTTY) + : isTTY(isTTY) { state_.lock()->active = isTTY; updateThread = std::thread([&]() { auto state(state_.lock()); + auto nextWakeup = std::chrono::milliseconds::max(); while (state->active) { if (!state->haveUpdate) - state.wait(updateCV); - draw(*state); + state.wait_for(updateCV, nextWakeup); + nextWakeup = draw(*state); state.wait_for(quitCV, std::chrono::milliseconds(50)); } }); @@ -118,18 +120,19 @@ public: updateThread.join(); } - bool isVerbose() override { + bool isVerbose() override + { return printBuildLogs; } - void log(Verbosity lvl, const FormatOrString & fs) override + void log(Verbosity lvl, std::string_view s) override { if (lvl > verbosity) return; auto state(state_.lock()); - log(*state, lvl, fs.s); + log(*state, lvl, s); } - void logEI(const ErrorInfo &ei) override + void logEI(const ErrorInfo & ei) override { auto state(state_.lock()); @@ -139,7 +142,7 @@ public: log(*state, ei.level, oss.str()); } - void log(State & state, Verbosity lvl, const std::string & s) + void log(State & state, Verbosity lvl, std::string_view s) { if (state.active) { writeToStderr("\r\e[K" + filterANSIEscapes(s, !isTTY) + ANSI_NORMAL "\n"); @@ -159,11 +162,13 @@ public: if (lvl <= verbosity && !s.empty() && type != actBuildWaiting) log(*state, lvl, s + "..."); - state->activities.emplace_back(ActInfo()); + state->activities.emplace_back(ActInfo { + .s = s, + .type = type, + .parent = parent, + .startTime = std::chrono::steady_clock::now() + }); auto i = std::prev(state->activities.end()); - i->s = s; - i->type = type; - i->parent = parent; state->its.emplace(act, i); state->activitiesByType[type].its.emplace(act, i); @@ -175,10 +180,12 @@ public: auto machineName = getS(fields, 1); if (machineName != "") i->s += fmt(" on " ANSI_BOLD "%s" ANSI_NORMAL, machineName); - auto curRound = getI(fields, 2); - auto nrRounds = getI(fields, 3); - if (nrRounds != 1) - i->s += fmt(" (round %d/%d)", curRound, nrRounds); + + // Used to be curRound and nrRounds, but the + // implementation was broken for a long time. + if (getI(fields, 2) != 1 || getI(fields, 3) != 1) { + throw Error("log message indicated repeating builds, but this is not currently implemented"); + } i->name = DrvName(name).name; } @@ -327,10 +334,12 @@ public: updateCV.notify_one(); } - void draw(State & state) + std::chrono::milliseconds draw(State & state) { + auto nextWakeup = std::chrono::milliseconds::max(); + state.haveUpdate = false; - if (!state.active) return; + if (!state.active) return nextWakeup; std::string line; @@ -341,12 +350,25 @@ public: line += "]"; } + auto now = std::chrono::steady_clock::now(); + if (!state.activities.empty()) { if (!status.empty()) line += " "; auto i = state.activities.rbegin(); - while (i != state.activities.rend() && (!i->visible || (i->s.empty() && i->lastLine.empty()))) + while (i != state.activities.rend()) { + if (i->visible && (!i->s.empty() || !i->lastLine.empty())) { + /* Don't show activities until some time has + passed, to avoid displaying very short + activities. */ + auto delay = std::chrono::milliseconds(10); + if (i->startTime + delay < now) + break; + else + nextWakeup = std::min(nextWakeup, std::chrono::duration_cast<std::chrono::milliseconds>(delay - (now - i->startTime))); + } ++i; + } if (i != state.activities.rend()) { line += i->s; @@ -366,6 +388,8 @@ public: if (width <= 0) width = std::numeric_limits<decltype(width)>::max(); writeToStderr("\r" + filterANSIEscapes(line, false, width) + ANSI_NORMAL + "\e[K"); + + return nextWakeup; } std::string getStatus(State & state) @@ -480,19 +504,21 @@ public: draw(*state); return s[0]; } + + void setPrintBuildLogs(bool printBuildLogs) override + { + this->printBuildLogs = printBuildLogs; + } }; -Logger * makeProgressBar(bool printBuildLogs) +Logger * makeProgressBar() { - return new ProgressBar( - printBuildLogs, - shouldANSI() - ); + return new ProgressBar(shouldANSI()); } -void startProgressBar(bool printBuildLogs) +void startProgressBar() { - logger = makeProgressBar(printBuildLogs); + logger = makeProgressBar(); } void stopProgressBar() diff --git a/src/libmain/progress-bar.hh b/src/libmain/progress-bar.hh index 7f0dafecf..3a76f8448 100644 --- a/src/libmain/progress-bar.hh +++ b/src/libmain/progress-bar.hh @@ -4,9 +4,9 @@ namespace nix { -Logger * makeProgressBar(bool printBuildLogs = false); +Logger * makeProgressBar(); -void startProgressBar(bool printBuildLogs = false); +void startProgressBar(); void stopProgressBar(); diff --git a/src/libmain/shared.cc b/src/libmain/shared.cc index 31454e49d..37664c065 100644 --- a/src/libmain/shared.cc +++ b/src/libmain/shared.cc @@ -4,6 +4,7 @@ #include "gc-store.hh" #include "util.hh" #include "loggers.hh" +#include "progress-bar.hh" #include <algorithm> #include <cctype> @@ -32,6 +33,7 @@ namespace nix { +char * * savedArgv; static bool gcWarning = true; @@ -82,8 +84,18 @@ void printMissing(ref<Store> store, const StorePathSet & willBuild, downloadSizeMiB, narSizeMiB); } - for (auto & i : willSubstitute) - printMsg(lvl, " %s", store->printStorePath(i)); + std::vector<const StorePath *> willSubstituteSorted = {}; + std::for_each(willSubstitute.begin(), willSubstitute.end(), + [&](const StorePath &p) { willSubstituteSorted.push_back(&p); }); + std::sort(willSubstituteSorted.begin(), willSubstituteSorted.end(), + [](const StorePath *lhs, const StorePath *rhs) { + if (lhs->name() == rhs->name()) + return lhs->to_string() < rhs->to_string(); + else + return lhs->name() < rhs->name(); + }); + for (auto p : willSubstituteSorted) + printMsg(lvl, " %s", store->printStorePath(*p)); } if (!unknown.empty()) { @@ -181,8 +193,9 @@ void initNix() /* Reset SIGCHLD to its default. */ struct sigaction act; sigemptyset(&act.sa_mask); - act.sa_handler = SIG_DFL; act.sa_flags = 0; + + act.sa_handler = SIG_DFL; if (sigaction(SIGCHLD, &act, 0)) throw SysError("resetting SIGCHLD"); @@ -194,9 +207,20 @@ void initNix() /* HACK: on darwin, we need can’t use sigprocmask with SIGWINCH. * Instead, add a dummy sigaction handler, and signalHandlerThread * can handle the rest. */ - struct sigaction sa; - sa.sa_handler = sigHandler; - if (sigaction(SIGWINCH, &sa, 0)) throw SysError("handling SIGWINCH"); + act.sa_handler = sigHandler; + if (sigaction(SIGWINCH, &act, 0)) throw SysError("handling SIGWINCH"); + + /* Disable SA_RESTART for interrupts, so that system calls on this thread + * error with EINTR like they do on Linux. + * Most signals on BSD systems default to SA_RESTART on, but Nix + * expects EINTR from syscalls to properly exit. */ + act.sa_handler = SIG_DFL; + if (sigaction(SIGINT, &act, 0)) throw SysError("handling SIGINT"); + if (sigaction(SIGTERM, &act, 0)) throw SysError("handling SIGTERM"); + if (sigaction(SIGHUP, &act, 0)) throw SysError("handling SIGHUP"); + if (sigaction(SIGPIPE, &act, 0)) throw SysError("handling SIGPIPE"); + if (sigaction(SIGQUIT, &act, 0)) throw SysError("handling SIGQUIT"); + if (sigaction(SIGTRAP, &act, 0)) throw SysError("handling SIGTRAP"); #endif /* Register a SIGSEGV handler to detect stack overflows. */ @@ -221,6 +245,7 @@ void initNix() #endif preloadNSS(); + initLibStore(); } @@ -332,7 +357,7 @@ void parseCmdLine(const std::string & programName, const Strings & args, void printVersion(const std::string & programName) { - std::cout << format("%1% (Nix) %2%") % programName % nixVersion << std::endl; + std::cout << fmt("%1% (Nix) %2%", programName, nixVersion) << std::endl; if (verbosity > lvlInfo) { Strings cfg; #if HAVE_BOEHMGC @@ -348,6 +373,7 @@ void printVersion(const std::string & programName) << "\n"; std::cout << "Store directory: " << settings.nixStore << "\n"; std::cout << "State directory: " << settings.nixStateDir << "\n"; + std::cout << "Data directory: " << settings.nixDataDir << "\n"; } throw Exit(); } @@ -388,8 +414,6 @@ int handleExceptions(const std::string & programName, std::function<void()> fun) return 1; } catch (BaseError & e) { logError(e.info()); - if (e.hasTrace() && !loggerSettings.showTrace.get()) - printError("(use '--show-trace' to show detailed location information)"); return e.status; } catch (std::bad_alloc & e) { printError(error + "out of memory"); @@ -410,6 +434,8 @@ RunPager::RunPager() if (!pager) pager = getenv("PAGER"); if (pager && ((std::string) pager == "" || (std::string) pager == "cat")) return; + stopProgressBar(); + Pipe toPager; toPager.create(); diff --git a/src/libmain/shared.hh b/src/libmain/shared.hh index 0cc56d47d..1715374a6 100644 --- a/src/libmain/shared.hh +++ b/src/libmain/shared.hh @@ -39,7 +39,6 @@ void printVersion(const std::string & programName); void printGCWarning(); class Store; -struct StorePathWithOutputs; void printMissing( ref<Store> store, @@ -113,5 +112,25 @@ struct PrintFreed /* Install a SIGSEGV handler to detect stack overflows. */ void detectStackOverflow(); +/* Pluggable behavior to run in case of a stack overflow. + + Default value: defaultStackOverflowHandler. + + This is called by the handler installed by detectStackOverflow(). + + This gives Nix library consumers a limit opportunity to report the error + condition. The handler should exit the process. + See defaultStackOverflowHandler() for a reference implementation. + + NOTE: Use with diligence, because this runs in the signal handler, with very + limited stack space and a potentially a corrupted heap, all while the failed + thread is blocked indefinitely. All functions called must be reentrant. */ +extern std::function<void(siginfo_t * info, void * ctx)> stackOverflowHandler; + +/* The default, robust implementation of stackOverflowHandler. + + Prints an error message directly to stderr using a syscall instead of the + logger. Exits the process immediately after. */ +void defaultStackOverflowHandler(siginfo_t * info, void * ctx); } diff --git a/src/libmain/stack.cc b/src/libmain/stack.cc index b0a4a4c5d..10f71c1dc 100644 --- a/src/libmain/stack.cc +++ b/src/libmain/stack.cc @@ -1,4 +1,5 @@ #include "error.hh" +#include "shared.hh" #include <cstring> #include <cstddef> @@ -29,9 +30,7 @@ static void sigsegvHandler(int signo, siginfo_t * info, void * ctx) ptrdiff_t diff = (char *) info->si_addr - sp; if (diff < 0) diff = -diff; if (diff < 4096) { - char msg[] = "error: stack overflow (possible infinite recursion)\n"; - [[gnu::unused]] auto res = write(2, msg, strlen(msg)); - _exit(1); // maybe abort instead? + nix::stackOverflowHandler(info, ctx); } } @@ -67,5 +66,12 @@ void detectStackOverflow() #endif } +std::function<void(siginfo_t * info, void * ctx)> stackOverflowHandler(defaultStackOverflowHandler); + +void defaultStackOverflowHandler(siginfo_t * info, void * ctx) { + char msg[] = "error: stack overflow (possible infinite recursion)\n"; + [[gnu::unused]] auto res = write(2, msg, strlen(msg)); + _exit(1); // maybe abort instead? +} } diff --git a/src/libstore/binary-cache-store.cc b/src/libstore/binary-cache-store.cc index 9226c4e19..751cf8c30 100644 --- a/src/libstore/binary-cache-store.cc +++ b/src/libstore/binary-cache-store.cc @@ -9,7 +9,6 @@ #include "remote-fs-accessor.hh" #include "nar-info-disk-cache.hh" #include "nar-accessor.hh" -#include "json.hh" #include "thread-pool.hh" #include "callback.hh" @@ -194,19 +193,12 @@ ref<const ValidPathInfo> BinaryCacheStore::addToStoreCommon( /* Optionally write a JSON file containing a listing of the contents of the NAR. */ if (writeNARListing) { - std::ostringstream jsonOut; - - { - JSONObject jsonRoot(jsonOut); - jsonRoot.attr("version", 1); - - { - auto res = jsonRoot.placeholder("root"); - listNar(res, ref<FSAccessor>(narAccessor), "", true); - } - } + nlohmann::json j = { + {"version", 1}, + {"root", listNar(ref<FSAccessor>(narAccessor), "", true)}, + }; - upsertFile(std::string(info.path.hashPart()) + ".ls", jsonOut.str(), "application/json"); + upsertFile(std::string(info.path.hashPart()) + ".ls", j.dump(), "application/json"); } /* Optionally maintain an index of DWARF debug info files @@ -331,6 +323,17 @@ bool BinaryCacheStore::isValidPathUncached(const StorePath & storePath) return fileExists(narInfoFileFor(storePath)); } +std::optional<StorePath> BinaryCacheStore::queryPathFromHashPart(const std::string & hashPart) +{ + auto pseudoPath = StorePath(hashPart + "-" + MissingName); + try { + auto info = queryPathInfo(pseudoPath); + return info->path; + } catch (InvalidPath &) { + return std::nullopt; + } +} + void BinaryCacheStore::narFromPath(const StorePath & storePath, Sink & sink) { auto info = queryPathInfo(storePath).cast<const NarInfo>(); @@ -343,7 +346,7 @@ void BinaryCacheStore::narFromPath(const StorePath & storePath, Sink & sink) try { getFile(info->url, *decompressor); } catch (NoSuchBinaryCacheFile & e) { - throw SubstituteGone(e.info()); + throw SubstituteGone(std::move(e.info())); } decompressor->finish(); @@ -367,7 +370,7 @@ void BinaryCacheStore::queryPathInfoUncached(const StorePath & storePath, auto callbackPtr = std::make_shared<decltype(callback)>(std::move(callback)); getFile(narInfoFile, - {[=](std::future<std::optional<std::string>> fut) { + {[=,this](std::future<std::optional<std::string>> fut) { try { auto data = fut.get(); @@ -499,22 +502,9 @@ void BinaryCacheStore::addSignatures(const StorePath & storePath, const StringSe writeNarInfo(narInfo); } -std::optional<std::string> BinaryCacheStore::getBuildLog(const StorePath & path) +std::optional<std::string> BinaryCacheStore::getBuildLogExact(const StorePath & path) { - auto drvPath = path; - - if (!path.isDerivation()) { - try { - auto info = queryPathInfo(path); - // FIXME: add a "Log" field to .narinfo - if (!info->deriver) return std::nullopt; - drvPath = *info->deriver; - } catch (InvalidPath &) { - return std::nullopt; - } - } - - auto logPath = "log/" + std::string(baseNameOf(printStorePath(drvPath))); + auto logPath = "log/" + std::string(baseNameOf(printStorePath(path))); debug("fetching build log from binary cache '%s/%s'", getUri(), logPath); diff --git a/src/libstore/binary-cache-store.hh b/src/libstore/binary-cache-store.hh index ca538b3cb..abd92a83c 100644 --- a/src/libstore/binary-cache-store.hh +++ b/src/libstore/binary-cache-store.hh @@ -95,8 +95,7 @@ public: void queryPathInfoUncached(const StorePath & path, Callback<std::shared_ptr<const ValidPathInfo>> callback) noexcept override; - std::optional<StorePath> queryPathFromHashPart(const std::string & hashPart) override - { unsupported("queryPathFromHashPart"); } + std::optional<StorePath> queryPathFromHashPart(const std::string & hashPart) override; void addToStore(const ValidPathInfo & info, Source & narSource, RepairFlag repair, CheckSigsFlag checkSigs) override; @@ -130,7 +129,7 @@ public: void addSignatures(const StorePath & storePath, const StringSet & sigs) override; - std::optional<std::string> getBuildLog(const StorePath & path) override; + std::optional<std::string> getBuildLogExact(const StorePath & path) override; void addBuildLog(const StorePath & drvPath, std::string_view log) override; diff --git a/src/libstore/build-result.hh b/src/libstore/build-result.hh index 24fb1f763..a5749cf33 100644 --- a/src/libstore/build-result.hh +++ b/src/libstore/build-result.hh @@ -5,7 +5,7 @@ #include <string> #include <chrono> - +#include <optional> namespace nix { @@ -78,6 +78,9 @@ struct BuildResult was repeated). */ time_t startTime = 0, stopTime = 0; + /* User and system CPU time the build took. */ + std::optional<std::chrono::microseconds> cpuUser, cpuSystem; + bool success() { return status == Built || status == Substituted || status == AlreadyValid || status == ResolvesToAlreadyValid; diff --git a/src/libstore/build/derivation-goal.cc b/src/libstore/build/derivation-goal.cc index 3fff2385f..38b73d531 100644 --- a/src/libstore/build/derivation-goal.cc +++ b/src/libstore/build/derivation-goal.cc @@ -7,7 +7,6 @@ #include "finally.hh" #include "util.hh" #include "archive.hh" -#include "json.hh" #include "compression.hh" #include "worker-protocol.hh" #include "topo-sort.hh" @@ -40,7 +39,6 @@ #include <sys/ioctl.h> #include <net/if.h> #include <netinet/ip.h> -#include <sys/personality.h> #include <sys/mman.h> #include <sched.h> #include <sys/param.h> @@ -65,7 +63,7 @@ namespace nix { DerivationGoal::DerivationGoal(const StorePath & drvPath, - const StringSet & wantedOutputs, Worker & worker, BuildMode buildMode) + const OutputsSpec & wantedOutputs, Worker & worker, BuildMode buildMode) : Goal(worker, DerivedPath::Built { .drvPath = drvPath, .outputs = wantedOutputs }) , useDerivation(true) , drvPath(drvPath) @@ -84,7 +82,7 @@ DerivationGoal::DerivationGoal(const StorePath & drvPath, DerivationGoal::DerivationGoal(const StorePath & drvPath, const BasicDerivation & drv, - const StringSet & wantedOutputs, Worker & worker, BuildMode buildMode) + const OutputsSpec & wantedOutputs, Worker & worker, BuildMode buildMode) : Goal(worker, DerivedPath::Built { .drvPath = drvPath, .outputs = wantedOutputs }) , useDerivation(false) , drvPath(drvPath) @@ -135,7 +133,7 @@ void DerivationGoal::killChild() void DerivationGoal::timedOut(Error && ex) { killChild(); - done(BuildResult::TimedOut, {}, ex); + done(BuildResult::TimedOut, {}, std::move(ex)); } @@ -144,18 +142,12 @@ void DerivationGoal::work() (this->*state)(); } -void DerivationGoal::addWantedOutputs(const StringSet & outputs) +void DerivationGoal::addWantedOutputs(const OutputsSpec & outputs) { - /* If we already want all outputs, there is nothing to do. */ - if (wantedOutputs.empty()) return; - - if (outputs.empty()) { - wantedOutputs.clear(); + auto newWanted = wantedOutputs.union_(outputs); + if (!newWanted.isSubsetOf(wantedOutputs)) needRestart = true; - } else - for (auto & i : outputs) - if (wantedOutputs.insert(i).second) - needRestart = true; + wantedOutputs = newWanted; } @@ -344,7 +336,7 @@ void DerivationGoal::gaveUpOnSubstitution() for (auto & i : dynamic_cast<Derivation *>(drv.get())->inputDrvs) { /* Ensure that pure, non-fixed-output derivations don't depend on impure derivations. */ - if (drv->type().isPure() && !drv->type().isFixed()) { + if (settings.isExperimentalFeatureEnabled(Xp::ImpureDerivations) && drv->type().isPure() && !drv->type().isFixed()) { auto inputDrv = worker.evalStore.readDerivation(i.first); if (!inputDrv.type().isPure()) throw Error("pure derivation '%s' depends on impure derivation '%s'", @@ -392,7 +384,7 @@ void DerivationGoal::repairClosure() auto outputs = queryDerivationOutputMap(); StorePathSet outputClosure; for (auto & i : outputs) { - if (!wantOutput(i.first, wantedOutputs)) continue; + if (!wantedOutputs.contains(i.first)) continue; worker.store.computeFSClosure(i.second, outputClosure); } @@ -424,7 +416,7 @@ void DerivationGoal::repairClosure() if (drvPath2 == outputsToDrv.end()) addWaitee(upcast_goal(worker.makePathSubstitutionGoal(i, Repair))); else - addWaitee(worker.makeDerivationGoal(drvPath2->second, StringSet(), bmRepair)); + addWaitee(worker.makeDerivationGoal(drvPath2->second, OutputsSpec::All(), bmRepair)); } if (waitees.empty()) { @@ -502,6 +494,14 @@ void DerivationGoal::inputsRealised() now-known results of dependencies. If so, we become a stub goal aliasing that resolved derivation goal. */ std::optional attempt = fullDrv.tryResolve(worker.store, inputDrvOutputs); + if (!attempt) { + /* TODO (impure derivations-induced tech debt) (see below): + The above attempt should have found it, but because we manage + inputDrvOutputs statefully, sometimes it gets out of sync with + the real source of truth (store). So we query the store + directly if there's a problem. */ + attempt = fullDrv.tryResolve(worker.store); + } assert(attempt); Derivation drvResolved { *std::move(attempt) }; @@ -528,13 +528,32 @@ void DerivationGoal::inputsRealised() /* Add the relevant output closures of the input derivation `i' as input paths. Only add the closures of output paths that are specified as inputs. */ - for (auto & j : wantedDepOutputs) - if (auto outPath = get(inputDrvOutputs, { depDrvPath, j })) + for (auto & j : wantedDepOutputs) { + /* TODO (impure derivations-induced tech debt): + Tracking input derivation outputs statefully through the + goals is error prone and has led to bugs. + For a robust nix, we need to move towards the `else` branch, + which does not rely on goal state to match up with the + reality of the store, which is our real source of truth. + However, the impure derivations feature still relies on this + fragile way of doing things, because its builds do not have + a representation in the store, which is a usability problem + in itself. When implementing this logic entirely with lookups + make sure that they're cached. */ + if (auto outPath = get(inputDrvOutputs, { depDrvPath, j })) { worker.store.computeFSClosure(*outPath, inputPaths); - else - throw Error( - "derivation '%s' requires non-existent output '%s' from input derivation '%s'", - worker.store.printStorePath(drvPath), j, worker.store.printStorePath(depDrvPath)); + } + else { + auto outMap = worker.evalStore.queryDerivationOutputMap(depDrvPath); + auto outMapPath = outMap.find(j); + if (outMapPath == outMap.end()) { + throw Error( + "derivation '%s' requires non-existent output '%s' from input derivation '%s'", + worker.store.printStorePath(drvPath), j, worker.store.printStorePath(depDrvPath)); + } + worker.store.computeFSClosure(outMapPath->second, inputPaths); + } + } } } @@ -546,10 +565,6 @@ void DerivationGoal::inputsRealised() /* What type of derivation are we building? */ derivationType = drv->type(); - /* Don't repeat fixed-output derivations since they're already - verified by their output hash.*/ - nrRounds = derivationType.isFixed() ? 1 : settings.buildRepeat + 1; - /* Okay, try to build. Note that here we don't wait for a build slot to become available, since we don't need one if there is a build hook. */ @@ -564,12 +579,11 @@ void DerivationGoal::started() auto msg = fmt( buildMode == bmRepair ? "repairing outputs of '%s'" : buildMode == bmCheck ? "checking outputs of '%s'" : - nrRounds > 1 ? "building '%s' (round %d/%d)" : - "building '%s'", worker.store.printStorePath(drvPath), curRound, nrRounds); + "building '%s'", worker.store.printStorePath(drvPath)); fmt("building '%s'", worker.store.printStorePath(drvPath)); if (hook) msg += fmt(" on '%s'", machineName); act = std::make_unique<Activity>(*logger, lvlInfo, actBuild, msg, - Logger::Fields{worker.store.printStorePath(drvPath), hook ? machineName : "", curRound, nrRounds}); + Logger::Fields{worker.store.printStorePath(drvPath), hook ? machineName : "", 1, 1}); mcRunningBuilds = std::make_unique<MaintainCount<uint64_t>>(worker.runningBuilds); worker.updateProgress(); } @@ -705,8 +719,7 @@ static void movePath(const Path & src, const Path & dst) if (changePerm) chmod_(src, st.st_mode | S_IWUSR); - if (rename(src.c_str(), dst.c_str())) - throw SysError("renaming '%1%' to '%2%'", src, dst); + renameFile(src, dst); if (changePerm) chmod_(dst, st.st_mode); @@ -719,7 +732,7 @@ void replaceValidPath(const Path & storePath, const Path & tmpPath) tmpPath (the replacement), so we have to move it out of the way first. We'd better not be interrupted here, because if we're repairing (say) Glibc, we end up with a broken system. */ - Path oldPath = (format("%1%.old-%2%-%3%") % storePath % getpid() % random()).str(); + Path oldPath = fmt("%1%.old-%2%-%3%", storePath, getpid(), random()); if (pathExists(storePath)) movePath(storePath, oldPath); @@ -870,6 +883,14 @@ void DerivationGoal::buildDone() cleanupPostChildKill(); + if (buildResult.cpuUser && buildResult.cpuSystem) { + debug("builder for '%s' terminated with status %d, user CPU %.3fs, system CPU %.3fs", + worker.store.printStorePath(drvPath), + status, + ((double) buildResult.cpuUser->count()) / 1000000, + ((double) buildResult.cpuSystem->count()) / 1000000); + } + bool diskFull = false; try { @@ -914,22 +935,8 @@ void DerivationGoal::buildDone() outputPaths ); - if (buildMode == bmCheck) { - cleanupPostOutputsRegisteredModeCheck(); - done(BuildResult::Built, std::move(builtOutputs)); - return; - } - cleanupPostOutputsRegisteredModeNonCheck(); - /* Repeat the build if necessary. */ - if (curRound++ < nrRounds) { - outputLocks.unlock(); - state = &DerivationGoal::tryToBuild; - worker.wakeUp(shared_from_this()); - return; - } - /* It is now safe to delete the lock files, since all future lockers will see that the output paths are valid; they will not create new lock files with the same names as the old @@ -958,7 +965,7 @@ void DerivationGoal::buildDone() BuildResult::PermanentFailure; } - done(st, {}, e); + done(st, {}, std::move(e)); return; } } @@ -978,10 +985,15 @@ void DerivationGoal::resolvedFinished() StorePathSet outputPaths; - // `wantedOutputs` might be empty, which means “all the outputs” - auto realWantedOutputs = wantedOutputs; - if (realWantedOutputs.empty()) - realWantedOutputs = resolvedDrv.outputNames(); + // `wantedOutputs` might merely indicate “all the outputs” + auto realWantedOutputs = std::visit(overloaded { + [&](const OutputsSpec::All &) { + return resolvedDrv.outputNames(); + }, + [&](const OutputsSpec::Names & names) { + return static_cast<std::set<std::string>>(names); + }, + }, wantedOutputs.raw()); for (auto & wantedOutput : realWantedOutputs) { auto initialOutput = get(initialOutputs, wantedOutput); @@ -990,22 +1002,34 @@ void DerivationGoal::resolvedFinished() throw Error( "derivation '%s' doesn't have expected output '%s' (derivation-goal.cc/resolvedFinished,resolve)", worker.store.printStorePath(drvPath), wantedOutput); - auto realisation = get(resolvedResult.builtOutputs, DrvOutput { *resolvedHash, wantedOutput }); - if (!realisation) - throw Error( - "derivation '%s' doesn't have expected output '%s' (derivation-goal.cc/resolvedFinished,realisation)", - worker.store.printStorePath(resolvedDrvGoal->drvPath), wantedOutput); + + auto realisation = [&]{ + auto take1 = get(resolvedResult.builtOutputs, DrvOutput { *resolvedHash, wantedOutput }); + if (take1) return *take1; + + /* The above `get` should work. But sateful tracking of + outputs in resolvedResult, this can get out of sync with the + store, which is our actual source of truth. For now we just + check the store directly if it fails. */ + auto take2 = worker.evalStore.queryRealisation(DrvOutput { *resolvedHash, wantedOutput }); + if (take2) return *take2; + + throw Error( + "derivation '%s' doesn't have expected output '%s' (derivation-goal.cc/resolvedFinished,realisation)", + worker.store.printStorePath(resolvedDrvGoal->drvPath), wantedOutput); + }(); + if (drv->type().isPure()) { - auto newRealisation = *realisation; + auto newRealisation = realisation; newRealisation.id = DrvOutput { initialOutput->outputHash, wantedOutput }; newRealisation.signatures.clear(); if (!drv->type().isFixed()) - newRealisation.dependentRealisations = drvOutputReferences(worker.store, *drv, realisation->outPath); + newRealisation.dependentRealisations = drvOutputReferences(worker.store, *drv, realisation.outPath); signRealisation(newRealisation); worker.store.registerDrvOutput(newRealisation); } - outputPaths.insert(realisation->outPath); - builtOutputs.emplace(realisation->id, *realisation); + outputPaths.insert(realisation.outPath); + builtOutputs.emplace(realisation.id, realisation); } runPostBuildHook( @@ -1297,7 +1321,14 @@ std::pair<bool, DrvOutputs> DerivationGoal::checkPathValidity() if (!drv->type().isPure()) return { false, {} }; bool checkHash = buildMode == bmRepair; - auto wantedOutputsLeft = wantedOutputs; + auto wantedOutputsLeft = std::visit(overloaded { + [&](const OutputsSpec::All &) { + return StringSet {}; + }, + [&](const OutputsSpec::Names & names) { + return static_cast<StringSet>(names); + }, + }, wantedOutputs.raw()); DrvOutputs validOutputs; for (auto & i : queryPartialDerivationOutputMap()) { @@ -1306,7 +1337,7 @@ std::pair<bool, DrvOutputs> DerivationGoal::checkPathValidity() // this is an invalid output, gets catched with (!wantedOutputsLeft.empty()) continue; auto & info = *initialOutput; - info.wanted = wantOutput(i.first, wantedOutputs); + info.wanted = wantedOutputs.contains(i.first); if (info.wanted) wantedOutputsLeft.erase(i.first); if (i.second) { @@ -1344,7 +1375,7 @@ std::pair<bool, DrvOutputs> DerivationGoal::checkPathValidity() validOutputs.emplace(drvOutput, Realisation { drvOutput, info.known->path }); } - // If we requested all the outputs via the empty set, we are always fine. + // If we requested all the outputs, we are always fine. // If we requested specific elements, the loop above removes all the valid // ones, so any that are left must be invalid. if (!wantedOutputsLeft.empty()) @@ -1409,7 +1440,7 @@ void DerivationGoal::done( fs << worker.store.printStorePath(drvPath) << "\t" << buildResult.toString() << std::endl; } - amDone(buildResult.success() ? ecSuccess : ecFailed, ex); + amDone(buildResult.success() ? ecSuccess : ecFailed, std::move(ex)); } diff --git a/src/libstore/build/derivation-goal.hh b/src/libstore/build/derivation-goal.hh index 2d8bfd592..707e38b4b 100644 --- a/src/libstore/build/derivation-goal.hh +++ b/src/libstore/build/derivation-goal.hh @@ -2,6 +2,7 @@ #include "parsed-derivations.hh" #include "lock.hh" +#include "outputs-spec.hh" #include "store-api.hh" #include "pathlocks.hh" #include "goal.hh" @@ -55,7 +56,7 @@ struct DerivationGoal : public Goal /* The specific outputs that we need to build. Empty means all of them. */ - StringSet wantedOutputs; + OutputsSpec wantedOutputs; /* Mapping from input derivations + output names to actual store paths. This is filled in by waiteeDone() as each dependency @@ -115,11 +116,6 @@ struct DerivationGoal : public Goal BuildMode buildMode; - /* The current round, if we're building multiple times. */ - size_t curRound = 1; - - size_t nrRounds; - std::unique_ptr<MaintainCount<uint64_t>> mcExpectedBuilds, mcRunningBuilds; std::unique_ptr<Activity> act; @@ -133,10 +129,10 @@ struct DerivationGoal : public Goal std::string machineName; DerivationGoal(const StorePath & drvPath, - const StringSet & wantedOutputs, Worker & worker, + const OutputsSpec & wantedOutputs, Worker & worker, BuildMode buildMode = bmNormal); DerivationGoal(const StorePath & drvPath, const BasicDerivation & drv, - const StringSet & wantedOutputs, Worker & worker, + const OutputsSpec & wantedOutputs, Worker & worker, BuildMode buildMode = bmNormal); virtual ~DerivationGoal(); @@ -147,7 +143,7 @@ struct DerivationGoal : public Goal void work() override; /* Add wanted outputs to an already existing derivation goal. */ - void addWantedOutputs(const StringSet & outputs); + void addWantedOutputs(const OutputsSpec & outputs); /* The states. */ void getDerivation(); diff --git a/src/libstore/build/drv-output-substitution-goal.cc b/src/libstore/build/drv-output-substitution-goal.cc index b7f7b5ab1..b30957c84 100644 --- a/src/libstore/build/drv-output-substitution-goal.cc +++ b/src/libstore/build/drv-output-substitution-goal.cc @@ -61,20 +61,25 @@ void DrvOutputSubstitutionGoal::tryNext() // FIXME: Make async // outputInfo = sub->queryRealisation(id); - outPipe.create(); - promise = decltype(promise)(); + + /* The callback of the curl download below can outlive `this` (if + some other error occurs), so it must not touch `this`. So put + the shared state in a separate refcounted object. */ + downloadState = std::make_shared<DownloadState>(); + downloadState->outPipe.create(); sub->queryRealisation( - id, { [&](std::future<std::shared_ptr<const Realisation>> res) { + id, + { [downloadState(downloadState)](std::future<std::shared_ptr<const Realisation>> res) { try { - Finally updateStats([this]() { outPipe.writeSide.close(); }); - promise.set_value(res.get()); + Finally updateStats([&]() { downloadState->outPipe.writeSide.close(); }); + downloadState->promise.set_value(res.get()); } catch (...) { - promise.set_exception(std::current_exception()); + downloadState->promise.set_exception(std::current_exception()); } } }); - worker.childStarted(shared_from_this(), {outPipe.readSide.get()}, true, false); + worker.childStarted(shared_from_this(), {downloadState->outPipe.readSide.get()}, true, false); state = &DrvOutputSubstitutionGoal::realisationFetched; } @@ -84,7 +89,7 @@ void DrvOutputSubstitutionGoal::realisationFetched() worker.childTerminated(this); try { - outputInfo = promise.get_future().get(); + outputInfo = downloadState->promise.get_future().get(); } catch (std::exception & e) { printError(e.what()); substituterFailed = true; @@ -155,7 +160,7 @@ void DrvOutputSubstitutionGoal::work() void DrvOutputSubstitutionGoal::handleEOF(int fd) { - if (fd == outPipe.readSide.get()) worker.wakeUp(shared_from_this()); + if (fd == downloadState->outPipe.readSide.get()) worker.wakeUp(shared_from_this()); } diff --git a/src/libstore/build/drv-output-substitution-goal.hh b/src/libstore/build/drv-output-substitution-goal.hh index 948dbda8f..e4b044790 100644 --- a/src/libstore/build/drv-output-substitution-goal.hh +++ b/src/libstore/build/drv-output-substitution-goal.hh @@ -16,7 +16,7 @@ class Worker; // 2. Substitute the corresponding output path // 3. Register the output info class DrvOutputSubstitutionGoal : public Goal { -private: + // The drv output we're trying to substitue DrvOutput id; @@ -30,9 +30,13 @@ private: /* The current substituter. */ std::shared_ptr<Store> sub; - Pipe outPipe; - std::thread thr; - std::promise<std::shared_ptr<const Realisation>> promise; + struct DownloadState + { + Pipe outPipe; + std::promise<std::shared_ptr<const Realisation>> promise; + }; + + std::shared_ptr<DownloadState> downloadState; /* Whether a substituter failed. */ bool substituterFailed = false; diff --git a/src/libstore/build/entry-points.cc b/src/libstore/build/entry-points.cc index bea7363db..2925fe3ca 100644 --- a/src/libstore/build/entry-points.cc +++ b/src/libstore/build/entry-points.cc @@ -30,7 +30,7 @@ void Store::buildPaths(const std::vector<DerivedPath> & reqs, BuildMode buildMod if (ex) logError(i->ex->info()); else - ex = i->ex; + ex = std::move(i->ex); } if (i->exitCode != Goal::ecSuccess) { if (auto i2 = dynamic_cast<DerivationGoal *>(i.get())) failed.insert(i2->drvPath); @@ -40,7 +40,7 @@ void Store::buildPaths(const std::vector<DerivedPath> & reqs, BuildMode buildMod if (failed.size() == 1 && ex) { ex->status = worker.exitStatus(); - throw *ex; + throw std::move(*ex); } else if (!failed.empty()) { if (ex) logError(ex->info()); throw Error(worker.exitStatus(), "build of %s failed", showPaths(failed)); @@ -80,7 +80,7 @@ BuildResult Store::buildDerivation(const StorePath & drvPath, const BasicDerivat BuildMode buildMode) { Worker worker(*this, *this); - auto goal = worker.makeBasicDerivationGoal(drvPath, drv, {}, buildMode); + auto goal = worker.makeBasicDerivationGoal(drvPath, drv, OutputsSpec::All {}, buildMode); try { worker.run(Goals{goal}); @@ -89,7 +89,10 @@ BuildResult Store::buildDerivation(const StorePath & drvPath, const BasicDerivat return BuildResult { .status = BuildResult::MiscFailure, .errorMsg = e.msg(), - .path = DerivedPath::Built { .drvPath = drvPath }, + .path = DerivedPath::Built { + .drvPath = drvPath, + .outputs = OutputsSpec::All { }, + }, }; }; } @@ -109,7 +112,7 @@ void Store::ensurePath(const StorePath & path) if (goal->exitCode != Goal::ecSuccess) { if (goal->ex) { goal->ex->status = worker.exitStatus(); - throw *goal->ex; + throw std::move(*goal->ex); } else throw Error(worker.exitStatus(), "path '%s' does not exist and cannot be created", printStorePath(path)); } @@ -130,7 +133,8 @@ void LocalStore::repairPath(const StorePath & path) auto info = queryPathInfo(path); if (info->deriver && isValidPath(*info->deriver)) { goals.clear(); - goals.insert(worker.makeDerivationGoal(*info->deriver, StringSet(), bmRepair)); + // FIXME: Should just build the specific output we need. + goals.insert(worker.makeDerivationGoal(*info->deriver, OutputsSpec::All { }, bmRepair)); worker.run(goals); } else throw Error(worker.exitStatus(), "cannot repair path '%s'", printStorePath(path)); diff --git a/src/libstore/build/goal.cc b/src/libstore/build/goal.cc index 58e805f55..d59b94797 100644 --- a/src/libstore/build/goal.cc +++ b/src/libstore/build/goal.cc @@ -78,9 +78,9 @@ void Goal::amDone(ExitCode result, std::optional<Error> ex) } -void Goal::trace(const FormatOrString & fs) +void Goal::trace(std::string_view s) { - debug("%1%: %2%", name, fs.s); + debug("%1%: %2%", name, s); } } diff --git a/src/libstore/build/goal.hh b/src/libstore/build/goal.hh index 35121c5d9..776eb86bc 100644 --- a/src/libstore/build/goal.hh +++ b/src/libstore/build/goal.hh @@ -88,7 +88,7 @@ struct Goal : public std::enable_shared_from_this<Goal> abort(); } - void trace(const FormatOrString & fs); + void trace(std::string_view s); std::string getName() { diff --git a/src/libstore/build/hook-instance.cc b/src/libstore/build/hook-instance.cc index 1f19ddccc..075ad554f 100644 --- a/src/libstore/build/hook-instance.cc +++ b/src/libstore/build/hook-instance.cc @@ -16,11 +16,11 @@ HookInstance::HookInstance() buildHookArgs.pop_front(); Strings args; + args.push_back(std::string(baseNameOf(buildHook))); for (auto & arg : buildHookArgs) args.push_back(arg); - args.push_back(std::string(baseNameOf(settings.buildHook.get()))); args.push_back(std::to_string(verbosity)); /* Create a pipe to get the output of the child. */ @@ -35,7 +35,10 @@ HookInstance::HookInstance() /* Fork the hook. */ pid = startProcess([&]() { - commonChildInit(fromHook); + if (dup2(fromHook.writeSide.get(), STDERR_FILENO) == -1) + throw SysError("cannot pipe standard error into log file"); + + commonChildInit(); if (chdir("/") == -1) throw SysError("changing into /"); diff --git a/src/libstore/build/local-derivation-goal.cc b/src/libstore/build/local-derivation-goal.cc index 79a241ae0..3484d2044 100644 --- a/src/libstore/build/local-derivation-goal.cc +++ b/src/libstore/build/local-derivation-goal.cc @@ -8,13 +8,15 @@ #include "finally.hh" #include "util.hh" #include "archive.hh" -#include "json.hh" #include "compression.hh" #include "daemon.hh" #include "worker-protocol.hh" #include "topo-sort.hh" #include "callback.hh" #include "json-utils.hh" +#include "cgroup.hh" +#include "personality.hh" +#include "namespaces.hh" #include <regex> #include <queue> @@ -24,7 +26,6 @@ #include <termios.h> #include <unistd.h> #include <sys/mman.h> -#include <sys/utsname.h> #include <sys/resource.h> #include <sys/socket.h> @@ -37,7 +38,6 @@ #include <sys/ioctl.h> #include <net/if.h> #include <netinet/ip.h> -#include <sys/personality.h> #include <sys/mman.h> #include <sched.h> #include <sys/param.h> @@ -56,6 +56,7 @@ #include <pwd.h> #include <grp.h> +#include <iostream> namespace nix { @@ -129,27 +130,46 @@ void LocalDerivationGoal::killChild() if (pid != -1) { worker.childTerminated(this); - if (buildUser) { - /* If we're using a build user, then there is a tricky - race condition: if we kill the build user before the - child has done its setuid() to the build user uid, then - it won't be killed, and we'll potentially lock up in - pid.wait(). So also send a conventional kill to the - child. */ - ::kill(-pid, SIGKILL); /* ignore the result */ - buildUser->kill(); - pid.wait(); - } else - pid.kill(); + /* If we're using a build user, then there is a tricky race + condition: if we kill the build user before the child has + done its setuid() to the build user uid, then it won't be + killed, and we'll potentially lock up in pid.wait(). So + also send a conventional kill to the child. */ + ::kill(-pid, SIGKILL); /* ignore the result */ + + killSandbox(true); - assert(pid == -1); + pid.wait(); } DerivationGoal::killChild(); } -void LocalDerivationGoal::tryLocalBuild() { +void LocalDerivationGoal::killSandbox(bool getStats) +{ + if (cgroup) { + #if __linux__ + auto stats = destroyCgroup(*cgroup); + if (getStats) { + buildResult.cpuUser = stats.cpuUser; + buildResult.cpuSystem = stats.cpuSystem; + } + #else + abort(); + #endif + } + + else if (buildUser) { + auto uid = buildUser->getUID(); + assert(uid != 0); + killUser(uid); + } +} + + +void LocalDerivationGoal::tryLocalBuild() +{ unsigned int curBuilds = worker.getNrLocalBuilds(); if (curBuilds >= settings.maxBuildJobs) { state = &DerivationGoal::tryToBuild; @@ -158,28 +178,57 @@ void LocalDerivationGoal::tryLocalBuild() { return; } - /* If `build-users-group' is not empty, then we have to build as - one of the members of that group. */ - if (settings.buildUsersGroup != "" && getuid() == 0) { -#if defined(__linux__) || defined(__APPLE__) - if (!buildUser) buildUser = std::make_unique<UserLock>(); + /* Are we doing a chroot build? */ + { + auto noChroot = parsedDrv->getBoolAttr("__noChroot"); + if (settings.sandboxMode == smEnabled) { + if (noChroot) + throw Error("derivation '%s' has '__noChroot' set, " + "but that's not allowed when 'sandbox' is 'true'", worker.store.printStorePath(drvPath)); +#if __APPLE__ + if (additionalSandboxProfile != "") + throw Error("derivation '%s' specifies a sandbox profile, " + "but this is only allowed when 'sandbox' is 'relaxed'", worker.store.printStorePath(drvPath)); +#endif + useChroot = true; + } + else if (settings.sandboxMode == smDisabled) + useChroot = false; + else if (settings.sandboxMode == smRelaxed) + useChroot = derivationType.isSandboxed() && !noChroot; + } + + auto & localStore = getLocalStore(); + if (localStore.storeDir != localStore.realStoreDir.get()) { + #if __linux__ + useChroot = true; + #else + throw Error("building using a diverted store is not supported on this platform"); + #endif + } - if (buildUser->findFreeUser()) { - /* Make sure that no other processes are executing under this - uid. */ - buildUser->kill(); - } else { + #if __linux__ + if (useChroot) { + if (!mountAndPidNamespacesSupported()) { + if (!settings.sandboxFallback) + throw Error("this system does not support the kernel namespaces that are required for sandboxing; use '--no-sandbox' to disable sandboxing"); + debug("auto-disabling sandboxing because the prerequisite namespaces are not available"); + useChroot = false; + } + } + #endif + + if (useBuildUsers()) { + if (!buildUser) + buildUser = acquireUserLock(parsedDrv->useUidRange() ? 65536 : 1, useChroot); + + if (!buildUser) { if (!actLock) actLock = std::make_unique<Activity>(*logger, lvlWarn, actBuildWaiting, fmt("waiting for UID to build '%s'", yellowtxt(worker.store.printStorePath(drvPath)))); worker.waitForAWhile(shared_from_this()); return; } -#else - /* Don't know how to block the creation of setuid/setgid - binaries on this platform. */ - throw Error("build users are not supported on this platform for security reasons"); -#endif } actLock.reset(); @@ -193,7 +242,7 @@ void LocalDerivationGoal::tryLocalBuild() { outputLocks.unlock(); buildUser.reset(); worker.permanentFailure = true; - done(BuildResult::InputRejected, {}, e); + done(BuildResult::InputRejected, {}, std::move(e)); return; } @@ -223,8 +272,7 @@ static void movePath(const Path & src, const Path & dst) if (changePerm) chmod_(src, st.st_mode | S_IWUSR); - if (rename(src.c_str(), dst.c_str())) - throw SysError("renaming '%1%' to '%2%'", src, dst); + renameFile(src, dst); if (changePerm) chmod_(dst, st.st_mode); @@ -244,7 +292,7 @@ void LocalDerivationGoal::closeReadPipes() if (hook) { DerivationGoal::closeReadPipes(); } else - builderOut.readSide = -1; + builderOut.close(); } @@ -271,7 +319,7 @@ void LocalDerivationGoal::cleanupPostChildKill() malicious user from leaving behind a process that keeps files open and modifies them after they have been chown'ed to root. */ - if (buildUser) buildUser->kill(); + killSandbox(true); /* Terminate the recursive Nix daemon. */ stopDaemon(); @@ -311,7 +359,7 @@ bool LocalDerivationGoal::cleanupDecideWhetherDiskFull() if (buildMode != bmCheck && status.known->isValid()) continue; auto p = worker.store.printStorePath(status.known->path); if (pathExists(chrootRootDir + p)) - rename((chrootRootDir + p).c_str(), p.c_str()); + renameFile((chrootRootDir + p), p); } return diskFull; @@ -337,12 +385,6 @@ void LocalDerivationGoal::cleanupPostOutputsRegisteredModeNonCheck() } -int childEntry(void * arg) -{ - ((LocalDerivationGoal *) arg)->runChild(); - return 1; -} - #if __linux__ static void linkOrCopy(const Path & from, const Path & to) { @@ -364,6 +406,64 @@ static void linkOrCopy(const Path & from, const Path & to) void LocalDerivationGoal::startBuilder() { + if ((buildUser && buildUser->getUIDCount() != 1) + #if __linux__ + || settings.useCgroups + #endif + ) + { + #if __linux__ + settings.requireExperimentalFeature(Xp::Cgroups); + + auto cgroupFS = getCgroupFS(); + if (!cgroupFS) + throw Error("cannot determine the cgroups file system"); + + auto ourCgroups = getCgroups("/proc/self/cgroup"); + auto ourCgroup = ourCgroups[""]; + if (ourCgroup == "") + throw Error("cannot determine cgroup name from /proc/self/cgroup"); + + auto ourCgroupPath = canonPath(*cgroupFS + "/" + ourCgroup); + + if (!pathExists(ourCgroupPath)) + throw Error("expected cgroup directory '%s'", ourCgroupPath); + + static std::atomic<unsigned int> counter{0}; + + cgroup = buildUser + ? fmt("%s/nix-build-uid-%d", ourCgroupPath, buildUser->getUID()) + : fmt("%s/nix-build-pid-%d-%d", ourCgroupPath, getpid(), counter++); + + debug("using cgroup '%s'", *cgroup); + + /* When using a build user, record the cgroup we used for that + user so that if we got interrupted previously, we can kill + any left-over cgroup first. */ + if (buildUser) { + auto cgroupsDir = settings.nixStateDir + "/cgroups"; + createDirs(cgroupsDir); + + auto cgroupFile = fmt("%s/%d", cgroupsDir, buildUser->getUID()); + + if (pathExists(cgroupFile)) { + auto prevCgroup = readFile(cgroupFile); + destroyCgroup(prevCgroup); + } + + writeFile(cgroupFile, *cgroup); + } + + #else + throw Error("cgroups are not supported on this platform"); + #endif + } + + /* Make sure that no other processes are executing under the + sandbox uids. This must be done before any chownToBuilder() + calls. */ + killSandbox(false); + /* Right platform? */ if (!parsedDrv->canBuildLocally(worker.store)) throw Error("a '%s' with features {%s} is required to build '%s', but I am a '%s' with features {%s}", @@ -377,35 +477,6 @@ void LocalDerivationGoal::startBuilder() additionalSandboxProfile = parsedDrv->getStringAttr("__sandboxProfile").value_or(""); #endif - /* Are we doing a chroot build? */ - { - auto noChroot = parsedDrv->getBoolAttr("__noChroot"); - if (settings.sandboxMode == smEnabled) { - if (noChroot) - throw Error("derivation '%s' has '__noChroot' set, " - "but that's not allowed when 'sandbox' is 'true'", worker.store.printStorePath(drvPath)); -#if __APPLE__ - if (additionalSandboxProfile != "") - throw Error("derivation '%s' specifies a sandbox profile, " - "but this is only allowed when 'sandbox' is 'relaxed'", worker.store.printStorePath(drvPath)); -#endif - useChroot = true; - } - else if (settings.sandboxMode == smDisabled) - useChroot = false; - else if (settings.sandboxMode == smRelaxed) - useChroot = derivationType.isSandboxed() && !noChroot; - } - - auto & localStore = getLocalStore(); - if (localStore.storeDir != localStore.realStoreDir.get()) { - #if __linux__ - useChroot = true; - #else - throw Error("building using a diverted store is not supported on this platform"); - #endif - } - /* Create a temporary directory where the build will take place. */ tmpDir = createTempDir("", "nix-build-" + std::string(drvPath.name()), false, false, 0700); @@ -579,12 +650,13 @@ void LocalDerivationGoal::startBuilder() /* Clean up the chroot directory automatically. */ autoDelChroot = std::make_shared<AutoDelete>(chrootRootDir); - printMsg(lvlChatty, format("setting up chroot environment in '%1%'") % chrootRootDir); + printMsg(lvlChatty, "setting up chroot environment in '%1%'", chrootRootDir); - if (mkdir(chrootRootDir.c_str(), 0750) == -1) + // FIXME: make this 0700 + if (mkdir(chrootRootDir.c_str(), buildUser && buildUser->getUIDCount() != 1 ? 0755 : 0750) == -1) throw SysError("cannot create '%1%'", chrootRootDir); - if (buildUser && chown(chrootRootDir.c_str(), 0, buildUser->getGID()) == -1) + if (buildUser && chown(chrootRootDir.c_str(), buildUser->getUIDCount() != 1 ? buildUser->getUID() : 0, buildUser->getGID()) == -1) throw SysError("cannot change ownership of '%1%'", chrootRootDir); /* Create a writable /tmp in the chroot. Many builders need @@ -598,6 +670,11 @@ void LocalDerivationGoal::startBuilder() nobody account. The latter is kind of a hack to support Samba-in-QEMU. */ createDirs(chrootRootDir + "/etc"); + if (parsedDrv->useUidRange()) + chownToBuilder(chrootRootDir + "/etc"); + + if (parsedDrv->useUidRange() && (!buildUser || buildUser->getUIDCount() < 65536)) + throw Error("feature 'uid-range' requires the setting '%s' to be enabled", settings.autoAllocateUids.name); /* Declare the build user's group so that programs get a consistent view of the system (e.g., "id -gn"). */ @@ -648,20 +725,35 @@ void LocalDerivationGoal::startBuilder() dirsInChroot.erase(worker.store.printStorePath(*i.second.second)); } -#elif __APPLE__ - /* We don't really have any parent prep work to do (yet?) - All work happens in the child, instead. */ + if (cgroup) { + if (mkdir(cgroup->c_str(), 0755) != 0) + throw SysError("creating cgroup '%s'", *cgroup); + chownToBuilder(*cgroup); + chownToBuilder(*cgroup + "/cgroup.procs"); + chownToBuilder(*cgroup + "/cgroup.threads"); + //chownToBuilder(*cgroup + "/cgroup.subtree_control"); + } + #else - throw Error("sandboxing builds is not supported on this platform"); + if (parsedDrv->useUidRange()) + throw Error("feature 'uid-range' is not supported on this platform"); + #if __APPLE__ + /* We don't really have any parent prep work to do (yet?) + All work happens in the child, instead. */ + #else + throw Error("sandboxing builds is not supported on this platform"); + #endif #endif + } else { + if (parsedDrv->useUidRange()) + throw Error("feature 'uid-range' is only supported in sandboxed builds"); } if (needsHashRewrite() && pathExists(homeDir)) throw Error("home directory '%1%' exists; please remove it to assure purity of builds without sandboxing", homeDir); if (useChroot && settings.preBuildHook != "" && dynamic_cast<Derivation *>(drv.get())) { - printMsg(lvlChatty, format("executing pre-build hook '%1%'") - % settings.preBuildHook); + printMsg(lvlChatty, "executing pre-build hook '%1%'", settings.preBuildHook); auto args = useChroot ? Strings({worker.store.printStorePath(drvPath), chrootRootDir}) : Strings({ worker.store.printStorePath(drvPath) }); enum BuildHookState { @@ -710,15 +802,13 @@ void LocalDerivationGoal::startBuilder() /* Create the log file. */ Path logFile = openLogFile(); - /* Create a pipe to get the output of the builder. */ - //builderOut.create(); - - builderOut.readSide = posix_openpt(O_RDWR | O_NOCTTY); - if (!builderOut.readSide) + /* Create a pseudoterminal to get the output of the builder. */ + builderOut = posix_openpt(O_RDWR | O_NOCTTY); + if (!builderOut) throw SysError("opening pseudoterminal master"); // FIXME: not thread-safe, use ptsname_r - std::string slaveName(ptsname(builderOut.readSide.get())); + std::string slaveName = ptsname(builderOut.get()); if (buildUser) { if (chmod(slaveName.c_str(), 0600)) @@ -729,34 +819,34 @@ void LocalDerivationGoal::startBuilder() } #if __APPLE__ else { - if (grantpt(builderOut.readSide.get())) + if (grantpt(builderOut.get())) throw SysError("granting access to pseudoterminal slave"); } #endif - #if 0 - // Mount the pt in the sandbox so that the "tty" command works. - // FIXME: this doesn't work with the new devpts in the sandbox. - if (useChroot) - dirsInChroot[slaveName] = {slaveName, false}; - #endif - - if (unlockpt(builderOut.readSide.get())) + if (unlockpt(builderOut.get())) throw SysError("unlocking pseudoterminal"); - builderOut.writeSide = open(slaveName.c_str(), O_RDWR | O_NOCTTY); - if (!builderOut.writeSide) - throw SysError("opening pseudoterminal slave"); + /* Open the slave side of the pseudoterminal and use it as stderr. */ + auto openSlave = [&]() + { + AutoCloseFD builderOut = open(slaveName.c_str(), O_RDWR | O_NOCTTY); + if (!builderOut) + throw SysError("opening pseudoterminal slave"); - // Put the pt into raw mode to prevent \n -> \r\n translation. - struct termios term; - if (tcgetattr(builderOut.writeSide.get(), &term)) - throw SysError("getting pseudoterminal attributes"); + // Put the pt into raw mode to prevent \n -> \r\n translation. + struct termios term; + if (tcgetattr(builderOut.get(), &term)) + throw SysError("getting pseudoterminal attributes"); - cfmakeraw(&term); + cfmakeraw(&term); - if (tcsetattr(builderOut.writeSide.get(), TCSANOW, &term)) - throw SysError("putting pseudoterminal into raw mode"); + if (tcsetattr(builderOut.get(), TCSANOW, &term)) + throw SysError("putting pseudoterminal into raw mode"); + + if (dup2(builderOut.get(), STDERR_FILENO) == -1) + throw SysError("cannot pipe standard error into log file"); + }; buildResult.startTime = time(0); @@ -803,14 +893,18 @@ void LocalDerivationGoal::startBuilder() userNamespaceSync.create(); - Path maxUserNamespaces = "/proc/sys/user/max_user_namespaces"; - static bool userNamespacesEnabled = - pathExists(maxUserNamespaces) - && trim(readFile(maxUserNamespaces)) != "0"; + usingUserNamespace = userNamespacesSupported(); - usingUserNamespace = userNamespacesEnabled; + Pipe sendPid; + sendPid.create(); Pid helper = startProcess([&]() { + sendPid.readSide.close(); + + /* We need to open the slave early, before + CLONE_NEWUSER. Otherwise we get EPERM when running as + root. */ + openSlave(); /* Drop additional groups here because we can't do it after we've created the new user namespace. FIXME: @@ -823,76 +917,22 @@ void LocalDerivationGoal::startBuilder() if (getuid() == 0 && setgroups(0, 0) == -1) throw SysError("setgroups failed"); - size_t stackSize = 1 * 1024 * 1024; - char * stack = (char *) mmap(0, stackSize, - PROT_WRITE | PROT_READ, MAP_PRIVATE | MAP_ANONYMOUS | MAP_STACK, -1, 0); - if (stack == MAP_FAILED) throw SysError("allocating stack"); - - int flags = CLONE_NEWPID | CLONE_NEWNS | CLONE_NEWIPC | CLONE_NEWUTS | CLONE_PARENT | SIGCHLD; + ProcessOptions options; + options.cloneFlags = CLONE_NEWPID | CLONE_NEWNS | CLONE_NEWIPC | CLONE_NEWUTS | CLONE_PARENT | SIGCHLD; if (privateNetwork) - flags |= CLONE_NEWNET; + options.cloneFlags |= CLONE_NEWNET; if (usingUserNamespace) - flags |= CLONE_NEWUSER; - - pid_t child = clone(childEntry, stack + stackSize, flags, this); - if (child == -1 && errno == EINVAL) { - /* Fallback for Linux < 2.13 where CLONE_NEWPID and - CLONE_PARENT are not allowed together. */ - flags &= ~CLONE_NEWPID; - child = clone(childEntry, stack + stackSize, flags, this); - } - if (usingUserNamespace && child == -1 && (errno == EPERM || errno == EINVAL)) { - /* Some distros patch Linux to not allow unprivileged - * user namespaces. If we get EPERM or EINVAL, try - * without CLONE_NEWUSER and see if that works. - * Details: https://salsa.debian.org/kernel-team/linux/-/commit/d98e00eda6bea437e39b9e80444eee84a32438a6 - */ - usingUserNamespace = false; - flags &= ~CLONE_NEWUSER; - child = clone(childEntry, stack + stackSize, flags, this); - } - if (child == -1) { - switch(errno) { - case EPERM: - case EINVAL: { - int errno_ = errno; - if (!userNamespacesEnabled && errno==EPERM) - notice("user namespaces appear to be disabled; they are required for sandboxing; check /proc/sys/user/max_user_namespaces"); - if (userNamespacesEnabled) { - Path procSysKernelUnprivilegedUsernsClone = "/proc/sys/kernel/unprivileged_userns_clone"; - if (pathExists(procSysKernelUnprivilegedUsernsClone) - && trim(readFile(procSysKernelUnprivilegedUsernsClone)) == "0") { - notice("user namespaces appear to be disabled; they are required for sandboxing; check /proc/sys/kernel/unprivileged_userns_clone"); - } - } - Path procSelfNsUser = "/proc/self/ns/user"; - if (!pathExists(procSelfNsUser)) - notice("/proc/self/ns/user does not exist; your kernel was likely built without CONFIG_USER_NS=y, which is required for sandboxing"); - /* Otherwise exit with EPERM so we can handle this in the - parent. This is only done when sandbox-fallback is set - to true (the default). */ - if (settings.sandboxFallback) - _exit(1); - /* Mention sandbox-fallback in the error message so the user - knows that having it disabled contributed to the - unrecoverability of this failure */ - throw SysError(errno_, "creating sandboxed builder process using clone(), without sandbox-fallback"); - } - default: - throw SysError("creating sandboxed builder process using clone()"); - } - } - writeFull(builderOut.writeSide.get(), - fmt("%d %d\n", usingUserNamespace, child)); + options.cloneFlags |= CLONE_NEWUSER; + + pid_t child = startProcess([&]() { runChild(); }, options); + + writeFull(sendPid.writeSide.get(), fmt("%d\n", child)); _exit(0); }); - int res = helper.wait(); - if (res != 0 && settings.sandboxFallback) { - useChroot = false; - initTmpDir(); - goto fallback; - } else if (res != 0) + sendPid.writeSide.close(); + + if (helper.wait() != 0) throw Error("unable to start build process"); userNamespaceSync.readSide = -1; @@ -903,10 +943,9 @@ void LocalDerivationGoal::startBuilder() userNamespaceSync.writeSide = -1; }); - auto ss = tokenizeString<std::vector<std::string>>(readLine(builderOut.readSide.get())); - assert(ss.size() == 2); - usingUserNamespace = ss[0] == "1"; - pid = string2Int<pid_t>(ss[1]).value(); + auto ss = tokenizeString<std::vector<std::string>>(readLine(sendPid.readSide.get())); + assert(ss.size() == 1); + pid = string2Int<pid_t>(ss[0]).value(); if (usingUserNamespace) { /* Set the UID/GID mapping of the builder's user namespace @@ -914,14 +953,16 @@ void LocalDerivationGoal::startBuilder() the calling user (if build users are disabled). */ uid_t hostUid = buildUser ? buildUser->getUID() : getuid(); uid_t hostGid = buildUser ? buildUser->getGID() : getgid(); + uid_t nrIds = buildUser ? buildUser->getUIDCount() : 1; writeFile("/proc/" + std::to_string(pid) + "/uid_map", - fmt("%d %d 1", sandboxUid(), hostUid)); + fmt("%d %d %d", sandboxUid(), hostUid, nrIds)); - writeFile("/proc/" + std::to_string(pid) + "/setgroups", "deny"); + if (!buildUser || buildUser->getUIDCount() == 1) + writeFile("/proc/" + std::to_string(pid) + "/setgroups", "deny"); writeFile("/proc/" + std::to_string(pid) + "/gid_map", - fmt("%d %d 1", sandboxGid(), hostGid)); + fmt("%d %d %d", sandboxGid(), hostGid, nrIds)); } else { debug("note: not using a user namespace"); if (!buildUser) @@ -948,31 +989,32 @@ void LocalDerivationGoal::startBuilder() throw SysError("getting sandbox user namespace"); } + /* Move the child into its own cgroup. */ + if (cgroup) + writeFile(*cgroup + "/cgroup.procs", fmt("%d", (pid_t) pid)); + /* Signal the builder that we've updated its user namespace. */ writeFull(userNamespaceSync.writeSide.get(), "1"); } else #endif { -#if __linux__ - fallback: -#endif pid = startProcess([&]() { + openSlave(); runChild(); }); } /* parent */ pid.setSeparatePG(true); - builderOut.writeSide = -1; - worker.childStarted(shared_from_this(), {builderOut.readSide.get()}, true, true); + worker.childStarted(shared_from_this(), {builderOut.get()}, true, true); /* Check if setting up the build environment failed. */ std::vector<std::string> msgs; while (true) { std::string msg = [&]() { try { - return readLine(builderOut.readSide.get()); + return readLine(builderOut.get()); } catch (Error & e) { auto status = pid.wait(); e.addTrace({}, "while waiting for the build environment for '%s' to initialize (%s, previous messages: %s)", @@ -984,7 +1026,7 @@ void LocalDerivationGoal::startBuilder() }(); if (msg.substr(0, 1) == "\2") break; if (msg.substr(0, 1) == "\1") { - FdSource source(builderOut.readSide.get()); + FdSource source(builderOut.get()); auto ex = readError(source); ex.addTrace({}, "while setting up the build environment"); throw ex; @@ -1068,7 +1110,7 @@ void LocalDerivationGoal::initEnv() env["NIX_STORE"] = worker.store.storeDir; /* The maximum number of cores to utilize for parallel building. */ - env["NIX_BUILD_CORES"] = (format("%d") % settings.buildCores).str(); + env["NIX_BUILD_CORES"] = fmt("%d", settings.buildCores); initTmpDir(); @@ -1119,10 +1161,10 @@ void LocalDerivationGoal::writeStructuredAttrs() writeFile(tmpDir + "/.attrs.sh", rewriteStrings(jsonSh, inputRewrites)); chownToBuilder(tmpDir + "/.attrs.sh"); - env["NIX_ATTRS_SH_FILE"] = tmpDir + "/.attrs.sh"; + env["NIX_ATTRS_SH_FILE"] = tmpDirInSandbox + "/.attrs.sh"; writeFile(tmpDir + "/.attrs.json", rewriteStrings(json.dump(), inputRewrites)); chownToBuilder(tmpDir + "/.attrs.json"); - env["NIX_ATTRS_JSON_FILE"] = tmpDir + "/.attrs.json"; + env["NIX_ATTRS_JSON_FILE"] = tmpDirInSandbox + "/.attrs.json"; } } @@ -1368,7 +1410,7 @@ struct RestrictedStore : public virtual RestrictedStoreConfig, public virtual Lo unknown, downloadSize, narSize); } - virtual std::optional<std::string> getBuildLog(const StorePath & path) override + virtual std::optional<std::string> getBuildLogExact(const StorePath & path) override { return std::nullopt; } virtual void addBuildLog(const StorePath & path, std::string_view log) override @@ -1425,8 +1467,7 @@ void LocalDerivationGoal::startDaemon() FdSink to(remote.get()); try { daemon::processConnection(store, from, to, - daemon::NotTrusted, daemon::Recursive, - [&](Store & store) { store.createUser("nobody", 65535); }); + daemon::NotTrusted, daemon::Recursive); debug("terminated daemon connection"); } catch (SysError &) { ignoreException(); @@ -1553,6 +1594,22 @@ void setupSeccomp() seccomp_arch_add(ctx, SCMP_ARCH_ARM) != 0) printError("unable to add ARM seccomp architecture; this may result in spurious build failures if running 32-bit ARM processes"); + if (nativeSystem == "mips64-linux" && + seccomp_arch_add(ctx, SCMP_ARCH_MIPS) != 0) + printError("unable to add mips seccomp architecture"); + + if (nativeSystem == "mips64-linux" && + seccomp_arch_add(ctx, SCMP_ARCH_MIPS64N32) != 0) + printError("unable to add mips64-*abin32 seccomp architecture"); + + if (nativeSystem == "mips64el-linux" && + seccomp_arch_add(ctx, SCMP_ARCH_MIPSEL) != 0) + printError("unable to add mipsel seccomp architecture"); + + if (nativeSystem == "mips64el-linux" && + seccomp_arch_add(ctx, SCMP_ARCH_MIPSEL64N32) != 0) + printError("unable to add mips64el-*abin32 seccomp architecture"); + /* Prevent builders from creating setuid/setgid binaries. */ for (int perm : { S_ISUID, S_ISGID }) { if (seccomp_rule_add(ctx, SCMP_ACT_ERRNO(EPERM), SCMP_SYS(chmod), 1, @@ -1595,9 +1652,11 @@ void LocalDerivationGoal::runChild() /* Warning: in the child we should absolutely not make any SQLite calls! */ + bool sendException = true; + try { /* child */ - commonChildInit(builderOut); + commonChildInit(); try { setupSeccomp(); @@ -1762,6 +1821,13 @@ void LocalDerivationGoal::runChild() if (mount("none", (chrootRootDir + "/proc").c_str(), "proc", 0, 0) == -1) throw SysError("mounting /proc"); + /* Mount sysfs on /sys. */ + if (buildUser && buildUser->getUIDCount() != 1) { + createDirs(chrootRootDir + "/sys"); + if (mount("none", (chrootRootDir + "/sys").c_str(), "sysfs", 0, 0) == -1) + throw SysError("mounting /sys"); + } + /* Mount a new tmpfs on /dev/shm to ensure that whatever the builder puts in /dev/shm is cleaned up automatically. */ if (pathExists("/dev/shm") && mount("none", (chrootRootDir + "/dev/shm").c_str(), "tmpfs", 0, @@ -1791,6 +1857,10 @@ void LocalDerivationGoal::runChild() } } + /* Make /etc unwritable */ + if (!parsedDrv->useUidRange()) + chmod_(chrootRootDir + "/etc", 0555); + /* Unshare this mount namespace. This is necessary because pivot_root() below changes the root of the mount namespace. This means that the call to setns() in @@ -1804,6 +1874,12 @@ void LocalDerivationGoal::runChild() if (unshare(CLONE_NEWNS) == -1) throw SysError("unsharing mount namespace"); + /* Unshare the cgroup namespace. This means + /proc/self/cgroup will show the child's cgroup as '/' + rather than whatever it is in the parent. */ + if (cgroup && unshare(CLONE_NEWCGROUP) == -1) + throw SysError("unsharing cgroup namespace"); + /* Do the chroot(). */ if (chdir(chrootRootDir.c_str()) == -1) throw SysError("cannot change directory to '%1%'", chrootRootDir); @@ -1841,33 +1917,7 @@ void LocalDerivationGoal::runChild() /* Close all other file descriptors. */ closeMostFDs({STDIN_FILENO, STDOUT_FILENO, STDERR_FILENO}); -#if __linux__ - /* Change the personality to 32-bit if we're doing an - i686-linux build on an x86_64-linux machine. */ - struct utsname utsbuf; - uname(&utsbuf); - if ((drv->platform == "i686-linux" - && (settings.thisSystem == "x86_64-linux" - || (!strcmp(utsbuf.sysname, "Linux") && !strcmp(utsbuf.machine, "x86_64")))) - || drv->platform == "armv7l-linux" - || drv->platform == "armv6l-linux") - { - if (personality(PER_LINUX32) == -1) - throw SysError("cannot set 32-bit personality"); - } - - /* Impersonate a Linux 2.6 machine to get some determinism in - builds that depend on the kernel version. */ - if ((drv->platform == "i686-linux" || drv->platform == "x86_64-linux") && settings.impersonateLinux26) { - int cur = personality(0xffffffff); - if (cur != -1) personality(cur | 0x0020000 /* == UNAME26 */); - } - - /* Disable address space randomization for improved - determinism. */ - int cur = personality(0xffffffff); - if (cur != -1) personality(cur | ADDR_NO_RANDOMIZE); -#endif + setPersonality(drv->platform); /* Disable core dumps by default. */ struct rlimit limit = { 0, RLIM_INFINITY }; @@ -1889,9 +1939,8 @@ void LocalDerivationGoal::runChild() if (setUser && buildUser) { /* Preserve supplementary groups of the build user, to allow admins to specify groups such as "kvm". */ - if (!buildUser->getSupplementaryGIDs().empty() && - setgroups(buildUser->getSupplementaryGIDs().size(), - buildUser->getSupplementaryGIDs().data()) == -1) + auto gids = buildUser->getSupplementaryGIDs(); + if (setgroups(gids.size(), gids.data()) == -1) throw SysError("cannot set supplementary groups of build user"); if (setgid(buildUser->getGID()) == -1 || @@ -1955,10 +2004,14 @@ void LocalDerivationGoal::runChild() sandboxProfile += "(deny default (with no-log))\n"; } - sandboxProfile += "(import \"sandbox-defaults.sb\")\n"; + sandboxProfile += + #include "sandbox-defaults.sb" + ; if (!derivationType.isSandboxed()) - sandboxProfile += "(import \"sandbox-network.sb\")\n"; + sandboxProfile += + #include "sandbox-network.sb" + ; /* Add the output paths we'll use at build-time to the chroot */ sandboxProfile += "(allow file-read* file-write* process-exec\n"; @@ -2001,7 +2054,9 @@ void LocalDerivationGoal::runChild() sandboxProfile += additionalSandboxProfile; } else - sandboxProfile += "(import \"sandbox-minimal.sb\")\n"; + sandboxProfile += + #include "sandbox-minimal.sb" + ; debug("Generated sandbox profile:"); debug(sandboxProfile); @@ -2014,7 +2069,7 @@ void LocalDerivationGoal::runChild() /* The tmpDir in scope points at the temporary build directory for our derivation. Some packages try different mechanisms to find temporary directories, so we want to open up a broader place for them to dump their files, if needed. */ - Path globalTmpDir = canonPath(getEnv("TMPDIR").value_or("/tmp"), true); + Path globalTmpDir = canonPath(getEnvNonEmpty("TMPDIR").value_or("/tmp"), true); /* They don't like trailing slashes on subpath directives */ if (globalTmpDir.back() == '/') globalTmpDir.pop_back(); @@ -2026,8 +2081,6 @@ void LocalDerivationGoal::runChild() args.push_back(sandboxFile); args.push_back("-D"); args.push_back("_GLOBAL_TMP_DIR=" + globalTmpDir); - args.push_back("-D"); - args.push_back("IMPORT_DIR=" + settings.nixDataDir + "/nix/sandbox/"); if (allowLocalNetworking) { args.push_back("-D"); args.push_back(std::string("_ALLOW_LOCAL_NETWORKING=1")); @@ -2051,6 +2104,8 @@ void LocalDerivationGoal::runChild() /* Indicate that we managed to set up the build environment. */ writeFull(STDERR_FILENO, std::string("\2\n")); + sendException = false; + /* Execute the program. This should not return. */ if (drv->isBuiltin()) { try { @@ -2104,10 +2159,13 @@ void LocalDerivationGoal::runChild() throw SysError("executing '%1%'", drv->builder); } catch (Error & e) { - writeFull(STDERR_FILENO, "\1\n"); - FdSink sink(STDERR_FILENO); - sink << e; - sink.flush(); + if (sendException) { + writeFull(STDERR_FILENO, "\1\n"); + FdSink sink(STDERR_FILENO); + sink << e; + sink.flush(); + } else + std::cerr << e.msg(); _exit(1); } } @@ -2133,7 +2191,6 @@ DrvOutputs LocalDerivationGoal::registerOutputs() InodesSeen inodesSeen; Path checkSuffix = ".check"; - bool keepPreviousRound = settings.keepFailed || settings.runDiffHook; std::exception_ptr delayedException; @@ -2215,13 +2272,33 @@ DrvOutputs LocalDerivationGoal::registerOutputs() /* Canonicalise first. This ensures that the path we're rewriting doesn't contain a hard link to /etc/shadow or something like that. */ - canonicalisePathMetaData(actualPath, buildUser ? buildUser->getUID() : -1, inodesSeen); + canonicalisePathMetaData( + actualPath, + buildUser ? std::optional(buildUser->getUIDRange()) : std::nullopt, + inodesSeen); - debug("scanning for references for output '%s' in temp location '%s'", outputName, actualPath); + bool discardReferences = false; + if (auto structuredAttrs = parsedDrv->getStructuredAttrs()) { + if (auto udr = get(*structuredAttrs, "unsafeDiscardReferences")) { + settings.requireExperimentalFeature(Xp::DiscardReferences); + if (auto output = get(*udr, outputName)) { + if (!output->is_boolean()) + throw Error("attribute 'unsafeDiscardReferences.\"%s\"' of derivation '%s' must be a Boolean", outputName, drvPath.to_string()); + discardReferences = output->get<bool>(); + } + } + } - /* Pass blank Sink as we are not ready to hash data at this stage. */ - NullSink blank; - auto references = scanForReferences(blank, actualPath, referenceablePaths); + StorePathSet references; + if (discardReferences) + debug("discarding references of output '%s'", outputName); + else { + debug("scanning for references for output '%s' in temp location '%s'", outputName, actualPath); + + /* Pass blank Sink as we are not ready to hash data at this stage. */ + NullSink blank; + references = scanForReferences(blank, actualPath, referenceablePaths); + } outputReferencesIfUnregistered.insert_or_assign( outputName, @@ -2308,6 +2385,10 @@ DrvOutputs LocalDerivationGoal::registerOutputs() sink.s = rewriteStrings(sink.s, outputRewrites); StringSource source(sink.s); restorePath(actualPath, source); + + /* FIXME: set proper permissions in restorePath() so + we don't have to do another traversal. */ + canonicalisePathMetaData(actualPath, {}, inodesSeen); } }; @@ -2375,10 +2456,8 @@ DrvOutputs LocalDerivationGoal::registerOutputs() if (*scratchPath != finalPath) { // Also rewrite the output path auto source = sinkToSource([&](Sink & nextSink) { - StringSink sink; - dumpPath(actualPath, sink); RewritingSink rsink2(oldHashPart, std::string(finalPath.hashPart()), nextSink); - rsink2(sink.s); + dumpPath(actualPath, rsink2); rsink2.flush(); }); Path tmpPath = actualPath + ".tmp"; @@ -2472,7 +2551,7 @@ DrvOutputs LocalDerivationGoal::registerOutputs() /* FIXME: set proper permissions in restorePath() so we don't have to do another traversal. */ - canonicalisePathMetaData(actualPath, -1, inodesSeen); + canonicalisePathMetaData(actualPath, {}, inodesSeen); /* Calculate where we'll move the output files. In the checking case we will leave leave them where they are, for now, rather than move to @@ -2556,10 +2635,8 @@ DrvOutputs LocalDerivationGoal::registerOutputs() debug("unreferenced input: '%1%'", worker.store.printStorePath(i)); } - if (curRound == nrRounds) { - localStore.optimisePath(actualPath, NoRepair); // FIXME: combine with scanForReferences() - worker.markContentsGood(newInfo.path); - } + localStore.optimisePath(actualPath, NoRepair); // FIXME: combine with scanForReferences() + worker.markContentsGood(newInfo.path); newInfo.deriver = drvPath; newInfo.ultimate = true; @@ -2588,62 +2665,6 @@ DrvOutputs LocalDerivationGoal::registerOutputs() /* Apply output checks. */ checkOutputs(infos); - /* Compare the result with the previous round, and report which - path is different, if any.*/ - if (curRound > 1 && prevInfos != infos) { - assert(prevInfos.size() == infos.size()); - for (auto i = prevInfos.begin(), j = infos.begin(); i != prevInfos.end(); ++i, ++j) - if (!(*i == *j)) { - buildResult.isNonDeterministic = true; - Path prev = worker.store.printStorePath(i->second.path) + checkSuffix; - bool prevExists = keepPreviousRound && pathExists(prev); - hintformat hint = prevExists - ? hintfmt("output '%s' of '%s' differs from '%s' from previous round", - worker.store.printStorePath(i->second.path), worker.store.printStorePath(drvPath), prev) - : hintfmt("output '%s' of '%s' differs from previous round", - worker.store.printStorePath(i->second.path), worker.store.printStorePath(drvPath)); - - handleDiffHook( - buildUser ? buildUser->getUID() : getuid(), - buildUser ? buildUser->getGID() : getgid(), - prev, worker.store.printStorePath(i->second.path), - worker.store.printStorePath(drvPath), tmpDir); - - if (settings.enforceDeterminism) - throw NotDeterministic(hint); - - printError(hint); - - curRound = nrRounds; // we know enough, bail out early - } - } - - /* If this is the first round of several, then move the output out of the way. */ - if (nrRounds > 1 && curRound == 1 && curRound < nrRounds && keepPreviousRound) { - for (auto & [_, outputStorePath] : finalOutputs) { - auto path = worker.store.printStorePath(outputStorePath); - Path prev = path + checkSuffix; - deletePath(prev); - Path dst = path + checkSuffix; - if (rename(path.c_str(), dst.c_str())) - throw SysError("renaming '%s' to '%s'", path, dst); - } - } - - if (curRound < nrRounds) { - prevInfos = std::move(infos); - return {}; - } - - /* Remove the .check directories if we're done. FIXME: keep them - if the result was not determistic? */ - if (curRound == nrRounds) { - for (auto & [_, outputStorePath] : finalOutputs) { - Path prev = worker.store.printStorePath(outputStorePath) + checkSuffix; - deletePath(prev); - } - } - /* Register each output path as valid, and register the sets of paths referenced by each of them. If there are cycles in the outputs, this will fail. */ @@ -2685,7 +2706,7 @@ DrvOutputs LocalDerivationGoal::registerOutputs() signRealisation(thisRealisation); worker.store.registerDrvOutput(thisRealisation); } - if (wantOutput(outputName, wantedOutputs)) + if (wantedOutputs.contains(outputName)) builtOutputs.emplace(thisRealisation.id, thisRealisation); } @@ -2877,7 +2898,7 @@ void LocalDerivationGoal::deleteTmpDir(bool force) bool LocalDerivationGoal::isReadDesc(int fd) { return (hook && DerivationGoal::isReadDesc(fd)) || - (!hook && fd == builderOut.readSide.get()); + (!hook && fd == builderOut.get()); } diff --git a/src/libstore/build/local-derivation-goal.hh b/src/libstore/build/local-derivation-goal.hh index d456e9cae..c9ecc8828 100644 --- a/src/libstore/build/local-derivation-goal.hh +++ b/src/libstore/build/local-derivation-goal.hh @@ -15,14 +15,18 @@ struct LocalDerivationGoal : public DerivationGoal /* The process ID of the builder. */ Pid pid; + /* The cgroup of the builder, if any. */ + std::optional<Path> cgroup; + /* The temporary directory. */ Path tmpDir; /* The path of the temporary directory in the sandbox. */ Path tmpDirInSandbox; - /* Pipe for the builder's standard output/error. */ - Pipe builderOut; + /* Master side of the pseudoterminal used for the builder's + standard output/error. */ + AutoCloseFD builderOut; /* Pipe for synchronising updates to the builder namespaces. */ Pipe userNamespaceSync; @@ -92,8 +96,8 @@ struct LocalDerivationGoal : public DerivationGoal result. */ std::map<Path, ValidPathInfo> prevInfos; - uid_t sandboxUid() { return usingUserNamespace ? 1000 : buildUser->getUID(); } - gid_t sandboxGid() { return usingUserNamespace ? 100 : buildUser->getGID(); } + uid_t sandboxUid() { return usingUserNamespace ? (!buildUser || buildUser->getUIDCount() == 1 ? 1000 : 0) : buildUser->getUID(); } + gid_t sandboxGid() { return usingUserNamespace ? (!buildUser || buildUser->getUIDCount() == 1 ? 100 : 0) : buildUser->getGID(); } const static Path homeDir; @@ -197,6 +201,10 @@ struct LocalDerivationGoal : public DerivationGoal /* Forcibly kill the child process, if any. */ void killChild() override; + /* Kill any processes running under the build user UID or in the + cgroup of the build. */ + void killSandbox(bool getStats); + /* Create alternative path calculated from but distinct from the input, so we can avoid overwriting outputs (or other store paths) that already exist. */ diff --git a/src/libstore/build/personality.cc b/src/libstore/build/personality.cc new file mode 100644 index 000000000..4ad477869 --- /dev/null +++ b/src/libstore/build/personality.cc @@ -0,0 +1,44 @@ +#include "personality.hh" +#include "globals.hh" + +#if __linux__ +#include <sys/utsname.h> +#include <sys/personality.h> +#endif + +#include <cstring> + +namespace nix { + +void setPersonality(std::string_view system) +{ +#if __linux__ + /* Change the personality to 32-bit if we're doing an + i686-linux build on an x86_64-linux machine. */ + struct utsname utsbuf; + uname(&utsbuf); + if ((system == "i686-linux" + && (std::string_view(SYSTEM) == "x86_64-linux" + || (!strcmp(utsbuf.sysname, "Linux") && !strcmp(utsbuf.machine, "x86_64")))) + || system == "armv7l-linux" + || system == "armv6l-linux") + { + if (personality(PER_LINUX32) == -1) + throw SysError("cannot set 32-bit personality"); + } + + /* Impersonate a Linux 2.6 machine to get some determinism in + builds that depend on the kernel version. */ + if ((system == "i686-linux" || system == "x86_64-linux") && settings.impersonateLinux26) { + int cur = personality(0xffffffff); + if (cur != -1) personality(cur | 0x0020000 /* == UNAME26 */); + } + + /* Disable address space randomization for improved + determinism. */ + int cur = personality(0xffffffff); + if (cur != -1) personality(cur | ADDR_NO_RANDOMIZE); +#endif +} + +} diff --git a/src/libstore/build/personality.hh b/src/libstore/build/personality.hh new file mode 100644 index 000000000..30e4f4062 --- /dev/null +++ b/src/libstore/build/personality.hh @@ -0,0 +1,11 @@ +#pragma once + +#include <string> + +namespace nix { + +void setPersonality(std::string_view system); + +} + + diff --git a/src/libstore/sandbox-defaults.sb b/src/libstore/build/sandbox-defaults.sb index d9d710559..77f013aea 100644 --- a/src/libstore/sandbox-defaults.sb +++ b/src/libstore/build/sandbox-defaults.sb @@ -1,3 +1,5 @@ +R""( + (define TMPDIR (param "_GLOBAL_TMP_DIR")) (deny default) @@ -104,3 +106,5 @@ (subpath "/System/Library/Apple/usr/libexec/oah") (subpath "/System/Library/LaunchDaemons/com.apple.oahd.plist") (subpath "/Library/Apple/System/Library/LaunchDaemons/com.apple.oahd.plist")) + +)"" diff --git a/src/libstore/sandbox-minimal.sb b/src/libstore/build/sandbox-minimal.sb index 65f5108b3..976a1f636 100644 --- a/src/libstore/sandbox-minimal.sb +++ b/src/libstore/build/sandbox-minimal.sb @@ -1,5 +1,9 @@ +R""( + (allow default) ; Disallow creating setuid/setgid binaries, since that ; would allow breaking build user isolation. (deny file-write-setugid) + +)"" diff --git a/src/libstore/sandbox-network.sb b/src/libstore/build/sandbox-network.sb index 56beec761..335edbaed 100644 --- a/src/libstore/sandbox-network.sb +++ b/src/libstore/build/sandbox-network.sb @@ -1,3 +1,5 @@ +R""( + ; Allow local and remote network traffic. (allow network* (local ip) (remote ip)) @@ -14,3 +16,9 @@ ; Allow DNS lookups. (allow network-outbound (remote unix-socket (path-literal "/private/var/run/mDNSResponder"))) + +; Allow access to trustd. +(allow mach-lookup (global-name "com.apple.trustd")) +(allow mach-lookup (global-name "com.apple.trustd.agent")) + +)"" diff --git a/src/libstore/build/worker.cc b/src/libstore/build/worker.cc index b192fbc77..f775f8486 100644 --- a/src/libstore/build/worker.cc +++ b/src/libstore/build/worker.cc @@ -42,7 +42,7 @@ Worker::~Worker() std::shared_ptr<DerivationGoal> Worker::makeDerivationGoalCommon( const StorePath & drvPath, - const StringSet & wantedOutputs, + const OutputsSpec & wantedOutputs, std::function<std::shared_ptr<DerivationGoal>()> mkDrvGoal) { std::weak_ptr<DerivationGoal> & goal_weak = derivationGoals[drvPath]; @@ -59,7 +59,7 @@ std::shared_ptr<DerivationGoal> Worker::makeDerivationGoalCommon( std::shared_ptr<DerivationGoal> Worker::makeDerivationGoal(const StorePath & drvPath, - const StringSet & wantedOutputs, BuildMode buildMode) + const OutputsSpec & wantedOutputs, BuildMode buildMode) { return makeDerivationGoalCommon(drvPath, wantedOutputs, [&]() -> std::shared_ptr<DerivationGoal> { return !dynamic_cast<LocalStore *>(&store) @@ -70,7 +70,7 @@ std::shared_ptr<DerivationGoal> Worker::makeDerivationGoal(const StorePath & drv std::shared_ptr<DerivationGoal> Worker::makeBasicDerivationGoal(const StorePath & drvPath, - const BasicDerivation & drv, const StringSet & wantedOutputs, BuildMode buildMode) + const BasicDerivation & drv, const OutputsSpec & wantedOutputs, BuildMode buildMode) { return makeDerivationGoalCommon(drvPath, wantedOutputs, [&]() -> std::shared_ptr<DerivationGoal> { return !dynamic_cast<LocalStore *>(&store) @@ -276,7 +276,7 @@ void Worker::run(const Goals & _topGoals) if (!children.empty() || !waitingForAWhile.empty()) waitForInput(); else { - if (awake.empty() && 0 == settings.maxBuildJobs) + if (awake.empty() && 0U == settings.maxBuildJobs) { if (getMachines().empty()) throw Error("unable to start any build; either increase '--max-jobs' " diff --git a/src/libstore/build/worker.hh b/src/libstore/build/worker.hh index a1e036a96..6d68d3cf1 100644 --- a/src/libstore/build/worker.hh +++ b/src/libstore/build/worker.hh @@ -140,15 +140,15 @@ public: /* derivation goal */ private: std::shared_ptr<DerivationGoal> makeDerivationGoalCommon( - const StorePath & drvPath, const StringSet & wantedOutputs, + const StorePath & drvPath, const OutputsSpec & wantedOutputs, std::function<std::shared_ptr<DerivationGoal>()> mkDrvGoal); public: std::shared_ptr<DerivationGoal> makeDerivationGoal( const StorePath & drvPath, - const StringSet & wantedOutputs, BuildMode buildMode = bmNormal); + const OutputsSpec & wantedOutputs, BuildMode buildMode = bmNormal); std::shared_ptr<DerivationGoal> makeBasicDerivationGoal( const StorePath & drvPath, const BasicDerivation & drv, - const StringSet & wantedOutputs, BuildMode buildMode = bmNormal); + const OutputsSpec & wantedOutputs, BuildMode buildMode = bmNormal); /* substitution goal */ std::shared_ptr<PathSubstitutionGoal> makePathSubstitutionGoal(const StorePath & storePath, RepairFlag repair = NoRepair, std::optional<ContentAddress> ca = std::nullopt); diff --git a/src/libstore/builtins/buildenv.cc b/src/libstore/builtins/buildenv.cc index 47458a388..7bba33fb9 100644 --- a/src/libstore/builtins/buildenv.cc +++ b/src/libstore/builtins/buildenv.cc @@ -92,13 +92,11 @@ static void createLinks(State & state, const Path & srcDir, const Path & dstDir, if (S_ISLNK(dstSt.st_mode)) { auto prevPriority = state.priorities[dstFile]; if (prevPriority == priority) - throw Error( - "files '%1%' and '%2%' have the same priority %3%; " - "use 'nix-env --set-flag priority NUMBER INSTALLED_PKGNAME' " - "or type 'nix profile install --help' if using 'nix profile' to find out how" - "to change the priority of one of the conflicting packages" - " (0 being the highest priority)", - srcFile, readLink(dstFile), priority); + throw BuildEnvFileConflictError( + readLink(dstFile), + srcFile, + priority + ); if (prevPriority < priority) continue; if (unlink(dstFile.c_str()) == -1) diff --git a/src/libstore/builtins/buildenv.hh b/src/libstore/builtins/buildenv.hh index 73c0f5f7f..a018de3af 100644 --- a/src/libstore/builtins/buildenv.hh +++ b/src/libstore/builtins/buildenv.hh @@ -12,6 +12,32 @@ struct Package { Package(const Path & path, bool active, int priority) : path{path}, active{active}, priority{priority} {} }; +class BuildEnvFileConflictError : public Error +{ +public: + const Path fileA; + const Path fileB; + int priority; + + BuildEnvFileConflictError( + const Path fileA, + const Path fileB, + int priority + ) + : Error( + "Unable to build profile. There is a conflict for the following files:\n" + "\n" + " %1%\n" + " %2%", + fileA, + fileB + ) + , fileA(fileA) + , fileB(fileB) + , priority(priority) + {} +}; + typedef std::vector<Package> Packages; void buildProfile(const Path & out, Packages && pkgs); diff --git a/src/libstore/builtins/unpack-channel.cc b/src/libstore/builtins/unpack-channel.cc index 426d58a53..ba04bb16c 100644 --- a/src/libstore/builtins/unpack-channel.cc +++ b/src/libstore/builtins/unpack-channel.cc @@ -22,8 +22,7 @@ void builtinUnpackChannel(const BasicDerivation & drv) auto entries = readDirectory(out); if (entries.size() != 1) throw Error("channel tarball '%s' contains more than one file", src); - if (rename((out + "/" + entries[0].name).c_str(), (out + "/" + channelName).c_str()) == -1) - throw SysError("renaming channel directory"); + renameFile((out + "/" + entries[0].name), (out + "/" + channelName)); } } diff --git a/src/libstore/daemon.cc b/src/libstore/daemon.cc index de69b50ee..7f8b0f905 100644 --- a/src/libstore/daemon.cc +++ b/src/libstore/daemon.cc @@ -67,12 +67,12 @@ struct TunnelLogger : public Logger state->pendingMsgs.push_back(s); } - void log(Verbosity lvl, const FormatOrString & fs) override + void log(Verbosity lvl, std::string_view s) override { if (lvl > verbosity) return; StringSink buf; - buf << STDERR_NEXT << (fs.s + "\n"); + buf << STDERR_NEXT << (s + "\n"); enqueueMsg(buf.s); } @@ -222,7 +222,8 @@ struct ClientSettings else if (!hasSuffix(s, "/") && trusted.count(s + "/")) subs.push_back(s + "/"); else - warn("ignoring untrusted substituter '%s'", s); + warn("ignoring untrusted substituter '%s', you are not a trusted user.\n" + "Run `man nix.conf` for more information on the `substituters` configuration option.", s); res = subs; return true; }; @@ -235,10 +236,15 @@ struct ClientSettings // the daemon, as that could cause some pretty weird stuff if (parseFeatures(tokenizeString<StringSet>(value)) != settings.experimentalFeatures.get()) debug("Ignoring the client-specified experimental features"); + } else if (name == settings.pluginFiles.name) { + if (tokenizeString<Paths>(value) != settings.pluginFiles.get()) + warn("Ignoring the client-specified plugin-files.\n" + "The client specifying plugins to the daemon never made sense, and was removed in Nix >=2.14."); } else if (trusted || name == settings.buildTimeout.name - || name == settings.buildRepeat.name + || name == settings.maxSilentTime.name + || name == settings.pollInterval.name || name == "connect-timeout" || (name == "builders" && value == "")) settings.set(name, value); @@ -523,7 +529,14 @@ static void performOp(TunnelLogger * logger, ref<Store> store, mode = (BuildMode) readInt(from); /* Repairing is not atomic, so disallowed for "untrusted" - clients. */ + clients. + + FIXME: layer violation in this message: the daemon code (i.e. + this file) knows whether a client/connection is trusted, but it + does not how how the client was authenticated. The mechanism + need not be getting the UID of the other end of a Unix Domain + Socket. + */ if (mode == bmRepair && !trusted) throw Error("repairing is not allowed because you are not in 'trusted-users'"); } @@ -540,7 +553,9 @@ static void performOp(TunnelLogger * logger, ref<Store> store, mode = (BuildMode) readInt(from); /* Repairing is not atomic, so disallowed for "untrusted" - clients. */ + clients. + + FIXME: layer violation; see above. */ if (mode == bmRepair && !trusted) throw Error("repairing is not allowed because you are not in 'trusted-users'"); @@ -979,8 +994,7 @@ void processConnection( FdSource & from, FdSink & to, TrustedFlag trusted, - RecursiveFlag recursive, - std::function<void(Store &)> authHook) + RecursiveFlag recursive) { auto monitor = !recursive ? std::make_unique<MonitorFdHup>(from.fd) : nullptr; @@ -1023,10 +1037,6 @@ void processConnection( try { - /* If we can't accept clientVersion, then throw an error - *here* (not above). */ - authHook(*store); - tunnelLogger->stopWork(); to.flush(); diff --git a/src/libstore/daemon.hh b/src/libstore/daemon.hh index 67755d54e..8c765615c 100644 --- a/src/libstore/daemon.hh +++ b/src/libstore/daemon.hh @@ -13,11 +13,6 @@ void processConnection( FdSource & from, FdSink & to, TrustedFlag trusted, - RecursiveFlag recursive, - /* Arbitrary hook to check authorization / initialize user data / whatever - after the protocol has been negotiated. The idea is that this function - and everything it calls doesn't know about this stuff, and the - `nix-daemon` handles that instead. */ - std::function<void(Store &)> authHook); + RecursiveFlag recursive); } diff --git a/src/libstore/derivations.cc b/src/libstore/derivations.cc index fe99c3c5e..05dc9a3cc 100644 --- a/src/libstore/derivations.cc +++ b/src/libstore/derivations.cc @@ -5,6 +5,7 @@ #include "worker-protocol.hh" #include "fs-accessor.hh" #include <boost/container/small_vector.hpp> +#include <nlohmann/json.hpp> namespace nix { @@ -448,7 +449,7 @@ std::string Derivation::unparse(const Store & store, bool maskOutputs, // FIXME: remove -bool isDerivation(const std::string & fileName) +bool isDerivation(std::string_view fileName) { return hasSuffix(fileName, drvExtension); } @@ -688,12 +689,6 @@ std::map<std::string, Hash> staticOutputHashes(Store & store, const Derivation & } -bool wantOutput(const std::string & output, const std::set<std::string> & wanted) -{ - return wanted.empty() || wanted.find(output) != wanted.end(); -} - - static DerivationOutput readDerivationOutput(Source & in, const Store & store) { const auto pathS = readString(in); @@ -896,4 +891,63 @@ std::optional<BasicDerivation> Derivation::tryResolve( const Hash impureOutputHash = hashString(htSHA256, "impure"); +nlohmann::json DerivationOutput::toJSON( + const Store & store, std::string_view drvName, std::string_view outputName) const +{ + nlohmann::json res = nlohmann::json::object(); + std::visit(overloaded { + [&](const DerivationOutput::InputAddressed & doi) { + res["path"] = store.printStorePath(doi.path); + }, + [&](const DerivationOutput::CAFixed & dof) { + res["path"] = store.printStorePath(dof.path(store, drvName, outputName)); + res["hashAlgo"] = dof.hash.printMethodAlgo(); + res["hash"] = dof.hash.hash.to_string(Base16, false); + }, + [&](const DerivationOutput::CAFloating & dof) { + res["hashAlgo"] = makeFileIngestionPrefix(dof.method) + printHashType(dof.hashType); + }, + [&](const DerivationOutput::Deferred &) {}, + [&](const DerivationOutput::Impure & doi) { + res["hashAlgo"] = makeFileIngestionPrefix(doi.method) + printHashType(doi.hashType); + res["impure"] = true; + }, + }, raw()); + return res; +} + +nlohmann::json Derivation::toJSON(const Store & store) const +{ + nlohmann::json res = nlohmann::json::object(); + + { + nlohmann::json & outputsObj = res["outputs"]; + outputsObj = nlohmann::json::object(); + for (auto & [outputName, output] : outputs) { + outputsObj[outputName] = output.toJSON(store, name, outputName); + } + } + + { + auto& inputsList = res["inputSrcs"]; + inputsList = nlohmann::json ::array(); + for (auto & input : inputSrcs) + inputsList.emplace_back(store.printStorePath(input)); + } + + { + auto& inputDrvsObj = res["inputDrvs"]; + inputDrvsObj = nlohmann::json ::object(); + for (auto & input : inputDrvs) + inputDrvsObj[store.printStorePath(input.first)] = input.second; + } + + res["system"] = platform; + res["builder"] = builder; + res["args"] = args; + res["env"] = env; + + return res; +} + } diff --git a/src/libstore/derivations.hh b/src/libstore/derivations.hh index af198a767..8456b29e7 100644 --- a/src/libstore/derivations.hh +++ b/src/libstore/derivations.hh @@ -13,6 +13,7 @@ namespace nix { +class Store; /* Abstract syntax of derivations. */ @@ -82,6 +83,11 @@ struct DerivationOutput : _DerivationOutputRaw inline const Raw & raw() const { return static_cast<const Raw &>(*this); } + + nlohmann::json toJSON( + const Store & store, + std::string_view drvName, + std::string_view outputName) const; }; typedef std::map<std::string, DerivationOutput> DerivationOutputs; @@ -209,6 +215,8 @@ struct Derivation : BasicDerivation Derivation() = default; Derivation(const BasicDerivation & bd) : BasicDerivation(bd) { } Derivation(BasicDerivation && bd) : BasicDerivation(std::move(bd)) { } + + nlohmann::json toJSON(const Store & store) const; }; @@ -224,7 +232,7 @@ StorePath writeDerivation(Store & store, Derivation parseDerivation(const Store & store, std::string && s, std::string_view name); // FIXME: remove -bool isDerivation(const std::string & fileName); +bool isDerivation(std::string_view fileName); /* Calculate the name that will be used for the store path for this output. @@ -294,8 +302,6 @@ typedef std::map<StorePath, DrvHash> DrvHashes; // FIXME: global, though at least thread-safe. extern Sync<DrvHashes> drvHashes; -bool wantOutput(const std::string & output, const std::set<std::string> & wanted); - struct Source; struct Sink; diff --git a/src/libstore/derived-path.cc b/src/libstore/derived-path.cc index 44587ae78..e0d86a42f 100644 --- a/src/libstore/derived-path.cc +++ b/src/libstore/derived-path.cc @@ -19,12 +19,13 @@ nlohmann::json DerivedPath::Built::toJSON(ref<Store> store) const { res["drvPath"] = store->printStorePath(drvPath); // Fallback for the input-addressed derivation case: We expect to always be // able to print the output paths, so let’s do it - const auto knownOutputs = store->queryPartialDerivationOutputMap(drvPath); - for (const auto& output : outputs) { - auto knownOutput = get(knownOutputs, output); - res["outputs"][output] = (knownOutput && *knownOutput) - ? store->printStorePath(**knownOutput) - : nullptr; + const auto outputMap = store->queryPartialDerivationOutputMap(drvPath); + for (const auto & [output, outputPathOpt] : outputMap) { + if (!outputs.contains(output)) continue; + if (outputPathOpt) + res["outputs"][output] = store->printStorePath(*outputPathOpt); + else + res["outputs"][output] = nullptr; } return res; } @@ -53,31 +54,16 @@ StorePathSet BuiltPath::outPaths() const ); } -template<typename T> -nlohmann::json stuffToJSON(const std::vector<T> & ts, ref<Store> store) { - auto res = nlohmann::json::array(); - for (const T & t : ts) { - std::visit([&res, store](const auto & t) { - res.push_back(t.toJSON(store)); - }, t.raw()); - } - return res; -} - -nlohmann::json derivedPathsWithHintsToJSON(const BuiltPaths & buildables, ref<Store> store) -{ return stuffToJSON<BuiltPath>(buildables, store); } -nlohmann::json derivedPathsToJSON(const DerivedPaths & paths, ref<Store> store) -{ return stuffToJSON<DerivedPath>(paths, store); } - - -std::string DerivedPath::Opaque::to_string(const Store & store) const { +std::string DerivedPath::Opaque::to_string(const Store & store) const +{ return store.printStorePath(path); } -std::string DerivedPath::Built::to_string(const Store & store) const { +std::string DerivedPath::Built::to_string(const Store & store) const +{ return store.printStorePath(drvPath) + "!" - + (outputs.empty() ? std::string { "*" } : concatStringsSep(",", outputs)); + + outputs.to_string(); } std::string DerivedPath::to_string(const Store & store) const @@ -93,16 +79,12 @@ DerivedPath::Opaque DerivedPath::Opaque::parse(const Store & store, std::string_ return {store.parseStorePath(s)}; } -DerivedPath::Built DerivedPath::Built::parse(const Store & store, std::string_view s) +DerivedPath::Built DerivedPath::Built::parse(const Store & store, std::string_view drvS, std::string_view outputsS) { - size_t n = s.find("!"); - assert(n != s.npos); - auto drvPath = store.parseStorePath(s.substr(0, n)); - auto outputsS = s.substr(n + 1); - std::set<std::string> outputs; - if (outputsS != "*") - outputs = tokenizeString<std::set<std::string>>(outputsS, ","); - return {drvPath, outputs}; + return { + .drvPath = store.parseStorePath(drvS), + .outputs = OutputsSpec::parse(outputsS), + }; } DerivedPath DerivedPath::parse(const Store & store, std::string_view s) @@ -110,7 +92,7 @@ DerivedPath DerivedPath::parse(const Store & store, std::string_view s) size_t n = s.find("!"); return n == s.npos ? (DerivedPath) DerivedPath::Opaque::parse(store, s) - : (DerivedPath) DerivedPath::Built::parse(store, s); + : (DerivedPath) DerivedPath::Built::parse(store, s.substr(0, n), s.substr(n + 1)); } RealisedPath::Set BuiltPath::toRealisedPaths(Store & store) const diff --git a/src/libstore/derived-path.hh b/src/libstore/derived-path.hh index 24a0ae773..9e0cce377 100644 --- a/src/libstore/derived-path.hh +++ b/src/libstore/derived-path.hh @@ -3,8 +3,10 @@ #include "util.hh" #include "path.hh" #include "realisation.hh" +#include "outputs-spec.hh" +#include "comparator.hh" -#include <optional> +#include <variant> #include <nlohmann/json_fwd.hpp> @@ -26,8 +28,7 @@ struct DerivedPathOpaque { std::string to_string(const Store & store) const; static DerivedPathOpaque parse(const Store & store, std::string_view); - bool operator < (const DerivedPathOpaque & b) const - { return path < b.path; } + GENERATE_CMP(DerivedPathOpaque, me->path); }; /** @@ -44,14 +45,13 @@ struct DerivedPathOpaque { */ struct DerivedPathBuilt { StorePath drvPath; - std::set<std::string> outputs; + OutputsSpec outputs; std::string to_string(const Store & store) const; - static DerivedPathBuilt parse(const Store & store, std::string_view); + static DerivedPathBuilt parse(const Store & store, std::string_view, std::string_view); nlohmann::json toJSON(ref<Store> store) const; - bool operator < (const DerivedPathBuilt & b) const - { return std::make_pair(drvPath, outputs) < std::make_pair(b.drvPath, b.outputs); } + GENERATE_CMP(DerivedPathBuilt, me->drvPath, me->outputs); }; using _DerivedPathRaw = std::variant< @@ -95,6 +95,8 @@ struct BuiltPathBuilt { nlohmann::json toJSON(ref<Store> store) const; static BuiltPathBuilt parse(const Store & store, std::string_view); + + GENERATE_CMP(BuiltPathBuilt, me->drvPath, me->outputs); }; using _BuiltPathRaw = std::variant< @@ -125,7 +127,4 @@ struct BuiltPath : _BuiltPathRaw { typedef std::vector<DerivedPath> DerivedPaths; typedef std::vector<BuiltPath> BuiltPaths; -nlohmann::json derivedPathsWithHintsToJSON(const BuiltPaths & buildables, ref<Store> store); -nlohmann::json derivedPathsToJSON(const DerivedPaths & , ref<Store> store); - } diff --git a/src/libstore/export-import.cc b/src/libstore/export-import.cc index 9875da909..4eb838b68 100644 --- a/src/libstore/export-import.cc +++ b/src/libstore/export-import.cc @@ -16,7 +16,7 @@ void Store::exportPaths(const StorePathSet & paths, Sink & sink) //logger->incExpected(doneLabel, sorted.size()); for (auto & path : sorted) { - //Activity act(*logger, lvlInfo, format("exporting path '%s'") % path); + //Activity act(*logger, lvlInfo, "exporting path '%s'", path); sink << 1; exportPath(path, sink); //logger->incProgress(doneLabel); @@ -71,7 +71,7 @@ StorePaths Store::importPaths(Source & source, CheckSigsFlag checkSigs) auto path = parseStorePath(readString(source)); - //Activity act(*logger, lvlInfo, format("importing path '%s'") % info.path); + //Activity act(*logger, lvlInfo, "importing path '%s'", info.path); auto references = worker_proto::read(*this, source, Phantom<StorePathSet> {}); auto deriver = readString(source); diff --git a/src/libstore/filetransfer.cc b/src/libstore/filetransfer.cc index 8454ad7d2..b5fe7c03b 100644 --- a/src/libstore/filetransfer.cc +++ b/src/libstore/filetransfer.cc @@ -33,14 +33,6 @@ FileTransferSettings fileTransferSettings; static GlobalConfig::Register rFileTransferSettings(&fileTransferSettings); -std::string resolveUri(std::string_view uri) -{ - if (uri.compare(0, 8, "channel:") == 0) - return "https://nixos.org/channels/" + std::string(uri.substr(8)) + "/nixexprs.tar.xz"; - else - return std::string(uri); -} - struct curlFileTransfer : public FileTransfer { CURLM * curlm = 0; @@ -96,6 +88,10 @@ struct curlFileTransfer : public FileTransfer {request.uri}, request.parentAct) , callback(std::move(callback)) , finalSink([this](std::string_view data) { + if (errorSink) { + (*errorSink)(data); + } + if (this->request.dataCallback) { auto httpStatus = getHTTPStatus(); @@ -109,6 +105,7 @@ struct curlFileTransfer : public FileTransfer this->result.data.append(data); }) { + requestHeaders = curl_slist_append(requestHeaders, "Accept-Encoding: zstd, br, gzip, deflate, bzip2, xz"); if (!request.expectedETag.empty()) requestHeaders = curl_slist_append(requestHeaders, ("If-None-Match: " + request.expectedETag).c_str()); if (!request.mimeType.empty()) @@ -142,9 +139,9 @@ struct curlFileTransfer : public FileTransfer } template<class T> - void fail(const T & e) + void fail(T && e) { - failEx(std::make_exception_ptr(e)); + failEx(std::make_exception_ptr(std::move(e))); } LambdaSink finalSink; @@ -170,8 +167,6 @@ struct curlFileTransfer : public FileTransfer } } - if (errorSink) - (*errorSink)({(char *) contents, realSize}); (*decompressionSink)({(char *) contents, realSize}); return realSize; @@ -190,7 +185,7 @@ struct curlFileTransfer : public FileTransfer { size_t realSize = size * nmemb; std::string line((char *) contents, realSize); - printMsg(lvlVomit, format("got header for '%s': %s") % request.uri % trim(line)); + printMsg(lvlVomit, "got header for '%s': %s", request.uri, trim(line)); static std::regex statusLine("HTTP/[^ ]+ +[0-9]+(.*)", std::regex::extended | std::regex::icase); std::smatch match; if (std::regex_match(line, match, statusLine)) { @@ -214,7 +209,7 @@ struct curlFileTransfer : public FileTransfer long httpStatus = 0; curl_easy_getinfo(req, CURLINFO_RESPONSE_CODE, &httpStatus); if (result.etag == request.expectedETag && httpStatus == 200) { - debug(format("shutting down on 200 HTTP response with expected ETag")); + debug("shutting down on 200 HTTP response with expected ETag"); return 0; } } else if (name == "content-encoding") @@ -308,6 +303,9 @@ struct curlFileTransfer : public FileTransfer curl_easy_setopt(req, CURLOPT_HTTPHEADER, requestHeaders); + if (settings.downloadSpeed.get() > 0) + curl_easy_setopt(req, CURLOPT_MAX_RECV_SPEED_LARGE, (curl_off_t) (settings.downloadSpeed.get() * 1024)); + if (request.head) curl_easy_setopt(req, CURLOPT_NOBODY, 1); @@ -319,7 +317,6 @@ struct curlFileTransfer : public FileTransfer } if (request.verifyTLS) { - debug("verify TLS: Nix CA file = '%s'", settings.caFile); if (settings.caFile != "") curl_easy_setopt(req, CURLOPT_CAINFO, settings.caFile.c_str()); } else { @@ -470,7 +467,7 @@ struct curlFileTransfer : public FileTransfer fileTransfer.enqueueItem(shared_from_this()); } else - fail(exc); + fail(std::move(exc)); } } }; @@ -834,7 +831,7 @@ void FileTransfer::download(FileTransferRequest && request, Sink & sink) { auto state(_state->lock()); - while (state->data.empty()) { + if (state->data.empty()) { if (state->quit) { if (state->exc) std::rethrow_exception(state->exc); @@ -842,6 +839,8 @@ void FileTransfer::download(FileTransferRequest && request, Sink & sink) } state.wait(state->avail); + + if (state->data.empty()) continue; } chunk = std::move(state->data); @@ -871,14 +870,4 @@ FileTransferError::FileTransferError(FileTransfer::Error error, std::optional<st err.msg = hf; } -bool isUri(std::string_view s) -{ - if (s.compare(0, 8, "channel:") == 0) return true; - size_t pos = s.find("://"); - if (pos == std::string::npos) return false; - std::string scheme(s, 0, pos); - return scheme == "http" || scheme == "https" || scheme == "file" || scheme == "channel" || scheme == "git" || scheme == "s3" || scheme == "ssh"; -} - - } diff --git a/src/libstore/filetransfer.hh b/src/libstore/filetransfer.hh index 40e7cf52c..07d58f53a 100644 --- a/src/libstore/filetransfer.hh +++ b/src/libstore/filetransfer.hh @@ -125,9 +125,4 @@ public: FileTransferError(FileTransfer::Error error, std::optional<std::string> response, const Args & ... args); }; -bool isUri(std::string_view s); - -/* Resolve deprecated 'channel:<foo>' URLs. */ -std::string resolveUri(std::string_view uri); - } diff --git a/src/libstore/gc.cc b/src/libstore/gc.cc index d58ed78b1..0038ec802 100644 --- a/src/libstore/gc.cc +++ b/src/libstore/gc.cc @@ -34,14 +34,11 @@ static void makeSymlink(const Path & link, const Path & target) createDirs(dirOf(link)); /* Create the new symlink. */ - Path tempLink = (format("%1%.tmp-%2%-%3%") - % link % getpid() % random()).str(); + Path tempLink = fmt("%1%.tmp-%2%-%3%", link, getpid(), random()); createSymlink(target, tempLink); /* Atomically replace the old one. */ - if (rename(tempLink.c_str(), link.c_str()) == -1) - throw SysError("cannot rename '%1%' to '%2%'", - tempLink , link); + renameFile(tempLink, link); } @@ -79,60 +76,73 @@ Path LocalFSStore::addPermRoot(const StorePath & storePath, const Path & _gcRoot } -void LocalStore::addTempRoot(const StorePath & path) +void LocalStore::createTempRootsFile() { - auto state(_state.lock()); + auto fdTempRoots(_fdTempRoots.lock()); /* Create the temporary roots file for this process. */ - if (!state->fdTempRoots) { - - while (1) { - if (pathExists(fnTempRoots)) - /* It *must* be stale, since there can be no two - processes with the same pid. */ - unlink(fnTempRoots.c_str()); + if (*fdTempRoots) return; - state->fdTempRoots = openLockFile(fnTempRoots, true); + while (1) { + if (pathExists(fnTempRoots)) + /* It *must* be stale, since there can be no two + processes with the same pid. */ + unlink(fnTempRoots.c_str()); - debug("acquiring write lock on '%s'", fnTempRoots); - lockFile(state->fdTempRoots.get(), ltWrite, true); + *fdTempRoots = openLockFile(fnTempRoots, true); - /* Check whether the garbage collector didn't get in our - way. */ - struct stat st; - if (fstat(state->fdTempRoots.get(), &st) == -1) - throw SysError("statting '%1%'", fnTempRoots); - if (st.st_size == 0) break; + debug("acquiring write lock on '%s'", fnTempRoots); + lockFile(fdTempRoots->get(), ltWrite, true); - /* The garbage collector deleted this file before we could - get a lock. (It won't delete the file after we get a - lock.) Try again. */ - } + /* Check whether the garbage collector didn't get in our + way. */ + struct stat st; + if (fstat(fdTempRoots->get(), &st) == -1) + throw SysError("statting '%1%'", fnTempRoots); + if (st.st_size == 0) break; + /* The garbage collector deleted this file before we could get + a lock. (It won't delete the file after we get a lock.) + Try again. */ } +} + + +void LocalStore::addTempRoot(const StorePath & path) +{ + createTempRootsFile(); - if (!state->fdGCLock) - state->fdGCLock = openGCLock(); + /* Open/create the global GC lock file. */ + { + auto fdGCLock(_fdGCLock.lock()); + if (!*fdGCLock) + *fdGCLock = openGCLock(); + } restart: - FdLock gcLock(state->fdGCLock.get(), ltRead, false, ""); + /* Try to acquire a shared global GC lock (non-blocking). This + only succeeds if the garbage collector is not currently + running. */ + FdLock gcLock(_fdGCLock.lock()->get(), ltRead, false, ""); if (!gcLock.acquired) { /* We couldn't get a shared global GC lock, so the garbage collector is running. So we have to connect to the garbage collector and inform it about our root. */ - if (!state->fdRootsSocket) { + auto fdRootsSocket(_fdRootsSocket.lock()); + + if (!*fdRootsSocket) { auto socketPath = stateDir.get() + gcSocketPath; debug("connecting to '%s'", socketPath); - state->fdRootsSocket = createUnixDomainSocket(); + *fdRootsSocket = createUnixDomainSocket(); try { - nix::connect(state->fdRootsSocket.get(), socketPath); + nix::connect(fdRootsSocket->get(), socketPath); } catch (SysError & e) { /* The garbage collector may have exited, so we need to restart. */ if (e.errNo == ECONNREFUSED) { debug("GC socket connection refused"); - state->fdRootsSocket.close(); + fdRootsSocket->close(); goto restart; } throw; @@ -141,30 +151,31 @@ void LocalStore::addTempRoot(const StorePath & path) try { debug("sending GC root '%s'", printStorePath(path)); - writeFull(state->fdRootsSocket.get(), printStorePath(path) + "\n", false); + writeFull(fdRootsSocket->get(), printStorePath(path) + "\n", false); char c; - readFull(state->fdRootsSocket.get(), &c, 1); + readFull(fdRootsSocket->get(), &c, 1); assert(c == '1'); debug("got ack for GC root '%s'", printStorePath(path)); } catch (SysError & e) { /* The garbage collector may have exited, so we need to restart. */ - if (e.errNo == EPIPE) { + if (e.errNo == EPIPE || e.errNo == ECONNRESET) { debug("GC socket disconnected"); - state->fdRootsSocket.close(); + fdRootsSocket->close(); goto restart; } throw; } catch (EndOfFile & e) { debug("GC socket disconnected"); - state->fdRootsSocket.close(); + fdRootsSocket->close(); goto restart; } } - /* Append the store path to the temporary roots file. */ + /* Record the store path in the temporary roots file so it will be + seen by a future run of the garbage collector. */ auto s = printStorePath(path) + '\0'; - writeFull(state->fdTempRoots.get(), s); + writeFull(_fdTempRoots.lock()->get(), s); } @@ -185,7 +196,7 @@ void LocalStore::findTempRoots(Roots & tempRoots, bool censor) pid_t pid = std::stoi(i.name); - debug(format("reading temporary root file '%1%'") % path); + debug("reading temporary root file '%1%'", path); AutoCloseFD fd(open(path.c_str(), O_CLOEXEC | O_RDWR, 0666)); if (!fd) { /* It's okay if the file has disappeared. */ @@ -251,7 +262,7 @@ void LocalStore::findRoots(const Path & path, unsigned char type, Roots & roots) target = absPath(target, dirOf(path)); if (!pathExists(target)) { if (isInDir(path, stateDir + "/" + gcRootsDir + "/auto")) { - printInfo(format("removing stale link from '%1%' to '%2%'") % path % target); + printInfo("removing stale link from '%1%' to '%2%'", path, target); unlink(path.c_str()); } } else { @@ -360,29 +371,29 @@ void LocalStore::findRuntimeRoots(Roots & roots, bool censor) while (errno = 0, ent = readdir(procDir.get())) { checkInterrupt(); if (std::regex_match(ent->d_name, digitsRegex)) { - readProcLink(fmt("/proc/%s/exe" ,ent->d_name), unchecked); - readProcLink(fmt("/proc/%s/cwd", ent->d_name), unchecked); - - auto fdStr = fmt("/proc/%s/fd", ent->d_name); - auto fdDir = AutoCloseDir(opendir(fdStr.c_str())); - if (!fdDir) { - if (errno == ENOENT || errno == EACCES) - continue; - throw SysError("opening %1%", fdStr); - } - struct dirent * fd_ent; - while (errno = 0, fd_ent = readdir(fdDir.get())) { - if (fd_ent->d_name[0] != '.') - readProcLink(fmt("%s/%s", fdStr, fd_ent->d_name), unchecked); - } - if (errno) { - if (errno == ESRCH) - continue; - throw SysError("iterating /proc/%1%/fd", ent->d_name); - } - fdDir.reset(); - try { + readProcLink(fmt("/proc/%s/exe" ,ent->d_name), unchecked); + readProcLink(fmt("/proc/%s/cwd", ent->d_name), unchecked); + + auto fdStr = fmt("/proc/%s/fd", ent->d_name); + auto fdDir = AutoCloseDir(opendir(fdStr.c_str())); + if (!fdDir) { + if (errno == ENOENT || errno == EACCES) + continue; + throw SysError("opening %1%", fdStr); + } + struct dirent * fd_ent; + while (errno = 0, fd_ent = readdir(fdDir.get())) { + if (fd_ent->d_name[0] != '.') + readProcLink(fmt("%s/%s", fdStr, fd_ent->d_name), unchecked); + } + if (errno) { + if (errno == ESRCH) + continue; + throw SysError("iterating /proc/%1%/fd", ent->d_name); + } + fdDir.reset(); + auto mapFile = fmt("/proc/%s/maps", ent->d_name); auto mapLines = tokenizeString<std::vector<std::string>>(readFile(mapFile), "\n"); for (const auto & line : mapLines) { @@ -508,6 +519,7 @@ void LocalStore::collectGarbage(const GCOptions & options, GCResults & results) Finally cleanup([&]() { debug("GC roots server shutting down"); + fdServer.close(); while (true) { auto item = remove_begin(*connections.lock()); if (!item) break; @@ -621,6 +633,17 @@ void LocalStore::collectGarbage(const GCOptions & options, GCResults & results) Path path = storeDir + "/" + std::string(baseName); Path realPath = realStoreDir + "/" + std::string(baseName); + /* There may be temp directories in the store that are still in use + by another process. We need to be sure that we can acquire an + exclusive lock before deleting them. */ + if (baseName.find("tmp-", 0) == 0) { + AutoCloseFD tmpDirFd = open(realPath.c_str(), O_RDONLY | O_DIRECTORY); + if (tmpDirFd.get() == -1 || !lockFile(tmpDirFd.get(), ltWrite, false)) { + debug("skipping locked tempdir '%s'", realPath); + return; + } + } + printInfo("deleting '%1%'", path); results.paths.insert(path); @@ -839,7 +862,7 @@ void LocalStore::collectGarbage(const GCOptions & options, GCResults & results) continue; } - printMsg(lvlTalkative, format("deleting unused link '%1%'") % path); + printMsg(lvlTalkative, "deleting unused link '%1%'", path); if (unlink(path.c_str()) == -1) throw SysError("deleting '%1%'", path); diff --git a/src/libstore/globals.cc b/src/libstore/globals.cc index d724897bb..fae79c1a0 100644 --- a/src/libstore/globals.cc +++ b/src/libstore/globals.cc @@ -30,15 +30,15 @@ static GlobalConfig::Register rSettings(&settings); Settings::Settings() : nixPrefix(NIX_PREFIX) - , nixStore(canonPath(getEnv("NIX_STORE_DIR").value_or(getEnv("NIX_STORE").value_or(NIX_STORE_DIR)))) - , nixDataDir(canonPath(getEnv("NIX_DATA_DIR").value_or(NIX_DATA_DIR))) - , nixLogDir(canonPath(getEnv("NIX_LOG_DIR").value_or(NIX_LOG_DIR))) - , nixStateDir(canonPath(getEnv("NIX_STATE_DIR").value_or(NIX_STATE_DIR))) - , nixConfDir(canonPath(getEnv("NIX_CONF_DIR").value_or(NIX_CONF_DIR))) + , nixStore(canonPath(getEnvNonEmpty("NIX_STORE_DIR").value_or(getEnvNonEmpty("NIX_STORE").value_or(NIX_STORE_DIR)))) + , nixDataDir(canonPath(getEnvNonEmpty("NIX_DATA_DIR").value_or(NIX_DATA_DIR))) + , nixLogDir(canonPath(getEnvNonEmpty("NIX_LOG_DIR").value_or(NIX_LOG_DIR))) + , nixStateDir(canonPath(getEnvNonEmpty("NIX_STATE_DIR").value_or(NIX_STATE_DIR))) + , nixConfDir(canonPath(getEnvNonEmpty("NIX_CONF_DIR").value_or(NIX_CONF_DIR))) , nixUserConfFiles(getUserConfigFiles()) - , nixBinDir(canonPath(getEnv("NIX_BIN_DIR").value_or(NIX_BIN_DIR))) + , nixBinDir(canonPath(getEnvNonEmpty("NIX_BIN_DIR").value_or(NIX_BIN_DIR))) , nixManDir(canonPath(NIX_MAN_DIR)) - , nixDaemonSocketFile(canonPath(getEnv("NIX_DAEMON_SOCKET_PATH").value_or(nixStateDir + DEFAULT_SOCKET_PATH))) + , nixDaemonSocketFile(canonPath(getEnvNonEmpty("NIX_DAEMON_SOCKET_PATH").value_or(nixStateDir + DEFAULT_SOCKET_PATH))) { buildUsersGroup = getuid() == 0 ? "nixbld" : ""; lockCPU = getEnv("NIX_AFFINITY_HACK") == "1"; @@ -131,6 +131,10 @@ StringSet Settings::getDefaultSystemFeatures() StringSet features{"nixos-test", "benchmark", "big-parallel"}; #if __linux__ + features.insert("uid-range"); + #endif + + #if __linux__ if (access("/dev/kvm", R_OK | W_OK) == 0) features.insert("kvm"); #endif @@ -154,13 +158,9 @@ StringSet Settings::getDefaultExtraPlatforms() // machines. Note that we can’t force processes from executing // x86_64 in aarch64 environments or vice versa since they can // always exec with their own binary preferences. - if (pathExists("/Library/Apple/System/Library/LaunchDaemons/com.apple.oahd.plist") || - pathExists("/System/Library/LaunchDaemons/com.apple.oahd.plist")) { - if (std::string{SYSTEM} == "x86_64-darwin") - extraPlatforms.insert("aarch64-darwin"); - else if (std::string{SYSTEM} == "aarch64-darwin") - extraPlatforms.insert("x86_64-darwin"); - } + if (std::string{SYSTEM} == "aarch64-darwin" && + runProgram(RunOptions {.program = "arch", .args = {"-arch", "x86_64", "/usr/bin/true"}, .mergeStderrToStdout = true}).first == 0) + extraPlatforms.insert("x86_64-darwin"); #endif return extraPlatforms; @@ -222,19 +222,19 @@ template<> void BaseSetting<SandboxMode>::convertToArg(Args & args, const std::s .longName = name, .description = "Enable sandboxing.", .category = category, - .handler = {[=]() { override(smEnabled); }} + .handler = {[this]() { override(smEnabled); }} }); args.addFlag({ .longName = "no-" + name, .description = "Disable sandboxing.", .category = category, - .handler = {[=]() { override(smDisabled); }} + .handler = {[this]() { override(smDisabled); }} }); args.addFlag({ .longName = "relaxed-" + name, .description = "Enable sandboxing, but allow builds to disable it.", .category = category, - .handler = {[=]() { override(smRelaxed); }} + .handler = {[this]() { override(smRelaxed); }} }); } @@ -291,4 +291,18 @@ void initPlugins() settings.pluginFiles.pluginsLoaded = true; } +static bool initLibStoreDone = false; + +void assertLibStoreInitialized() { + if (!initLibStoreDone) { + printError("The program must call nix::initNix() before calling any libstore library functions."); + abort(); + }; +} + +void initLibStore() { + initLibStoreDone = true; +} + + } diff --git a/src/libstore/globals.hh b/src/libstore/globals.hh index d7f351166..93086eaf8 100644 --- a/src/libstore/globals.hh +++ b/src/libstore/globals.hh @@ -46,6 +46,14 @@ struct PluginFilesSetting : public BaseSetting<Paths> void set(const std::string & str, bool append = false) override; }; +const uint32_t maxIdsPerBuild = + #if __linux__ + 1 << 16 + #else + 1 + #endif + ; + class Settings : public Config { unsigned int getDefaultCores(); @@ -193,7 +201,16 @@ public: {"build-timeout"}}; PathSetting buildHook{this, true, "", "build-hook", - "The path of the helper program that executes builds to remote machines."}; + R"( + The path to the helper program that executes remote builds. + + Nix communicates with the build hook over `stdio` using a custom protocol to request builds that cannot be performed directly by the Nix daemon. + The default value is the internal Nix binary that implements remote building. + + > **Important** + > + > Change this setting only if you really know what you’re doing. + )"}; Setting<std::string> builders{ this, "@" + nixConfDir + "/machines", "builders", @@ -271,9 +288,70 @@ public: If the build users group is empty, builds will be performed under the uid of the Nix process (that is, the uid of the caller if `NIX_REMOTE` is empty, the uid under which the Nix daemon runs if - `NIX_REMOTE` is `daemon`). Obviously, this should not be used in - multi-user settings with untrusted users. + `NIX_REMOTE` is `daemon`). Obviously, this should not be used + with a nix daemon accessible to untrusted clients. + + Defaults to `nixbld` when running as root, *empty* otherwise. + )", + {}, false}; + + Setting<bool> autoAllocateUids{this, false, "auto-allocate-uids", + R"( + Whether to select UIDs for builds automatically, instead of using the + users in `build-users-group`. + + UIDs are allocated starting at 872415232 (0x34000000) on Linux and 56930 on macOS. + + > **Warning** + > This is an experimental feature. + + To enable it, add the following to [`nix.conf`](#): + + ``` + extra-experimental-features = auto-allocate-uids + auto-allocate-uids = true + ``` + )"}; + + Setting<uint32_t> startId{this, + #if __linux__ + 0x34000000, + #else + 56930, + #endif + "start-id", + "The first UID and GID to use for dynamic ID allocation."}; + + Setting<uint32_t> uidCount{this, + #if __linux__ + maxIdsPerBuild * 128, + #else + 128, + #endif + "id-count", + "The number of UIDs/GIDs to use for dynamic ID allocation."}; + + #if __linux__ + Setting<bool> useCgroups{ + this, false, "use-cgroups", + R"( + Whether to execute builds inside cgroups. + This is only supported on Linux. + + Cgroups are required and enabled automatically for derivations + that require the `uid-range` system feature. + + > **Warning** + > This is an experimental feature. + + To enable it, add the following to [`nix.conf`](#): + + ``` + extra-experimental-features = cgroups + use-cgroups = true + ``` )"}; + #endif Setting<bool> impersonateLinux26{this, false, "impersonate-linux-26", "Whether to impersonate a Linux 2.6 machine on newer kernels.", @@ -307,11 +385,6 @@ public: )", {"build-max-log-size"}}; - /* When buildRepeat > 0 and verboseBuild == true, whether to print - repeated builds (i.e. builds other than the first one) to - stderr. Hack to prevent Hydra logs from being polluted. */ - bool printRepeatedBuilds = true; - Setting<unsigned int> pollInterval{this, 5, "build-poll-interval", "How often (in seconds) to poll for locks."}; @@ -427,6 +500,9 @@ public: for example, `/dev/nvidiactl?` specifies that `/dev/nvidiactl` will only be mounted in the sandbox if it exists in the host filesystem. + If the source is in the Nix store, then its closure will be added to + the sandbox as well. + Depending on how Nix was built, the default value for this option may be empty or provide `/bin/sh` as a bind-mount of `bash`. )", @@ -435,19 +511,6 @@ public: Setting<bool> sandboxFallback{this, true, "sandbox-fallback", "Whether to disable sandboxing when the kernel doesn't allow it."}; - Setting<size_t> buildRepeat{ - this, 0, "repeat", - R"( - How many times to repeat builds to check whether they are - deterministic. The default value is 0. If the value is non-zero, - every build is repeated the specified number of times. If the - contents of any of the runs differs from the previous ones and - `enforce-determinism` is true, the build is rejected and the - resulting store paths are not registered as “valid” in Nix’s - database. - )", - {"build-repeat"}}; - #if __linux__ Setting<std::string> sandboxShmSize{ this, "50%", "sandbox-dev-shm-size", @@ -511,20 +574,20 @@ public: configuration file, and cannot be passed at the command line. )"}; - Setting<bool> enforceDeterminism{ - this, true, "enforce-determinism", - "Whether to fail if repeated builds produce different output. See `repeat`."}; - Setting<Strings> trustedPublicKeys{ this, {"cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY="}, "trusted-public-keys", R"( - A whitespace-separated list of public keys. When paths are copied - from another Nix store (such as a binary cache), they must be - signed with one of these keys. For example: - `cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY= - hydra.nixos.org-1:CNHJZBh9K4tP3EKF6FkkgeVYsS3ohTl+oS0Qa8bezVs=`. + A whitespace-separated list of public keys. + + At least one of the following condition must be met + for Nix to accept copying a store object from another + Nix store (such as a substituter): + + - the store object has been signed using a key in the trusted keys list + - the [`require-sigs`](#conf-require-sigs) option has been set to `false` + - the store object is [output-addressed](@docroot@/glossary.md#gloss-output-addressed-store-object) )", {"binary-cache-public-keys"}}; @@ -560,9 +623,15 @@ public: R"( If set to `true` (the default), any non-content-addressed path added or copied to the Nix store (e.g. when substituting from a binary - cache) must have a valid signature, that is, be signed using one of - the keys listed in `trusted-public-keys` or `secret-key-files`. Set - to `false` to disable signature checking. + cache) must have a signature by a trusted key. A trusted key is one + listed in `trusted-public-keys`, or a public key counterpart to a + private key stored in a file listed in `secret-key-files`. + + Set to `false` to disable signature checking and trust all + non-content-addressed paths unconditionally. + + (Content-addressed paths are inherently trustworthy and thus + unaffected by this configuration option.) )"}; Setting<StringSet> extraPlatforms{ @@ -613,6 +682,15 @@ public: are tried based on their Priority value, which each substituter can set independently. Lower value means higher priority. The default is `https://cache.nixos.org`, with a Priority of 40. + + At least one of the following conditions must be met for Nix to use + a substituter: + + - the substituter is in the [`trusted-substituters`](#conf-trusted-substituters) list + - the user calling Nix is in the [`trusted-users`](#conf-trusted-users) list + + In addition, each store path should be trusted as described + in [`trusted-public-keys`](#conf-trusted-public-keys) )", {"binary-caches"}}; @@ -627,24 +705,6 @@ public: )", {"trusted-binary-caches"}}; - Setting<Strings> trustedUsers{ - this, {"root"}, "trusted-users", - R"( - A list of names of users (separated by whitespace) that have - additional rights when connecting to the Nix daemon, such as the - ability to specify additional binary caches, or to import unsigned - NARs. You can also specify groups by prefixing them with `@`; for - instance, `@wheel` means all users in the `wheel` group. The default - is `root`. - - > **Warning** - > - > Adding a user to `trusted-users` is essentially equivalent to - > giving that user root access to the system. For example, the user - > can set `sandbox-paths` and thereby obtain read access to - > directories that are otherwise inacessible to them. - )"}; - Setting<unsigned int> ttlNegativeNarInfoCache{ this, 3600, "narinfo-cache-negative-ttl", R"( @@ -667,18 +727,6 @@ public: mismatch if the build isn't reproducible. )"}; - /* ?Who we trust to use the daemon in safe ways */ - Setting<Strings> allowedUsers{ - this, {"*"}, "allowed-users", - R"( - A list of names of users (separated by whitespace) that are allowed - to connect to the Nix daemon. As with the `trusted-users` option, - you can specify groups by prefixing them with `@`. Also, you can - allow all users by specifying `*`. The default is `*`. - - Note that trusted users are always allowed to connect. - )"}; - Setting<bool> printMissing{this, true, "print-missing", "Whether to print what paths need to be built or downloaded."}; @@ -746,6 +794,13 @@ public: /nix/store/xfghy8ixrhz3kyy6p724iv3cxji088dx-bash-4.4-p23`. )"}; + Setting<unsigned int> downloadSpeed { + this, 0, "download-speed", + R"( + Specify the maximum transfer rate in kilobytes per second you want + Nix to use for downloads. + )"}; + Setting<std::string> netrcFile{ this, fmt("%s/%s", nixConfDir, "netrc"), "netrc-file", R"( @@ -899,6 +954,27 @@ public: resolves to a different location from that of the build machine. You can enable this setting if you are sure you're not going to do that. )"}; + + Setting<bool> useXDGBaseDirectories{ + this, false, "use-xdg-base-directories", + R"( + If set to `true`, Nix will conform to the [XDG Base Directory Specification] for files in `$HOME`. + The environment variables used to implement this are documented in the [Environment Variables section](@docroot@/installation/env-variables.md). + + [XDG Base Directory Specification]: https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html + + > **Warning** + > This changes the location of some well-known symlinks that Nix creates, which might break tools that rely on the old, non-XDG-conformant locations. + + In particular, the following locations change: + + | Old | New | + |-------------------|--------------------------------| + | `~/.nix-profile` | `$XDG_STATE_HOME/nix/profile` | + | `~/.nix-defexpr` | `$XDG_STATE_HOME/nix/defexpr` | + | `~/.nix-channels` | `$XDG_STATE_HOME/nix/channels` | + )" + }; }; @@ -916,4 +992,12 @@ std::vector<Path> getUserConfigFiles(); extern const std::string nixVersion; +/* NB: This is not sufficient. You need to call initNix() */ +void initLibStore(); + +/* It's important to initialize before doing _anything_, which is why we + call upon the programmer to handle this correctly. However, we only add + this in a key locations, so as not to litter the code. */ +void assertLibStoreInitialized(); + } diff --git a/src/libstore/http-binary-cache-store.cc b/src/libstore/http-binary-cache-store.cc index 73bcd6e81..1479822a9 100644 --- a/src/libstore/http-binary-cache-store.cc +++ b/src/libstore/http-binary-cache-store.cc @@ -56,7 +56,7 @@ public: void init() override { // FIXME: do this lazily? - if (auto cacheInfo = diskCache->cacheExists(cacheUri)) { + if (auto cacheInfo = diskCache->upToDateCacheExists(cacheUri)) { wantMassQuery.setDefault(cacheInfo->wantMassQuery); priority.setDefault(cacheInfo->priority); } else { diff --git a/src/libstore/legacy-ssh-store.cc b/src/libstore/legacy-ssh-store.cc index dd34b19c6..2c9dd2680 100644 --- a/src/libstore/legacy-ssh-store.cc +++ b/src/libstore/legacy-ssh-store.cc @@ -134,7 +134,6 @@ struct LegacySSHStore : public virtual LegacySSHStoreConfig, public virtual Stor /* Hash will be set below. FIXME construct ValidPathInfo at end. */ auto info = std::make_shared<ValidPathInfo>(path, Hash::dummy); - PathSet references; auto deriver = readString(conn->from); if (deriver != "") info->deriver = parseStorePath(deriver); @@ -255,8 +254,8 @@ private: << settings.maxLogSize; if (GET_PROTOCOL_MINOR(conn.remoteVersion) >= 3) conn.to - << settings.buildRepeat - << settings.enforceDeterminism; + << 0 // buildRepeat hasn't worked for ages anyway + << 0; if (GET_PROTOCOL_MINOR(conn.remoteVersion) >= 7) { conn.to << ((int) settings.keepFailed); @@ -279,7 +278,12 @@ public: conn->to.flush(); - BuildResult status { .path = DerivedPath::Built { .drvPath = drvPath } }; + BuildResult status { + .path = DerivedPath::Built { + .drvPath = drvPath, + .outputs = OutputsSpec::All { }, + }, + }; status.status = (BuildResult::Status) readInt(conn->from); conn->from >> status.errorMsg; diff --git a/src/libstore/local-binary-cache-store.cc b/src/libstore/local-binary-cache-store.cc index ba4416f6d..f20b1fa02 100644 --- a/src/libstore/local-binary-cache-store.cc +++ b/src/libstore/local-binary-cache-store.cc @@ -57,8 +57,7 @@ protected: AutoDelete del(tmp, false); StreamToSourceAdapter source(istream); writeFile(tmp, source); - if (rename(tmp.c_str(), path2.c_str())) - throw SysError("renaming '%1%' to '%2%'", tmp, path2); + renameFile(tmp, path2); del.cancel(); } diff --git a/src/libstore/local-fs-store.cc b/src/libstore/local-fs-store.cc index c5ae7536f..b224fc3e9 100644 --- a/src/libstore/local-fs-store.cc +++ b/src/libstore/local-fs-store.cc @@ -87,20 +87,8 @@ void LocalFSStore::narFromPath(const StorePath & path, Sink & sink) const std::string LocalFSStore::drvsLogDir = "drvs"; -std::optional<std::string> LocalFSStore::getBuildLog(const StorePath & path_) +std::optional<std::string> LocalFSStore::getBuildLogExact(const StorePath & path) { - auto path = path_; - - if (!path.isDerivation()) { - try { - auto info = queryPathInfo(path); - if (!info->deriver) return std::nullopt; - path = *info->deriver; - } catch (InvalidPath &) { - return std::nullopt; - } - } - auto baseName = path.to_string(); for (int j = 0; j < 2; j++) { diff --git a/src/libstore/local-fs-store.hh b/src/libstore/local-fs-store.hh index e6fb3201a..947707341 100644 --- a/src/libstore/local-fs-store.hh +++ b/src/libstore/local-fs-store.hh @@ -50,7 +50,7 @@ public: return getRealStoreDir() + "/" + std::string(storePath, storeDir.size() + 1); } - std::optional<std::string> getBuildLog(const StorePath & path) override; + std::optional<std::string> getBuildLogExact(const StorePath & path) override; }; diff --git a/src/libstore/local-store.cc b/src/libstore/local-store.cc index eba3b0fa5..c9a466ee8 100644 --- a/src/libstore/local-store.cc +++ b/src/libstore/local-store.cc @@ -91,6 +91,7 @@ void migrateCASchema(SQLite& db, Path schemaPath, AutoCloseFD& lockFd) if (!lockFile(lockFd.get(), ltWrite, false)) { printInfo("waiting for exclusive access to the Nix store for ca drvs..."); + lockFile(lockFd.get(), ltNone, false); // We have acquired a shared lock; release it to prevent deadlocks lockFile(lockFd.get(), ltWrite, true); } @@ -158,7 +159,7 @@ void migrateCASchema(SQLite& db, Path schemaPath, AutoCloseFD& lockFd) txn.commit(); } - writeFile(schemaPath, fmt("%d", nixCASchemaVersion)); + writeFile(schemaPath, fmt("%d", nixCASchemaVersion), 0666, true); lockFile(lockFd.get(), ltRead, true); } } @@ -200,8 +201,6 @@ LocalStore::LocalStore(const Params & params) throw SysError("could not set permissions on '%s' to 755", perUserDir); } - createUser(getUserName(), getuid()); - /* Optionally, create directories and set permissions for a multi-user install. */ if (getuid() == 0 && settings.buildUsersGroup != "") { @@ -281,7 +280,7 @@ LocalStore::LocalStore(const Params & params) else if (curSchema == 0) { /* new store */ curSchema = nixSchemaVersion; openDB(*state, true); - writeFile(schemaPath, (format("%1%") % nixSchemaVersion).str()); + writeFile(schemaPath, fmt("%1%", nixSchemaVersion), 0666, true); } else if (curSchema < nixSchemaVersion) { @@ -299,6 +298,7 @@ LocalStore::LocalStore(const Params & params) if (!lockFile(globalLock.get(), ltWrite, false)) { printInfo("waiting for exclusive access to the Nix store..."); + lockFile(globalLock.get(), ltNone, false); // We have acquired a shared lock; release it to prevent deadlocks lockFile(globalLock.get(), ltWrite, true); } @@ -329,7 +329,7 @@ LocalStore::LocalStore(const Params & params) txn.commit(); } - writeFile(schemaPath, (format("%1%") % nixSchemaVersion).str()); + writeFile(schemaPath, fmt("%1%", nixSchemaVersion), 0666, true); lockFile(globalLock.get(), ltRead, true); } @@ -439,9 +439,9 @@ LocalStore::~LocalStore() } try { - auto state(_state.lock()); - if (state->fdTempRoots) { - state->fdTempRoots = -1; + auto fdTempRoots(_fdTempRoots.lock()); + if (*fdTempRoots) { + *fdTempRoots = -1; unlink(fnTempRoots.c_str()); } } catch (...) { @@ -583,7 +583,10 @@ void canonicaliseTimestampAndPermissions(const Path & path) } -static void canonicalisePathMetaData_(const Path & path, uid_t fromUid, InodesSeen & inodesSeen) +static void canonicalisePathMetaData_( + const Path & path, + std::optional<std::pair<uid_t, uid_t>> uidRange, + InodesSeen & inodesSeen) { checkInterrupt(); @@ -630,7 +633,7 @@ static void canonicalisePathMetaData_(const Path & path, uid_t fromUid, InodesSe However, ignore files that we chown'ed ourselves previously to ensure that we don't fail on hard links within the same build (i.e. "touch $out/foo; ln $out/foo $out/bar"). */ - if (fromUid != (uid_t) -1 && st.st_uid != fromUid) { + if (uidRange && (st.st_uid < uidRange->first || st.st_uid > uidRange->second)) { if (S_ISDIR(st.st_mode) || !inodesSeen.count(Inode(st.st_dev, st.st_ino))) throw BuildError("invalid ownership on file '%1%'", path); mode_t mode = st.st_mode & ~S_IFMT; @@ -663,14 +666,17 @@ static void canonicalisePathMetaData_(const Path & path, uid_t fromUid, InodesSe if (S_ISDIR(st.st_mode)) { DirEntries entries = readDirectory(path); for (auto & i : entries) - canonicalisePathMetaData_(path + "/" + i.name, fromUid, inodesSeen); + canonicalisePathMetaData_(path + "/" + i.name, uidRange, inodesSeen); } } -void canonicalisePathMetaData(const Path & path, uid_t fromUid, InodesSeen & inodesSeen) +void canonicalisePathMetaData( + const Path & path, + std::optional<std::pair<uid_t, uid_t>> uidRange, + InodesSeen & inodesSeen) { - canonicalisePathMetaData_(path, fromUid, inodesSeen); + canonicalisePathMetaData_(path, uidRange, inodesSeen); /* On platforms that don't have lchown(), the top-level path can't be a symlink, since we can't change its ownership. */ @@ -683,10 +689,11 @@ void canonicalisePathMetaData(const Path & path, uid_t fromUid, InodesSeen & ino } -void canonicalisePathMetaData(const Path & path, uid_t fromUid) +void canonicalisePathMetaData(const Path & path, + std::optional<std::pair<uid_t, uid_t>> uidRange) { InodesSeen inodesSeen; - canonicalisePathMetaData(path, fromUid, inodesSeen); + canonicalisePathMetaData(path, uidRange, inodesSeen); } @@ -751,7 +758,7 @@ void LocalStore::registerDrvOutput(const Realisation & info, CheckSigsFlag check if (checkSigs == NoCheckSigs || !realisationIsUntrusted(info)) registerDrvOutput(info); else - throw Error("cannot register realisation '%s' because it lacks a valid signature", info.outPath.to_string()); + throw Error("cannot register realisation '%s' because it lacks a signature by a trusted key", info.outPath.to_string()); } void LocalStore::registerDrvOutput(const Realisation & info) @@ -1266,7 +1273,7 @@ void LocalStore::addToStore(const ValidPathInfo & info, Source & source, RepairFlag repair, CheckSigsFlag checkSigs) { if (checkSigs && pathInfoIsUntrusted(info)) - throw Error("cannot add path '%s' because it lacks a valid signature", printStorePath(info.path)); + throw Error("cannot add path '%s' because it lacks a signature by a trusted key", printStorePath(info.path)); addTempRoot(info.path); @@ -1331,7 +1338,7 @@ void LocalStore::addToStore(const ValidPathInfo & info, Source & source, autoGC(); - canonicalisePathMetaData(realPath, -1); + canonicalisePathMetaData(realPath, {}); optimisePath(realPath, repair); // FIXME: combine with hashPath() @@ -1382,13 +1389,15 @@ StorePath LocalStore::addToStoreFromDump(Source & source0, std::string_view name std::unique_ptr<AutoDelete> delTempDir; Path tempPath; + Path tempDir; + AutoCloseFD tempDirFd; if (!inMemory) { /* Drain what we pulled so far, and then keep on pulling */ StringSource dumpSource { dump }; ChainSource bothSource { dumpSource, source }; - auto tempDir = createTempDir(realStoreDir, "add"); + std::tie(tempDir, tempDirFd) = createTempDirInStore(); delTempDir = std::make_unique<AutoDelete>(tempDir); tempPath = tempDir + "/x"; @@ -1430,8 +1439,7 @@ StorePath LocalStore::addToStoreFromDump(Source & source0, std::string_view name writeFile(realPath, dumpSource); } else { /* Move the temporary path we restored above. */ - if (rename(tempPath.c_str(), realPath.c_str())) - throw Error("renaming '%s' to '%s'", tempPath, realPath); + moveFile(tempPath, realPath); } /* For computing the nar hash. In recursive SHA-256 mode, this @@ -1443,7 +1451,7 @@ StorePath LocalStore::addToStoreFromDump(Source & source0, std::string_view name narHash = narSink.finish(); } - canonicalisePathMetaData(realPath, -1); // FIXME: merge into restorePath + canonicalisePathMetaData(realPath, {}); // FIXME: merge into restorePath optimisePath(realPath, repair); @@ -1485,7 +1493,7 @@ StorePath LocalStore::addTextToStore( writeFile(realPath, s); - canonicalisePathMetaData(realPath, -1); + canonicalisePathMetaData(realPath, {}); StringSink sink; dumpString(s, sink); @@ -1508,18 +1516,24 @@ StorePath LocalStore::addTextToStore( /* Create a temporary directory in the store that won't be - garbage-collected. */ -Path LocalStore::createTempDirInStore() + garbage-collected until the returned FD is closed. */ +std::pair<Path, AutoCloseFD> LocalStore::createTempDirInStore() { - Path tmpDir; + Path tmpDirFn; + AutoCloseFD tmpDirFd; + bool lockedByUs = false; do { /* There is a slight possibility that `tmpDir' gets deleted by - the GC between createTempDir() and addTempRoot(), so repeat - until `tmpDir' exists. */ - tmpDir = createTempDir(realStoreDir); - addTempRoot(parseStorePath(tmpDir)); - } while (!pathExists(tmpDir)); - return tmpDir; + the GC between createTempDir() and when we acquire a lock on it. + We'll repeat until 'tmpDir' exists and we've locked it. */ + tmpDirFn = createTempDir(realStoreDir, "tmp"); + tmpDirFd = open(tmpDirFn.c_str(), O_RDONLY | O_DIRECTORY); + if (tmpDirFd.get() < 0) { + continue; + } + lockedByUs = lockFile(tmpDirFd.get(), ltWrite, true); + } while (!pathExists(tmpDirFn) || !lockedByUs); + return {tmpDirFn, std::move(tmpDirFd)}; } @@ -1546,7 +1560,7 @@ void LocalStore::invalidatePathChecked(const StorePath & path) bool LocalStore::verifyStore(bool checkContents, RepairFlag repair) { - printInfo(format("reading the Nix store...")); + printInfo("reading the Nix store..."); bool errors = false; @@ -1808,20 +1822,6 @@ void LocalStore::signPathInfo(ValidPathInfo & info) } -void LocalStore::createUser(const std::string & userName, uid_t userId) -{ - for (auto & dir : { - fmt("%s/profiles/per-user/%s", stateDir, userName), - fmt("%s/gcroots/per-user/%s", stateDir, userName) - }) { - createDirs(dir); - if (chmod(dir.c_str(), 0755) == -1) - throw SysError("changing permissions of directory '%s'", dir); - if (chown(dir.c_str(), userId, getgid()) == -1) - throw SysError("changing owner of directory '%s'", dir); - } -} - std::optional<std::pair<int64_t, Realisation>> LocalStore::queryRealisationCore_( LocalStore::State & state, const DrvOutput & id) @@ -1942,8 +1942,7 @@ void LocalStore::addBuildLog(const StorePath & drvPath, std::string_view log) writeFile(tmpFile, compress("bzip2", log)); - if (rename(tmpFile.c_str(), logPath.c_str()) != 0) - throw SysError("renaming '%1%' to '%2%'", tmpFile, logPath); + renameFile(tmpFile, logPath); } std::optional<std::string> LocalStore::getVersion() diff --git a/src/libstore/local-store.hh b/src/libstore/local-store.hh index 70d225be3..a84eb7c26 100644 --- a/src/libstore/local-store.hh +++ b/src/libstore/local-store.hh @@ -59,15 +59,6 @@ private: struct Stmts; std::unique_ptr<Stmts> stmts; - /* The global GC lock */ - AutoCloseFD fdGCLock; - - /* The file to which we write our temporary roots. */ - AutoCloseFD fdTempRoots; - - /* Connection to the garbage collector. */ - AutoCloseFD fdRootsSocket; - /* The last time we checked whether to do an auto-GC, or an auto-GC finished. */ std::chrono::time_point<std::chrono::steady_clock> lastGCCheck; @@ -156,6 +147,21 @@ public: void addTempRoot(const StorePath & path) override; +private: + + void createTempRootsFile(); + + /* The file to which we write our temporary roots. */ + Sync<AutoCloseFD> _fdTempRoots; + + /* The global GC lock. */ + Sync<AutoCloseFD> _fdGCLock; + + /* Connection to the garbage collector. */ + Sync<AutoCloseFD> _fdRootsSocket; + +public: + void addIndirectRoot(const Path & path) override; private: @@ -256,7 +262,7 @@ private: void findRuntimeRoots(Roots & roots, bool censor); - Path createTempDirInStore(); + std::pair<Path, AutoCloseFD> createTempDirInStore(); void checkDerivationOutputs(const StorePath & drvPath, const Derivation & drv); @@ -275,8 +281,6 @@ private: void signPathInfo(ValidPathInfo & info); void signRealisation(Realisation &); - void createUser(const std::string & userName, uid_t userId) override; - // XXX: Make a generic `Store` method FixedOutputHash hashCAPath( const FileIngestionMethod & method, @@ -310,9 +314,18 @@ typedef std::set<Inode> InodesSeen; - the permissions are set of 444 or 555 (i.e., read-only with or without execute permission; setuid bits etc. are cleared) - the owner and group are set to the Nix user and group, if we're - running as root. */ -void canonicalisePathMetaData(const Path & path, uid_t fromUid, InodesSeen & inodesSeen); -void canonicalisePathMetaData(const Path & path, uid_t fromUid); + running as root. + If uidRange is not empty, this function will throw an error if it + encounters files owned by a user outside of the closed interval + [uidRange->first, uidRange->second]. +*/ +void canonicalisePathMetaData( + const Path & path, + std::optional<std::pair<uid_t, uid_t>> uidRange, + InodesSeen & inodesSeen); +void canonicalisePathMetaData( + const Path & path, + std::optional<std::pair<uid_t, uid_t>> uidRange); void canonicaliseTimestampAndPermissions(const Path & path); diff --git a/src/libstore/local.mk b/src/libstore/local.mk index 1d26ac918..e5e24501e 100644 --- a/src/libstore/local.mk +++ b/src/libstore/local.mk @@ -13,14 +13,10 @@ ifdef HOST_LINUX libstore_LDFLAGS += -ldl endif -ifdef HOST_DARWIN -libstore_FILES = sandbox-defaults.sb sandbox-minimal.sb sandbox-network.sb -endif - $(foreach file,$(libstore_FILES),$(eval $(call install-data-in,$(d)/$(file),$(datadir)/nix/sandbox))) ifeq ($(ENABLE_S3), 1) - libstore_LDFLAGS += -laws-cpp-sdk-transfer -laws-cpp-sdk-s3 -laws-cpp-sdk-core + libstore_LDFLAGS += -laws-cpp-sdk-transfer -laws-cpp-sdk-s3 -laws-cpp-sdk-core -laws-crt-cpp endif ifdef HOST_SOLARIS diff --git a/src/libstore/lock.cc b/src/libstore/lock.cc index fa718f55d..4fe1fcf56 100644 --- a/src/libstore/lock.cc +++ b/src/libstore/lock.cc @@ -2,105 +2,201 @@ #include "globals.hh" #include "pathlocks.hh" -#include <grp.h> #include <pwd.h> - -#include <fcntl.h> -#include <unistd.h> +#include <grp.h> namespace nix { -UserLock::UserLock() +struct SimpleUserLock : UserLock { - assert(settings.buildUsersGroup != ""); - createDirs(settings.nixStateDir + "/userpool"); -} + AutoCloseFD fdUserLock; + uid_t uid; + gid_t gid; + std::vector<gid_t> supplementaryGIDs; + + uid_t getUID() override { assert(uid); return uid; } + uid_t getUIDCount() override { return 1; } + gid_t getGID() override { assert(gid); return gid; } + + std::vector<gid_t> getSupplementaryGIDs() override { return supplementaryGIDs; } + + static std::unique_ptr<UserLock> acquire() + { + assert(settings.buildUsersGroup != ""); + createDirs(settings.nixStateDir + "/userpool"); + + /* Get the members of the build-users-group. */ + struct group * gr = getgrnam(settings.buildUsersGroup.get().c_str()); + if (!gr) + throw Error("the group '%s' specified in 'build-users-group' does not exist", settings.buildUsersGroup); + + /* Copy the result of getgrnam. */ + Strings users; + for (char * * p = gr->gr_mem; *p; ++p) { + debug("found build user '%s'", *p); + users.push_back(*p); + } -bool UserLock::findFreeUser() { - if (enabled()) return true; - - /* Get the members of the build-users-group. */ - struct group * gr = getgrnam(settings.buildUsersGroup.get().c_str()); - if (!gr) - throw Error("the group '%1%' specified in 'build-users-group' does not exist", - settings.buildUsersGroup); - gid = gr->gr_gid; - - /* Copy the result of getgrnam. */ - Strings users; - for (char * * p = gr->gr_mem; *p; ++p) { - debug("found build user '%1%'", *p); - users.push_back(*p); + if (users.empty()) + throw Error("the build users group '%s' has no members", settings.buildUsersGroup); + + /* Find a user account that isn't currently in use for another + build. */ + for (auto & i : users) { + debug("trying user '%s'", i); + + struct passwd * pw = getpwnam(i.c_str()); + if (!pw) + throw Error("the user '%s' in the group '%s' does not exist", i, settings.buildUsersGroup); + + auto fnUserLock = fmt("%s/userpool/%s", settings.nixStateDir,pw->pw_uid); + + AutoCloseFD fd = open(fnUserLock.c_str(), O_RDWR | O_CREAT | O_CLOEXEC, 0600); + if (!fd) + throw SysError("opening user lock '%s'", fnUserLock); + + if (lockFile(fd.get(), ltWrite, false)) { + auto lock = std::make_unique<SimpleUserLock>(); + + lock->fdUserLock = std::move(fd); + lock->uid = pw->pw_uid; + lock->gid = gr->gr_gid; + + /* Sanity check... */ + if (lock->uid == getuid() || lock->uid == geteuid()) + throw Error("the Nix user should not be a member of '%s'", settings.buildUsersGroup); + + #if __linux__ + /* Get the list of supplementary groups of this build + user. This is usually either empty or contains a + group such as "kvm". */ + int ngroups = 32; // arbitrary initial guess + std::vector<gid_t> gids; + gids.resize(ngroups); + + int err = getgrouplist( + pw->pw_name, pw->pw_gid, + gids.data(), + &ngroups); + + /* Our initial size of 32 wasn't sufficient, the + correct size has been stored in ngroups, so we try + again. */ + if (err == -1) { + gids.resize(ngroups); + err = getgrouplist( + pw->pw_name, pw->pw_gid, + gids.data(), + &ngroups); + } + + // If it failed once more, then something must be broken. + if (err == -1) + throw Error("failed to get list of supplementary groups for '%s'", pw->pw_name); + + // Finally, trim back the GID list to its real size. + for (auto i = 0; i < ngroups; i++) + if (gids[i] != lock->gid) + lock->supplementaryGIDs.push_back(gids[i]); + #endif + + return lock; + } + } + + return nullptr; } +}; + +struct AutoUserLock : UserLock +{ + AutoCloseFD fdUserLock; + uid_t firstUid = 0; + gid_t firstGid = 0; + uid_t nrIds = 1; - if (users.empty()) - throw Error("the build users group '%1%' has no members", - settings.buildUsersGroup); + uid_t getUID() override { assert(firstUid); return firstUid; } - /* Find a user account that isn't currently in use for another - build. */ - for (auto & i : users) { - debug("trying user '%1%'", i); + gid_t getUIDCount() override { return nrIds; } - struct passwd * pw = getpwnam(i.c_str()); - if (!pw) - throw Error("the user '%1%' in the group '%2%' does not exist", - i, settings.buildUsersGroup); + gid_t getGID() override { assert(firstGid); return firstGid; } + std::vector<gid_t> getSupplementaryGIDs() override { return {}; } - fnUserLock = (format("%1%/userpool/%2%") % settings.nixStateDir % pw->pw_uid).str(); + static std::unique_ptr<UserLock> acquire(uid_t nrIds, bool useUserNamespace) + { + #if !defined(__linux__) + useUserNamespace = false; + #endif - AutoCloseFD fd = open(fnUserLock.c_str(), O_RDWR | O_CREAT | O_CLOEXEC, 0600); - if (!fd) - throw SysError("opening user lock '%1%'", fnUserLock); + settings.requireExperimentalFeature(Xp::AutoAllocateUids); + assert(settings.startId > 0); + assert(settings.uidCount % maxIdsPerBuild == 0); + assert((uint64_t) settings.startId + (uint64_t) settings.uidCount <= std::numeric_limits<uid_t>::max()); + assert(nrIds <= maxIdsPerBuild); - if (lockFile(fd.get(), ltWrite, false)) { - fdUserLock = std::move(fd); - user = i; - uid = pw->pw_uid; + createDirs(settings.nixStateDir + "/userpool2"); - /* Sanity check... */ - if (uid == getuid() || uid == geteuid()) - throw Error("the Nix user should not be a member of '%1%'", - settings.buildUsersGroup); + size_t nrSlots = settings.uidCount / maxIdsPerBuild; -#if __linux__ - /* Get the list of supplementary groups of this build user. This - is usually either empty or contains a group such as "kvm". */ - int ngroups = 32; // arbitrary initial guess - supplementaryGIDs.resize(ngroups); + for (size_t i = 0; i < nrSlots; i++) { + debug("trying user slot '%d'", i); - int err = getgrouplist(pw->pw_name, pw->pw_gid, supplementaryGIDs.data(), - &ngroups); + createDirs(settings.nixStateDir + "/userpool2"); - // Our initial size of 32 wasn't sufficient, the correct size has - // been stored in ngroups, so we try again. - if (err == -1) { - supplementaryGIDs.resize(ngroups); - err = getgrouplist(pw->pw_name, pw->pw_gid, supplementaryGIDs.data(), - &ngroups); - } + auto fnUserLock = fmt("%s/userpool2/slot-%d", settings.nixStateDir, i); + + AutoCloseFD fd = open(fnUserLock.c_str(), O_RDWR | O_CREAT | O_CLOEXEC, 0600); + if (!fd) + throw SysError("opening user lock '%s'", fnUserLock); - // If it failed once more, then something must be broken. - if (err == -1) - throw Error("failed to get list of supplementary groups for '%1%'", - pw->pw_name); + if (lockFile(fd.get(), ltWrite, false)) { - // Finally, trim back the GID list to its real size - supplementaryGIDs.resize(ngroups); -#endif + auto firstUid = settings.startId + i * maxIdsPerBuild; - isEnabled = true; - return true; + auto pw = getpwuid(firstUid); + if (pw) + throw Error("auto-allocated UID %d clashes with existing user account '%s'", firstUid, pw->pw_name); + + auto lock = std::make_unique<AutoUserLock>(); + lock->fdUserLock = std::move(fd); + lock->firstUid = firstUid; + if (useUserNamespace) + lock->firstGid = firstUid; + else { + struct group * gr = getgrnam(settings.buildUsersGroup.get().c_str()); + if (!gr) + throw Error("the group '%s' specified in 'build-users-group' does not exist", settings.buildUsersGroup); + lock->firstGid = gr->gr_gid; + } + lock->nrIds = nrIds; + return lock; + } } + + return nullptr; } +}; - return false; +std::unique_ptr<UserLock> acquireUserLock(uid_t nrIds, bool useUserNamespace) +{ + if (settings.autoAllocateUids) + return AutoUserLock::acquire(nrIds, useUserNamespace); + else + return SimpleUserLock::acquire(); } -void UserLock::kill() +bool useBuildUsers() { - killUser(uid); + #if __linux__ + static bool b = (settings.buildUsersGroup != "" || settings.autoAllocateUids) && getuid() == 0; + return b; + #elif __APPLE__ + static bool b = settings.buildUsersGroup != "" && getuid() == 0; + return b; + #else + return false; + #endif } } diff --git a/src/libstore/lock.hh b/src/libstore/lock.hh index 3d29a7b5b..7f1934510 100644 --- a/src/libstore/lock.hh +++ b/src/libstore/lock.hh @@ -1,37 +1,38 @@ #pragma once -#include "sync.hh" #include "types.hh" -#include "util.hh" + +#include <optional> + +#include <sys/types.h> namespace nix { -class UserLock +struct UserLock { -private: - Path fnUserLock; - AutoCloseFD fdUserLock; + virtual ~UserLock() { } - bool isEnabled = false; - std::string user; - uid_t uid = 0; - gid_t gid = 0; - std::vector<gid_t> supplementaryGIDs; + /* Get the first and last UID. */ + std::pair<uid_t, uid_t> getUIDRange() + { + auto first = getUID(); + return {first, first + getUIDCount() - 1}; + } -public: - UserLock(); + /* Get the first UID. */ + virtual uid_t getUID() = 0; - void kill(); + virtual uid_t getUIDCount() = 0; - std::string getUser() { return user; } - uid_t getUID() { assert(uid); return uid; } - uid_t getGID() { assert(gid); return gid; } - std::vector<gid_t> getSupplementaryGIDs() { return supplementaryGIDs; } + virtual gid_t getGID() = 0; - bool findFreeUser(); + virtual std::vector<gid_t> getSupplementaryGIDs() = 0; +}; - bool enabled() { return isEnabled; } +/* Acquire a user lock for a UID range of size `nrIds`. Note that this + may return nullptr if no user is available. */ +std::unique_ptr<UserLock> acquireUserLock(uid_t nrIds, bool useUserNamespace); -}; +bool useBuildUsers(); } diff --git a/src/libstore/log-store.cc b/src/libstore/log-store.cc new file mode 100644 index 000000000..8a26832ab --- /dev/null +++ b/src/libstore/log-store.cc @@ -0,0 +1,12 @@ +#include "log-store.hh" + +namespace nix { + +std::optional<std::string> LogStore::getBuildLog(const StorePath & path) { + auto maybePath = getBuildDerivationPath(path); + if (!maybePath) + return std::nullopt; + return getBuildLogExact(maybePath.value()); +} + +} diff --git a/src/libstore/log-store.hh b/src/libstore/log-store.hh index ff1b92e17..e4d95bab6 100644 --- a/src/libstore/log-store.hh +++ b/src/libstore/log-store.hh @@ -11,7 +11,9 @@ struct LogStore : public virtual Store /* Return the build log of the specified store path, if available, or null otherwise. */ - virtual std::optional<std::string> getBuildLog(const StorePath & path) = 0; + std::optional<std::string> getBuildLog(const StorePath & path); + + virtual std::optional<std::string> getBuildLogExact(const StorePath & path) = 0; virtual void addBuildLog(const StorePath & path, std::string_view log) = 0; diff --git a/src/libstore/misc.cc b/src/libstore/misc.cc index fb985c97b..b28768459 100644 --- a/src/libstore/misc.cc +++ b/src/libstore/misc.cc @@ -185,7 +185,7 @@ void Store::queryMissing(const std::vector<DerivedPath> & targets, knownOutputPaths = false; break; } - if (wantOutput(outputName, bfd.outputs) && !isValidPath(*pathOpt)) + if (bfd.outputs.contains(outputName) && !isValidPath(*pathOpt)) invalid.insert(*pathOpt); } if (knownOutputPaths && invalid.empty()) return; @@ -301,4 +301,47 @@ std::map<DrvOutput, StorePath> drvOutputReferences( return drvOutputReferences(Realisation::closure(store, inputRealisations), info->references); } +OutputPathMap resolveDerivedPath(Store & store, const DerivedPath::Built & bfd, Store * evalStore_) +{ + auto & evalStore = evalStore_ ? *evalStore_ : store; + + OutputPathMap outputs; + auto drv = evalStore.readDerivation(bfd.drvPath); + auto outputHashes = staticOutputHashes(store, drv); + auto drvOutputs = drv.outputsAndOptPaths(store); + auto outputNames = std::visit(overloaded { + [&](const OutputsSpec::All &) { + StringSet names; + for (auto & [outputName, _] : drv.outputs) + names.insert(outputName); + return names; + }, + [&](const OutputsSpec::Names & names) { + return static_cast<std::set<std::string>>(names); + }, + }, bfd.outputs.raw()); + for (auto & output : outputNames) { + auto outputHash = get(outputHashes, output); + if (!outputHash) + throw Error( + "the derivation '%s' doesn't have an output named '%s'", + store.printStorePath(bfd.drvPath), output); + if (settings.isExperimentalFeatureEnabled(Xp::CaDerivations)) { + DrvOutput outputId { *outputHash, output }; + auto realisation = store.queryRealisation(outputId); + if (!realisation) + throw MissingRealisation(outputId); + outputs.insert_or_assign(output, realisation->outPath); + } else { + // If ca-derivations isn't enabled, assume that + // the output path is statically known. + auto drvOutput = get(drvOutputs, output); + assert(drvOutput); + assert(drvOutput->second); + outputs.insert_or_assign(output, *drvOutput->second); + } + } + return outputs; +} + } diff --git a/src/libstore/nar-accessor.cc b/src/libstore/nar-accessor.cc index 72d41cc94..9a0003588 100644 --- a/src/libstore/nar-accessor.cc +++ b/src/libstore/nar-accessor.cc @@ -1,6 +1,5 @@ #include "nar-accessor.hh" #include "archive.hh" -#include "json.hh" #include <map> #include <stack> @@ -75,6 +74,9 @@ struct NarAccessor : public FSAccessor createMember(path, {FSAccessor::Type::tRegular, false, 0, 0}); } + void closeRegularFile() override + { } + void isExecutable() override { parents.top()->isExecutable = true; @@ -240,42 +242,43 @@ ref<FSAccessor> makeLazyNarAccessor(const std::string & listing, return make_ref<NarAccessor>(listing, getNarBytes); } -void listNar(JSONPlaceholder & res, ref<FSAccessor> accessor, - const Path & path, bool recurse) +using nlohmann::json; +json listNar(ref<FSAccessor> accessor, const Path & path, bool recurse) { auto st = accessor->stat(path); - auto obj = res.object(); + json obj = json::object(); switch (st.type) { case FSAccessor::Type::tRegular: - obj.attr("type", "regular"); - obj.attr("size", st.fileSize); + obj["type"] = "regular"; + obj["size"] = st.fileSize; if (st.isExecutable) - obj.attr("executable", true); + obj["executable"] = true; if (st.narOffset) - obj.attr("narOffset", st.narOffset); + obj["narOffset"] = st.narOffset; break; case FSAccessor::Type::tDirectory: - obj.attr("type", "directory"); + obj["type"] = "directory"; { - auto res2 = obj.object("entries"); + obj["entries"] = json::object(); + json &res2 = obj["entries"]; for (auto & name : accessor->readDirectory(path)) { if (recurse) { - auto res3 = res2.placeholder(name); - listNar(res3, accessor, path + "/" + name, true); + res2[name] = listNar(accessor, path + "/" + name, true); } else - res2.object(name); + res2[name] = json::object(); } } break; case FSAccessor::Type::tSymlink: - obj.attr("type", "symlink"); - obj.attr("target", accessor->readLink(path)); + obj["type"] = "symlink"; + obj["target"] = accessor->readLink(path); break; default: throw Error("path '%s' does not exist in NAR", path); } + return obj; } } diff --git a/src/libstore/nar-accessor.hh b/src/libstore/nar-accessor.hh index c2241a04c..7d998ae0b 100644 --- a/src/libstore/nar-accessor.hh +++ b/src/libstore/nar-accessor.hh @@ -2,6 +2,7 @@ #include <functional> +#include <nlohmann/json_fwd.hpp> #include "fs-accessor.hh" namespace nix { @@ -24,11 +25,8 @@ ref<FSAccessor> makeLazyNarAccessor( const std::string & listing, GetNarBytes getNarBytes); -class JSONPlaceholder; - /* Write a JSON representation of the contents of a NAR (except file contents). */ -void listNar(JSONPlaceholder & res, ref<FSAccessor> accessor, - const Path & path, bool recurse); +nlohmann::json listNar(ref<FSAccessor> accessor, const Path & path, bool recurse); } diff --git a/src/libstore/nar-info-disk-cache.cc b/src/libstore/nar-info-disk-cache.cc index f4ea739b0..2645f468b 100644 --- a/src/libstore/nar-info-disk-cache.cc +++ b/src/libstore/nar-info-disk-cache.cc @@ -84,11 +84,10 @@ public: Sync<State> _state; - NarInfoDiskCacheImpl() + NarInfoDiskCacheImpl(Path dbPath = getCacheDir() + "/nix/binary-cache-v6.sqlite") { auto state(_state.lock()); - Path dbPath = getCacheDir() + "/nix/binary-cache-v6.sqlite"; createDirs(dirOf(dbPath)); state->db = SQLite(dbPath); @@ -98,7 +97,7 @@ public: state->db.exec(schema); state->insertCache.create(state->db, - "insert or replace into BinaryCaches(url, timestamp, storeDir, wantMassQuery, priority) values (?, ?, ?, ?, ?)"); + "insert into BinaryCaches(url, timestamp, storeDir, wantMassQuery, priority) values (?1, ?2, ?3, ?4, ?5) on conflict (url) do update set timestamp = ?2, storeDir = ?3, wantMassQuery = ?4, priority = ?5 returning id;"); state->queryCache.create(state->db, "select id, storeDir, wantMassQuery, priority from BinaryCaches where url = ? and timestamp > ?"); @@ -166,38 +165,71 @@ public: return i->second; } - void createCache(const std::string & uri, const Path & storeDir, bool wantMassQuery, int priority) override +private: + + std::optional<Cache> queryCacheRaw(State & state, const std::string & uri) { - retrySQLite<void>([&]() { + auto i = state.caches.find(uri); + if (i == state.caches.end()) { + auto queryCache(state.queryCache.use()(uri)(time(0) - cacheInfoTtl)); + if (!queryCache.next()) + return std::nullopt; + auto cache = Cache { + .id = (int) queryCache.getInt(0), + .storeDir = queryCache.getStr(1), + .wantMassQuery = queryCache.getInt(2) != 0, + .priority = (int) queryCache.getInt(3), + }; + state.caches.emplace(uri, cache); + } + return getCache(state, uri); + } + +public: + int createCache(const std::string & uri, const Path & storeDir, bool wantMassQuery, int priority) override + { + return retrySQLite<int>([&]() { auto state(_state.lock()); + SQLiteTxn txn(state->db); + + // To avoid the race, we have to check if maybe someone hasn't yet created + // the cache for this URI in the meantime. + auto cache(queryCacheRaw(*state, uri)); + + if (cache) + return cache->id; + + Cache ret { + .id = -1, // set below + .storeDir = storeDir, + .wantMassQuery = wantMassQuery, + .priority = priority, + }; + + { + auto r(state->insertCache.use()(uri)(time(0))(storeDir)(wantMassQuery)(priority)); + assert(r.next()); + ret.id = (int) r.getInt(0); + } - // FIXME: race + state->caches[uri] = ret; - state->insertCache.use()(uri)(time(0))(storeDir)(wantMassQuery)(priority).exec(); - assert(sqlite3_changes(state->db) == 1); - state->caches[uri] = Cache{(int) sqlite3_last_insert_rowid(state->db), storeDir, wantMassQuery, priority}; + txn.commit(); + return ret.id; }); } - std::optional<CacheInfo> cacheExists(const std::string & uri) override + std::optional<CacheInfo> upToDateCacheExists(const std::string & uri) override { return retrySQLite<std::optional<CacheInfo>>([&]() -> std::optional<CacheInfo> { auto state(_state.lock()); - - auto i = state->caches.find(uri); - if (i == state->caches.end()) { - auto queryCache(state->queryCache.use()(uri)(time(0) - cacheInfoTtl)); - if (!queryCache.next()) - return std::nullopt; - state->caches.emplace(uri, - Cache{(int) queryCache.getInt(0), queryCache.getStr(1), queryCache.getInt(2) != 0, (int) queryCache.getInt(3)}); - } - - auto & cache(getCache(*state, uri)); - + auto cache(queryCacheRaw(*state, uri)); + if (!cache) + return std::nullopt; return CacheInfo { - .wantMassQuery = cache.wantMassQuery, - .priority = cache.priority + .id = cache->id, + .wantMassQuery = cache->wantMassQuery, + .priority = cache->priority }; }); } @@ -359,4 +391,9 @@ ref<NarInfoDiskCache> getNarInfoDiskCache() return cache; } +ref<NarInfoDiskCache> getTestNarInfoDiskCache(Path dbPath) +{ + return make_ref<NarInfoDiskCacheImpl>(dbPath); +} + } diff --git a/src/libstore/nar-info-disk-cache.hh b/src/libstore/nar-info-disk-cache.hh index 2dcaa76a4..4877f56d8 100644 --- a/src/libstore/nar-info-disk-cache.hh +++ b/src/libstore/nar-info-disk-cache.hh @@ -13,16 +13,17 @@ public: virtual ~NarInfoDiskCache() { } - virtual void createCache(const std::string & uri, const Path & storeDir, + virtual int createCache(const std::string & uri, const Path & storeDir, bool wantMassQuery, int priority) = 0; struct CacheInfo { + int id; bool wantMassQuery; int priority; }; - virtual std::optional<CacheInfo> cacheExists(const std::string & uri) = 0; + virtual std::optional<CacheInfo> upToDateCacheExists(const std::string & uri) = 0; virtual std::pair<Outcome, std::shared_ptr<NarInfo>> lookupNarInfo( const std::string & uri, const std::string & hashPart) = 0; @@ -45,4 +46,6 @@ public: multiple threads. */ ref<NarInfoDiskCache> getNarInfoDiskCache(); +ref<NarInfoDiskCache> getTestNarInfoDiskCache(Path dbPath); + } diff --git a/src/libstore/nix-store.pc.in b/src/libstore/nix-store.pc.in index 6d67b1e03..dc42d0bca 100644 --- a/src/libstore/nix-store.pc.in +++ b/src/libstore/nix-store.pc.in @@ -6,4 +6,4 @@ Name: Nix Description: Nix Package Manager Version: @PACKAGE_VERSION@ Libs: -L${libdir} -lnixstore -lnixutil -Cflags: -I${includedir}/nix -std=c++17 +Cflags: -I${includedir}/nix -std=c++2a diff --git a/src/libstore/optimise-store.cc b/src/libstore/optimise-store.cc index 8af9b1dde..4a79cf4a1 100644 --- a/src/libstore/optimise-store.cc +++ b/src/libstore/optimise-store.cc @@ -55,7 +55,7 @@ LocalStore::InodeHash LocalStore::loadInodeHash() } if (errno) throw SysError("reading directory '%1%'", linksDir); - printMsg(lvlTalkative, format("loaded %1% hash inodes") % inodeHash.size()); + printMsg(lvlTalkative, "loaded %1% hash inodes", inodeHash.size()); return inodeHash; } @@ -73,7 +73,7 @@ Strings LocalStore::readDirectoryIgnoringInodes(const Path & path, const InodeHa checkInterrupt(); if (inodeHash.count(dirent->d_ino)) { - debug(format("'%1%' is already linked") % dirent->d_name); + debug("'%1%' is already linked", dirent->d_name); continue; } @@ -102,7 +102,7 @@ void LocalStore::optimisePath_(Activity * act, OptimiseStats & stats, if (std::regex_search(path, std::regex("\\.app/Contents/.+$"))) { - debug(format("'%1%' is not allowed to be linked in macOS") % path); + debug("'%1%' is not allowed to be linked in macOS", path); return; } #endif @@ -146,7 +146,7 @@ void LocalStore::optimisePath_(Activity * act, OptimiseStats & stats, contents of the symlink (i.e. the result of readlink()), not the contents of the target (which may not even exist). */ Hash hash = hashPath(htSHA256, path).first; - debug(format("'%1%' has hash '%2%'") % path % hash.to_string(Base32, true)); + debug("'%1%' has hash '%2%'", path, hash.to_string(Base32, true)); /* Check if this is a known hash. */ Path linkPath = linksDir + "/" + hash.to_string(Base32, false); @@ -196,11 +196,11 @@ void LocalStore::optimisePath_(Activity * act, OptimiseStats & stats, auto stLink = lstat(linkPath); if (st.st_ino == stLink.st_ino) { - debug(format("'%1%' is already linked to '%2%'") % path % linkPath); + debug("'%1%' is already linked to '%2%'", path, linkPath); return; } - printMsg(lvlTalkative, format("linking '%1%' to '%2%'") % path % linkPath); + printMsg(lvlTalkative, "linking '%1%' to '%2%'", path, linkPath); /* Make the containing directory writable, but only if it's not the store itself (we don't want or need to mess with its @@ -213,8 +213,7 @@ void LocalStore::optimisePath_(Activity * act, OptimiseStats & stats, its timestamp back to 0. */ MakeReadOnly makeReadOnly(mustToggle ? dirOfPath : ""); - Path tempLink = (format("%1%/.tmp-link-%2%-%3%") - % realStoreDir % getpid() % random()).str(); + Path tempLink = fmt("%1%/.tmp-link-%2%-%3%", realStoreDir, getpid(), random()); if (link(linkPath.c_str(), tempLink.c_str()) == -1) { if (errno == EMLINK) { @@ -222,14 +221,16 @@ void LocalStore::optimisePath_(Activity * act, OptimiseStats & stats, systems). This is likely to happen with empty files. Just shrug and ignore. */ if (st.st_size) - printInfo(format("'%1%' has maximum number of links") % linkPath); + printInfo("'%1%' has maximum number of links", linkPath); return; } throw SysError("cannot link '%1%' to '%2%'", tempLink, linkPath); } /* Atomically replace the old file with the new hard link. */ - if (rename(tempLink.c_str(), path.c_str()) == -1) { + try { + renameFile(tempLink, path); + } catch (SysError & e) { if (unlink(tempLink.c_str()) == -1) printError("unable to unlink '%1%'", tempLink); if (errno == EMLINK) { @@ -240,7 +241,7 @@ void LocalStore::optimisePath_(Activity * act, OptimiseStats & stats, debug("'%s' has reached maximum number of links", linkPath); return; } - throw SysError("cannot rename '%1%' to '%2%'", tempLink, path); + throw; } stats.filesLinked++; diff --git a/src/libstore/outputs-spec.cc b/src/libstore/outputs-spec.cc new file mode 100644 index 000000000..e26c38138 --- /dev/null +++ b/src/libstore/outputs-spec.cc @@ -0,0 +1,195 @@ +#include <regex> +#include <nlohmann/json.hpp> + +#include "util.hh" +#include "regex-combinators.hh" +#include "outputs-spec.hh" +#include "path-regex.hh" + +namespace nix { + +bool OutputsSpec::contains(const std::string & outputName) const +{ + return std::visit(overloaded { + [&](const OutputsSpec::All &) { + return true; + }, + [&](const OutputsSpec::Names & outputNames) { + return outputNames.count(outputName) > 0; + }, + }, raw()); +} + +static std::string outputSpecRegexStr = + regex::either( + regex::group(R"(\*)"), + regex::group(regex::list(nameRegexStr))); + +std::optional<OutputsSpec> OutputsSpec::parseOpt(std::string_view s) +{ + static std::regex regex(std::string { outputSpecRegexStr }); + + std::smatch match; + std::string s2 { s }; // until some improves std::regex + if (!std::regex_match(s2, match, regex)) + return std::nullopt; + + if (match[1].matched) + return { OutputsSpec::All {} }; + + if (match[2].matched) + return OutputsSpec::Names { tokenizeString<StringSet>(match[2].str(), ",") }; + + assert(false); +} + + +OutputsSpec OutputsSpec::parse(std::string_view s) +{ + std::optional spec = parseOpt(s); + if (!spec) + throw Error("invalid outputs specifier '%s'", s); + return *spec; +} + + +std::optional<std::pair<std::string_view, ExtendedOutputsSpec>> ExtendedOutputsSpec::parseOpt(std::string_view s) +{ + auto found = s.rfind('^'); + + if (found == std::string::npos) + return std::pair { s, ExtendedOutputsSpec::Default {} }; + + auto specOpt = OutputsSpec::parseOpt(s.substr(found + 1)); + if (!specOpt) + return std::nullopt; + return std::pair { s.substr(0, found), ExtendedOutputsSpec::Explicit { *std::move(specOpt) } }; +} + + +std::pair<std::string_view, ExtendedOutputsSpec> ExtendedOutputsSpec::parse(std::string_view s) +{ + std::optional spec = parseOpt(s); + if (!spec) + throw Error("invalid extended outputs specifier '%s'", s); + return *spec; +} + + +std::string OutputsSpec::to_string() const +{ + return std::visit(overloaded { + [&](const OutputsSpec::All &) -> std::string { + return "*"; + }, + [&](const OutputsSpec::Names & outputNames) -> std::string { + return concatStringsSep(",", outputNames); + }, + }, raw()); +} + + +std::string ExtendedOutputsSpec::to_string() const +{ + return std::visit(overloaded { + [&](const ExtendedOutputsSpec::Default &) -> std::string { + return ""; + }, + [&](const ExtendedOutputsSpec::Explicit & outputSpec) -> std::string { + return "^" + outputSpec.to_string(); + }, + }, raw()); +} + + +OutputsSpec OutputsSpec::union_(const OutputsSpec & that) const +{ + return std::visit(overloaded { + [&](const OutputsSpec::All &) -> OutputsSpec { + return OutputsSpec::All { }; + }, + [&](const OutputsSpec::Names & theseNames) -> OutputsSpec { + return std::visit(overloaded { + [&](const OutputsSpec::All &) -> OutputsSpec { + return OutputsSpec::All {}; + }, + [&](const OutputsSpec::Names & thoseNames) -> OutputsSpec { + OutputsSpec::Names ret = theseNames; + ret.insert(thoseNames.begin(), thoseNames.end()); + return ret; + }, + }, that.raw()); + }, + }, raw()); +} + + +bool OutputsSpec::isSubsetOf(const OutputsSpec & that) const +{ + return std::visit(overloaded { + [&](const OutputsSpec::All &) { + return true; + }, + [&](const OutputsSpec::Names & thoseNames) { + return std::visit(overloaded { + [&](const OutputsSpec::All &) { + return false; + }, + [&](const OutputsSpec::Names & theseNames) { + bool ret = true; + for (auto & o : theseNames) + if (thoseNames.count(o) == 0) + ret = false; + return ret; + }, + }, raw()); + }, + }, that.raw()); +} + +} + +namespace nlohmann { + +using namespace nix; + +OutputsSpec adl_serializer<OutputsSpec>::from_json(const json & json) { + auto names = json.get<StringSet>(); + if (names == StringSet({"*"})) + return OutputsSpec::All {}; + else + return OutputsSpec::Names { std::move(names) }; +} + +void adl_serializer<OutputsSpec>::to_json(json & json, OutputsSpec t) { + std::visit(overloaded { + [&](const OutputsSpec::All &) { + json = std::vector<std::string>({"*"}); + }, + [&](const OutputsSpec::Names & names) { + json = names; + }, + }, t.raw()); +} + + +ExtendedOutputsSpec adl_serializer<ExtendedOutputsSpec>::from_json(const json & json) { + if (json.is_null()) + return ExtendedOutputsSpec::Default {}; + else { + return ExtendedOutputsSpec::Explicit { json.get<OutputsSpec>() }; + } +} + +void adl_serializer<ExtendedOutputsSpec>::to_json(json & json, ExtendedOutputsSpec t) { + std::visit(overloaded { + [&](const ExtendedOutputsSpec::Default &) { + json = nullptr; + }, + [&](const ExtendedOutputsSpec::Explicit & e) { + adl_serializer<OutputsSpec>::to_json(json, e); + }, + }, t.raw()); +} + +} diff --git a/src/libstore/outputs-spec.hh b/src/libstore/outputs-spec.hh new file mode 100644 index 000000000..46bc35ebc --- /dev/null +++ b/src/libstore/outputs-spec.hh @@ -0,0 +1,95 @@ +#pragma once + +#include <cassert> +#include <optional> +#include <set> +#include <variant> + +#include "json-impls.hh" + +namespace nix { + +struct OutputNames : std::set<std::string> { + using std::set<std::string>::set; + + /* These need to be "inherited manually" */ + + OutputNames(const std::set<std::string> & s) + : std::set<std::string>(s) + { assert(!empty()); } + + OutputNames(std::set<std::string> && s) + : std::set<std::string>(s) + { assert(!empty()); } + + /* This set should always be non-empty, so we delete this + constructor in order make creating empty ones by mistake harder. + */ + OutputNames() = delete; +}; + +struct AllOutputs : std::monostate { }; + +typedef std::variant<AllOutputs, OutputNames> _OutputsSpecRaw; + +struct OutputsSpec : _OutputsSpecRaw { + using Raw = _OutputsSpecRaw; + using Raw::Raw; + + /* Force choosing a variant */ + OutputsSpec() = delete; + + using Names = OutputNames; + using All = AllOutputs; + + inline const Raw & raw() const { + return static_cast<const Raw &>(*this); + } + + inline Raw & raw() { + return static_cast<Raw &>(*this); + } + + bool contains(const std::string & output) const; + + /* Create a new OutputsSpec which is the union of this and that. */ + OutputsSpec union_(const OutputsSpec & that) const; + + /* Whether this OutputsSpec is a subset of that. */ + bool isSubsetOf(const OutputsSpec & outputs) const; + + /* Parse a string of the form 'output1,...outputN' or + '*', returning the outputs spec. */ + static OutputsSpec parse(std::string_view s); + static std::optional<OutputsSpec> parseOpt(std::string_view s); + + std::string to_string() const; +}; + +struct DefaultOutputs : std::monostate { }; + +typedef std::variant<DefaultOutputs, OutputsSpec> _ExtendedOutputsSpecRaw; + +struct ExtendedOutputsSpec : _ExtendedOutputsSpecRaw { + using Raw = _ExtendedOutputsSpecRaw; + using Raw::Raw; + + using Default = DefaultOutputs; + using Explicit = OutputsSpec; + + inline const Raw & raw() const { + return static_cast<const Raw &>(*this); + } + + /* Parse a string of the form 'prefix^output1,...outputN' or + 'prefix^*', returning the prefix and the extended outputs spec. */ + static std::pair<std::string_view, ExtendedOutputsSpec> parse(std::string_view s); + static std::optional<std::pair<std::string_view, ExtendedOutputsSpec>> parseOpt(std::string_view s); + + std::string to_string() const; +}; + +} + +JSON_IMPL(OutputsSpec) +JSON_IMPL(ExtendedOutputsSpec) diff --git a/src/libstore/parsed-derivations.cc b/src/libstore/parsed-derivations.cc index f2288a04e..cc4a94fab 100644 --- a/src/libstore/parsed-derivations.cc +++ b/src/libstore/parsed-derivations.cc @@ -2,7 +2,6 @@ #include <nlohmann/json.hpp> #include <regex> -#include "json.hh" namespace nix { @@ -90,6 +89,7 @@ std::optional<Strings> ParsedDerivation::getStringsAttr(const std::string & name StringSet ParsedDerivation::getRequiredSystemFeatures() const { + // FIXME: cache this? StringSet res; for (auto & i : getStringsAttr("requiredSystemFeatures").value_or(Strings())) res.insert(i); @@ -125,6 +125,11 @@ bool ParsedDerivation::substitutesAllowed() const return getBoolAttr("allowSubstitutes", true); } +bool ParsedDerivation::useUidRange() const +{ + return getRequiredSystemFeatures().count("uid-range"); +} + static std::regex shVarName("[A-Za-z_][A-Za-z0-9_]*"); std::optional<nlohmann::json> ParsedDerivation::prepareStructuredAttrs(Store & store, const StorePathSet & inputPaths) @@ -144,16 +149,11 @@ std::optional<nlohmann::json> ParsedDerivation::prepareStructuredAttrs(Store & s auto e = json.find("exportReferencesGraph"); if (e != json.end() && e->is_object()) { for (auto i = e->begin(); i != e->end(); ++i) { - std::ostringstream str; - { - JSONPlaceholder jsonRoot(str, true); - StorePathSet storePaths; - for (auto & p : *i) - storePaths.insert(store.parseStorePath(p.get<std::string>())); - store.pathInfoToJSON(jsonRoot, - store.exportReferences(storePaths, inputPaths), false, true); - } - json[i.key()] = nlohmann::json::parse(str.str()); // urgh + StorePathSet storePaths; + for (auto & p : *i) + storePaths.insert(store.parseStorePath(p.get<std::string>())); + json[i.key()] = store.pathInfoToJSON( + store.exportReferences(storePaths, inputPaths), false, true); } } diff --git a/src/libstore/parsed-derivations.hh b/src/libstore/parsed-derivations.hh index 95bec21e8..bfb3857c0 100644 --- a/src/libstore/parsed-derivations.hh +++ b/src/libstore/parsed-derivations.hh @@ -38,6 +38,8 @@ public: bool substitutesAllowed() const; + bool useUidRange() const; + std::optional<nlohmann::json> prepareStructuredAttrs(Store & store, const StorePathSet & inputPaths); }; diff --git a/src/libstore/path-info.cc b/src/libstore/path-info.cc index fda55b2b6..bd55a9d06 100644 --- a/src/libstore/path-info.cc +++ b/src/libstore/path-info.cc @@ -3,6 +3,80 @@ namespace nix { +std::string ValidPathInfo::fingerprint(const Store & store) const +{ + if (narSize == 0) + throw Error("cannot calculate fingerprint of path '%s' because its size is not known", + store.printStorePath(path)); + return + "1;" + store.printStorePath(path) + ";" + + narHash.to_string(Base32, true) + ";" + + std::to_string(narSize) + ";" + + concatStringsSep(",", store.printStorePathSet(references)); +} + + +void ValidPathInfo::sign(const Store & store, const SecretKey & secretKey) +{ + sigs.insert(secretKey.signDetached(fingerprint(store))); +} + + +bool ValidPathInfo::isContentAddressed(const Store & store) const +{ + if (! ca) return false; + + auto caPath = std::visit(overloaded { + [&](const TextHash & th) { + return store.makeTextPath(path.name(), th.hash, references); + }, + [&](const FixedOutputHash & fsh) { + auto refs = references; + bool hasSelfReference = false; + if (refs.count(path)) { + hasSelfReference = true; + refs.erase(path); + } + return store.makeFixedOutputPath(fsh.method, fsh.hash, path.name(), refs, hasSelfReference); + } + }, *ca); + + bool res = caPath == path; + + if (!res) + printError("warning: path '%s' claims to be content-addressed but isn't", store.printStorePath(path)); + + return res; +} + + +size_t ValidPathInfo::checkSignatures(const Store & store, const PublicKeys & publicKeys) const +{ + if (isContentAddressed(store)) return maxSigs; + + size_t good = 0; + for (auto & sig : sigs) + if (checkSignature(store, publicKeys, sig)) + good++; + return good; +} + + +bool ValidPathInfo::checkSignature(const Store & store, const PublicKeys & publicKeys, const std::string & sig) const +{ + return verifyDetached(fingerprint(store), sig, publicKeys); +} + + +Strings ValidPathInfo::shortRefs() const +{ + Strings refs; + for (auto & r : references) + refs.push_back(std::string(r.to_string())); + return refs; +} + + ValidPathInfo ValidPathInfo::read(Source & source, const Store & store, unsigned int format) { return read(source, store, format, store.parseStorePath(readString(source))); @@ -24,6 +98,7 @@ ValidPathInfo ValidPathInfo::read(Source & source, const Store & store, unsigned return info; } + void ValidPathInfo::write( Sink & sink, const Store & store, diff --git a/src/libstore/path-info.hh b/src/libstore/path-info.hh index b4b54e593..a7fcbd232 100644 --- a/src/libstore/path-info.hh +++ b/src/libstore/path-info.hh @@ -81,12 +81,6 @@ struct ValidPathInfo /* Return true iff the path is verifiably content-addressed. */ bool isContentAddressed(const Store & store) const; - /* Functions to view references + hasSelfReference as one set, mainly for - compatibility's sake. */ - StorePathSet referencesPossiblyToSelf() const; - void insertReferencePossiblyToSelf(StorePath && ref); - void setReferencesPossiblyToSelf(StorePathSet && refs); - static const size_t maxSigs = std::numeric_limits<size_t>::max(); /* Return the number of signatures on this .narinfo that were diff --git a/src/libstore/path-regex.hh b/src/libstore/path-regex.hh new file mode 100644 index 000000000..6893c3876 --- /dev/null +++ b/src/libstore/path-regex.hh @@ -0,0 +1,7 @@ +#pragma once + +namespace nix { + +static constexpr std::string_view nameRegexStr = R"([0-9a-zA-Z\+\-\._\?=]+)"; + +} diff --git a/src/libstore/path-with-outputs.cc b/src/libstore/path-with-outputs.cc index d6d67ea05..869b490ad 100644 --- a/src/libstore/path-with-outputs.cc +++ b/src/libstore/path-with-outputs.cc @@ -1,6 +1,5 @@ #include "path-with-outputs.hh" #include "store-api.hh" -#include "nlohmann/json.hpp" #include <regex> @@ -16,10 +15,14 @@ std::string StorePathWithOutputs::to_string(const Store & store) const DerivedPath StorePathWithOutputs::toDerivedPath() const { - if (!outputs.empty() || path.isDerivation()) - return DerivedPath::Built { path, outputs }; - else + if (!outputs.empty()) { + return DerivedPath::Built { path, OutputsSpec::Names { outputs } }; + } else if (path.isDerivation()) { + assert(outputs.empty()); + return DerivedPath::Built { path, OutputsSpec::All { } }; + } else { return DerivedPath::Opaque { path }; + } } @@ -42,7 +45,18 @@ std::variant<StorePathWithOutputs, StorePath> StorePathWithOutputs::tryFromDeriv return StorePathWithOutputs { bo.path }; }, [&](const DerivedPath::Built & bfd) -> std::variant<StorePathWithOutputs, StorePath> { - return StorePathWithOutputs { bfd.drvPath, bfd.outputs }; + return StorePathWithOutputs { + .path = bfd.drvPath, + // Use legacy encoding of wildcard as empty set + .outputs = std::visit(overloaded { + [&](const OutputsSpec::All &) -> StringSet { + return {}; + }, + [&](const OutputsSpec::Names & outputs) { + return static_cast<StringSet>(outputs); + }, + }, bfd.outputs.raw()), + }; }, }, p.raw()); } @@ -53,8 +67,8 @@ std::pair<std::string_view, StringSet> parsePathWithOutputs(std::string_view s) size_t n = s.find("!"); return n == s.npos ? std::make_pair(s, std::set<std::string>()) - : std::make_pair(((std::string_view) s).substr(0, n), - tokenizeString<std::set<std::string>>(((std::string_view) s).substr(n + 1), ",")); + : std::make_pair(s.substr(0, n), + tokenizeString<std::set<std::string>>(s.substr(n + 1), ",")); } @@ -71,57 +85,4 @@ StorePathWithOutputs followLinksToStorePathWithOutputs(const Store & store, std: return StorePathWithOutputs { store.followLinksToStorePath(path), std::move(outputs) }; } -std::pair<std::string, OutputsSpec> parseOutputsSpec(const std::string & s) -{ - static std::regex regex(R"((.*)\^((\*)|([a-z]+(,[a-z]+)*)))"); - - std::smatch match; - if (!std::regex_match(s, match, regex)) - return {s, DefaultOutputs()}; - - if (match[3].matched) - return {match[1], AllOutputs()}; - - return {match[1], tokenizeString<OutputNames>(match[4].str(), ",")}; -} - -std::string printOutputsSpec(const OutputsSpec & outputsSpec) -{ - if (std::get_if<DefaultOutputs>(&outputsSpec)) - return ""; - - if (std::get_if<AllOutputs>(&outputsSpec)) - return "^*"; - - if (auto outputNames = std::get_if<OutputNames>(&outputsSpec)) - return "^" + concatStringsSep(",", *outputNames); - - assert(false); -} - -void to_json(nlohmann::json & json, const OutputsSpec & outputsSpec) -{ - if (std::get_if<DefaultOutputs>(&outputsSpec)) - json = nullptr; - - else if (std::get_if<AllOutputs>(&outputsSpec)) - json = std::vector<std::string>({"*"}); - - else if (auto outputNames = std::get_if<OutputNames>(&outputsSpec)) - json = *outputNames; -} - -void from_json(const nlohmann::json & json, OutputsSpec & outputsSpec) -{ - if (json.is_null()) - outputsSpec = DefaultOutputs(); - else { - auto names = json.get<OutputNames>(); - if (names == OutputNames({"*"})) - outputsSpec = AllOutputs(); - else - outputsSpec = names; - } -} - } diff --git a/src/libstore/path-with-outputs.hh b/src/libstore/path-with-outputs.hh index 0cb5eb223..5d25656a5 100644 --- a/src/libstore/path-with-outputs.hh +++ b/src/libstore/path-with-outputs.hh @@ -1,13 +1,17 @@ #pragma once -#include <variant> - #include "path.hh" #include "derived-path.hh" -#include "nlohmann/json_fwd.hpp" namespace nix { +/* This is a deprecated old type just for use by the old CLI, and older + versions of the RPC protocols. In new code don't use it; you want + `DerivedPath` instead. + + `DerivedPath` is better because it handles more cases, and does so more + explicitly without devious punning tricks. +*/ struct StorePathWithOutputs { StorePath path; @@ -33,25 +37,4 @@ StorePathWithOutputs parsePathWithOutputs(const Store & store, std::string_view StorePathWithOutputs followLinksToStorePathWithOutputs(const Store & store, std::string_view pathWithOutputs); -typedef std::set<std::string> OutputNames; - -struct AllOutputs { - bool operator < (const AllOutputs & _) const { return false; } -}; - -struct DefaultOutputs { - bool operator < (const DefaultOutputs & _) const { return false; } -}; - -typedef std::variant<DefaultOutputs, AllOutputs, OutputNames> OutputsSpec; - -/* Parse a string of the form 'prefix^output1,...outputN' or - 'prefix^*', returning the prefix and the outputs spec. */ -std::pair<std::string, OutputsSpec> parseOutputsSpec(const std::string & s); - -std::string printOutputsSpec(const OutputsSpec & outputsSpec); - -void to_json(nlohmann::json &, const OutputsSpec &); -void from_json(const nlohmann::json &, OutputsSpec &); - } diff --git a/src/libstore/path.cc b/src/libstore/path.cc index 392db225e..46be54281 100644 --- a/src/libstore/path.cc +++ b/src/libstore/path.cc @@ -8,8 +8,10 @@ static void checkName(std::string_view path, std::string_view name) { if (name.empty()) throw BadStorePath("store path '%s' has an empty name", path); - if (name.size() > 211) - throw BadStorePath("store path '%s' has a name longer than 211 characters", path); + if (name.size() > StorePath::MaxPathLen) + throw BadStorePath("store path '%s' has a name longer than '%d characters", + StorePath::MaxPathLen, path); + // See nameRegexStr for the definition for (auto c : name) if (!((c >= '0' && c <= '9') || (c >= 'a' && c <= 'z') diff --git a/src/libstore/path.hh b/src/libstore/path.hh index 77fd0f8dc..1e5579b90 100644 --- a/src/libstore/path.hh +++ b/src/libstore/path.hh @@ -1,11 +1,11 @@ #pragma once -#include "content-address.hh" +#include <string_view> + #include "types.hh" namespace nix { -class Store; struct Hash; class StorePath @@ -17,6 +17,8 @@ public: /* Size of the hash part of store paths, in base-32 characters. */ constexpr static size_t HashLen = 32; // i.e. 160 bits + constexpr static size_t MaxPathLen = 211; + StorePath() = delete; StorePath(std::string_view baseName); @@ -64,9 +66,6 @@ public: typedef std::set<StorePath> StorePathSet; typedef std::vector<StorePath> StorePaths; -typedef std::map<std::string, StorePath> OutputPathMap; - -typedef std::map<StorePath, std::optional<ContentAddress>> StorePathCAMap; /* Extension of derivations in the Nix store. */ const std::string drvExtension = ".drv"; diff --git a/src/libstore/pathlocks.cc b/src/libstore/pathlocks.cc index 42023cd0a..adc763e6a 100644 --- a/src/libstore/pathlocks.cc +++ b/src/libstore/pathlocks.cc @@ -96,7 +96,7 @@ bool PathLocks::lockPaths(const PathSet & paths, checkInterrupt(); Path lockPath = path + ".lock"; - debug(format("locking path '%1%'") % path); + debug("locking path '%1%'", path); AutoCloseFD fd; @@ -118,7 +118,7 @@ bool PathLocks::lockPaths(const PathSet & paths, } } - debug(format("lock acquired on '%1%'") % lockPath); + debug("lock acquired on '%1%'", lockPath); /* Check that the lock file hasn't become stale (i.e., hasn't been unlinked). */ @@ -130,7 +130,7 @@ bool PathLocks::lockPaths(const PathSet & paths, a lock on a deleted file. This means that other processes may create and acquire a lock on `lockPath', and proceed. So we must retry. */ - debug(format("open lock file '%1%' has become stale") % lockPath); + debug("open lock file '%1%' has become stale", lockPath); else break; } @@ -163,7 +163,7 @@ void PathLocks::unlock() "error (ignored): cannot close lock file on '%1%'", i.second); - debug(format("lock released on '%1%'") % i.second); + debug("lock released on '%1%'", i.second); } fds.clear(); diff --git a/src/libstore/profiles.cc b/src/libstore/profiles.cc index 3e4188188..179161ff7 100644 --- a/src/libstore/profiles.cc +++ b/src/libstore/profiles.cc @@ -64,7 +64,7 @@ std::pair<Generations, std::optional<GenerationNumber>> findGenerations(Path pro static void makeName(const Path & profile, GenerationNumber num, Path & outLink) { - Path prefix = (format("%1%-%2%") % profile % num).str(); + Path prefix = fmt("%1%-%2%", profile, num); outLink = prefix + "-link"; } @@ -269,7 +269,7 @@ void switchGeneration( void lockProfile(PathLocks & lock, const Path & profile) { - lock.lockPaths({profile}, (format("waiting for lock on profile '%1%'") % profile).str()); + lock.lockPaths({profile}, fmt("waiting for lock on profile '%1%'", profile)); lock.setDeletion(true); } @@ -280,16 +280,24 @@ std::string optimisticLockProfile(const Path & profile) } +Path profilesDir() +{ + auto profileRoot = createNixStateDir() + "/profiles"; + createDirs(profileRoot); + return profileRoot; +} + + Path getDefaultProfile() { - Path profileLink = getHome() + "/.nix-profile"; + Path profileLink = settings.useXDGBaseDirectories ? createNixStateDir() + "/profile" : getHome() + "/.nix-profile"; try { + auto profile = + getuid() == 0 + ? settings.nixStateDir + "/profiles/default" + : profilesDir() + "/profile"; if (!pathExists(profileLink)) { - replaceSymlink( - getuid() == 0 - ? settings.nixStateDir + "/profiles/default" - : fmt("%s/profiles/per-user/%s/profile", settings.nixStateDir, getUserName()), - profileLink); + replaceSymlink(profile, profileLink); } return absPath(readLink(profileLink), dirOf(profileLink)); } catch (Error &) { diff --git a/src/libstore/profiles.hh b/src/libstore/profiles.hh index 408ca039c..fbf95b850 100644 --- a/src/libstore/profiles.hh +++ b/src/libstore/profiles.hh @@ -68,8 +68,13 @@ void lockProfile(PathLocks & lock, const Path & profile); rebuilt. */ std::string optimisticLockProfile(const Path & profile); -/* Resolve ~/.nix-profile. If ~/.nix-profile doesn't exist yet, create - it. */ +/* Creates and returns the path to a directory suitable for storing the user’s + profiles. */ +Path profilesDir(); + +/* Resolve the default profile (~/.nix-profile by default, $XDG_STATE_HOME/ + nix/profile if XDG Base Directory Support is enabled), and create if doesn't + exist */ Path getDefaultProfile(); } diff --git a/src/libstore/realisation.hh b/src/libstore/realisation.hh index 9070a6ee2..48d0283de 100644 --- a/src/libstore/realisation.hh +++ b/src/libstore/realisation.hh @@ -1,5 +1,8 @@ #pragma once +#include <variant> + +#include "hash.hh" #include "path.hh" #include <nlohmann/json_fwd.hpp> #include "comparator.hh" @@ -7,6 +10,8 @@ namespace nix { +class Store; + struct DrvOutput { // The hash modulo of the derivation Hash drvHash; @@ -93,4 +98,14 @@ struct RealisedPath { GENERATE_CMP(RealisedPath, me->raw); }; +class MissingRealisation : public Error +{ +public: + MissingRealisation(DrvOutput & outputId) + : Error( "cannot operate on an output of the " + "unbuilt derivation '%s'", + outputId.to_string()) + {} +}; + } diff --git a/src/libstore/references.cc b/src/libstore/references.cc index 34dce092c..345f4528b 100644 --- a/src/libstore/references.cc +++ b/src/libstore/references.cc @@ -39,8 +39,7 @@ static void search( if (!match) continue; std::string ref(s.substr(i, refLength)); if (hashes.erase(ref)) { - debug(format("found reference to '%1%' at offset '%2%'") - % ref % i); + debug("found reference to '%1%' at offset '%2%'", ref, i); seen.insert(ref); } ++i; @@ -67,20 +66,12 @@ void RefScanSink::operator () (std::string_view data) } -std::pair<StorePathSet, HashResult> scanForReferences( - const std::string & path, - const StorePathSet & refs) -{ - HashSink hashSink { htSHA256 }; - auto found = scanForReferences(hashSink, path, refs); - auto hash = hashSink.finish(); - return std::pair<StorePathSet, HashResult>(found, hash); -} +PathRefScanSink::PathRefScanSink(StringSet && hashes, std::map<std::string, StorePath> && backMap) + : RefScanSink(std::move(hashes)) + , backMap(std::move(backMap)) +{ } -StorePathSet scanForReferences( - Sink & toTee, - const Path & path, - const StorePathSet & refs) +PathRefScanSink PathRefScanSink::fromPaths(const StorePathSet & refs) { StringSet hashes; std::map<std::string, StorePath> backMap; @@ -92,14 +83,14 @@ StorePathSet scanForReferences( hashes.insert(hashPart); } - /* Look for the hashes in the NAR dump of the path. */ - RefScanSink refsSink(std::move(hashes)); - TeeSink sink { refsSink, toTee }; - dumpPath(path, sink); + return PathRefScanSink(std::move(hashes), std::move(backMap)); +} +StorePathSet PathRefScanSink::getResultPaths() +{ /* Map the hashes found back to their store paths. */ StorePathSet found; - for (auto & i : refsSink.getResult()) { + for (auto & i : getResult()) { auto j = backMap.find(i); assert(j != backMap.end()); found.insert(j->second); @@ -109,6 +100,31 @@ StorePathSet scanForReferences( } +std::pair<StorePathSet, HashResult> scanForReferences( + const std::string & path, + const StorePathSet & refs) +{ + HashSink hashSink { htSHA256 }; + auto found = scanForReferences(hashSink, path, refs); + auto hash = hashSink.finish(); + return std::pair<StorePathSet, HashResult>(found, hash); +} + +StorePathSet scanForReferences( + Sink & toTee, + const Path & path, + const StorePathSet & refs) +{ + PathRefScanSink refsSink = PathRefScanSink::fromPaths(refs); + TeeSink sink { refsSink, toTee }; + + /* Look for the hashes in the NAR dump of the path. */ + dumpPath(path, sink); + + return refsSink.getResultPaths(); +} + + RewritingSink::RewritingSink(const std::string & from, const std::string & to, Sink & nextSink) : from(from), to(to), nextSink(nextSink) { diff --git a/src/libstore/references.hh b/src/libstore/references.hh index a6119c861..6f381f96c 100644 --- a/src/libstore/references.hh +++ b/src/libstore/references.hh @@ -27,6 +27,19 @@ public: void operator () (std::string_view data) override; }; +class PathRefScanSink : public RefScanSink +{ + std::map<std::string, StorePath> backMap; + + PathRefScanSink(StringSet && hashes, std::map<std::string, StorePath> && backMap); + +public: + + static PathRefScanSink fromPaths(const StorePathSet & refs); + + StorePathSet getResultPaths(); +}; + struct RewritingSink : Sink { std::string from, to, prev; diff --git a/src/libstore/remote-fs-accessor.cc b/src/libstore/remote-fs-accessor.cc index 0ce335646..fcfb527f5 100644 --- a/src/libstore/remote-fs-accessor.cc +++ b/src/libstore/remote-fs-accessor.cc @@ -1,6 +1,6 @@ +#include <nlohmann/json.hpp> #include "remote-fs-accessor.hh" #include "nar-accessor.hh" -#include "json.hh" #include <sys/types.h> #include <sys/stat.h> @@ -38,10 +38,8 @@ ref<FSAccessor> RemoteFSAccessor::addToCache(std::string_view hashPart, std::str if (cacheDir != "") { try { - std::ostringstream str; - JSONPlaceholder jsonRoot(str); - listNar(jsonRoot, narAccessor, "", true); - writeFile(makeCacheFile(hashPart, "ls"), str.str()); + nlohmann::json j = listNar(narAccessor, "", true); + writeFile(makeCacheFile(hashPart, "ls"), j.dump()); } catch (...) { ignoreException(); } diff --git a/src/libstore/remote-store.cc b/src/libstore/remote-store.cc index bc36aef5d..d1296627a 100644 --- a/src/libstore/remote-store.cc +++ b/src/libstore/remote-store.cc @@ -266,6 +266,7 @@ void RemoteStore::setOptions(Connection & conn) overrides.erase(settings.useSubstitutes.name); overrides.erase(loggerSettings.showTrace.name); overrides.erase(settings.experimentalFeatures.name); + overrides.erase(settings.pluginFiles.name); conn.to << overrides.size(); for (auto & i : overrides) conn.to << i.first << i.second.value; @@ -447,7 +448,7 @@ void RemoteStore::queryPathInfoUncached(const StorePath & path, } catch (Error & e) { // Ugly backwards compatibility hack. if (e.msg().find("is not valid") != std::string::npos) - throw InvalidPath(e.info()); + throw InvalidPath(std::move(e.info())); throw; } if (GET_PROTOCOL_MINOR(conn->daemonVersion) >= 17) { @@ -580,7 +581,6 @@ ref<const ValidPathInfo> RemoteStore::addCAToStore( try { conn->to.written = 0; - conn->to.warn = true; connections->incCapacity(); { Finally cleanup([&]() { connections->decCapacity(); }); @@ -591,7 +591,6 @@ ref<const ValidPathInfo> RemoteStore::addCAToStore( dumpString(contents, conn->to); } } - conn->to.warn = false; conn.processStderr(); } catch (SysError & e) { /* Daemon closed while we were sending the path. Probably OOM @@ -674,6 +673,23 @@ void RemoteStore::addToStore(const ValidPathInfo & info, Source & source, void RemoteStore::addMultipleToStore( + PathsSource & pathsToCopy, + Activity & act, + RepairFlag repair, + CheckSigsFlag checkSigs) +{ + auto source = sinkToSource([&](Sink & sink) { + sink << pathsToCopy.size(); + for (auto & [pathInfo, pathSource] : pathsToCopy) { + pathInfo.write(sink, *this, 16); + pathSource->drainInto(sink); + } + }); + + addMultipleToStore(*source, repair, checkSigs); +} + +void RemoteStore::addMultipleToStore( Source & source, RepairFlag repair, CheckSigsFlag checkSigs) @@ -852,8 +868,8 @@ std::vector<BuildResult> RemoteStore::buildPathsWithResults( OutputPathMap outputs; auto drv = evalStore->readDerivation(bfd.drvPath); const auto outputHashes = staticOutputHashes(*evalStore, drv); // FIXME: expensive - const auto drvOutputs = drv.outputsAndOptPaths(*this); - for (auto & output : bfd.outputs) { + auto built = resolveDerivedPath(*this, bfd, &*evalStore); + for (auto & [output, outputPath] : built) { auto outputHash = get(outputHashes, output); if (!outputHash) throw Error( @@ -864,22 +880,14 @@ std::vector<BuildResult> RemoteStore::buildPathsWithResults( auto realisation = queryRealisation(outputId); if (!realisation) - throw Error( - "cannot operate on an output of unbuilt " - "content-addressed derivation '%s'", - outputId.to_string()); + throw MissingRealisation(outputId); res.builtOutputs.emplace(realisation->id, *realisation); } else { - // If ca-derivations isn't enabled, assume that - // the output path is statically known. - const auto drvOutput = get(drvOutputs, output); - assert(drvOutput); - assert(drvOutput->second); res.builtOutputs.emplace( outputId, Realisation { .id = outputId, - .outPath = *drvOutput->second, + .outPath = outputPath, }); } } @@ -903,7 +911,12 @@ BuildResult RemoteStore::buildDerivation(const StorePath & drvPath, const BasicD writeDerivation(conn->to, *this, drv); conn->to << buildMode; conn.processStderr(); - BuildResult res { .path = DerivedPath::Built { .drvPath = drvPath } }; + BuildResult res { + .path = DerivedPath::Built { + .drvPath = drvPath, + .outputs = OutputsSpec::All { }, + }, + }; res.status = (BuildResult::Status) readInt(conn->from); conn->from >> res.errorMsg; if (GET_PROTOCOL_MINOR(conn->daemonVersion) >= 29) { diff --git a/src/libstore/remote-store.hh b/src/libstore/remote-store.hh index 8493be6fc..8cd7cc822 100644 --- a/src/libstore/remote-store.hh +++ b/src/libstore/remote-store.hh @@ -38,8 +38,6 @@ class RemoteStore : public virtual RemoteStoreConfig, { public: - virtual bool sameMachine() = 0; - RemoteStore(const Params & params); /* Implementations of abstract store API methods. */ @@ -88,6 +86,12 @@ public: RepairFlag repair, CheckSigsFlag checkSigs) override; + void addMultipleToStore( + PathsSource & pathsToCopy, + Activity & act, + RepairFlag repair, + CheckSigsFlag checkSigs) override; + StorePath addTextToStore( std::string_view name, std::string_view s, diff --git a/src/libstore/s3-binary-cache-store.cc b/src/libstore/s3-binary-cache-store.cc index 844553ad3..8006bd733 100644 --- a/src/libstore/s3-binary-cache-store.cc +++ b/src/libstore/s3-binary-cache-store.cc @@ -40,12 +40,12 @@ struct S3Error : public Error /* Helper: given an Outcome<R, E>, return R in case of success, or throw an exception in case of an error. */ template<typename R, typename E> -R && checkAws(const FormatOrString & fs, Aws::Utils::Outcome<R, E> && outcome) +R && checkAws(std::string_view s, Aws::Utils::Outcome<R, E> && outcome) { if (!outcome.IsSuccess()) throw S3Error( outcome.GetError().GetErrorType(), - fs.s + ": " + outcome.GetError().GetMessage()); + s + ": " + outcome.GetError().GetMessage()); return outcome.GetResultWithOwnership(); } @@ -238,7 +238,7 @@ struct S3BinaryCacheStoreImpl : virtual S3BinaryCacheStoreConfig, public virtual void init() override { - if (auto cacheInfo = diskCache->cacheExists(getUri())) { + if (auto cacheInfo = diskCache->upToDateCacheExists(getUri())) { wantMassQuery.setDefault(cacheInfo->wantMassQuery); priority.setDefault(cacheInfo->priority); } else { @@ -430,9 +430,9 @@ struct S3BinaryCacheStoreImpl : virtual S3BinaryCacheStoreConfig, public virtual std::string marker; do { - debug(format("listing bucket 's3://%s' from key '%s'...") % bucketName % marker); + debug("listing bucket 's3://%s' from key '%s'...", bucketName, marker); - auto res = checkAws(format("AWS error listing bucket '%s'") % bucketName, + auto res = checkAws(fmt("AWS error listing bucket '%s'", bucketName), s3Helper.client->ListObjects( Aws::S3::Model::ListObjectsRequest() .WithBucket(bucketName) @@ -441,8 +441,8 @@ struct S3BinaryCacheStoreImpl : virtual S3BinaryCacheStoreConfig, public virtual auto & contents = res.GetContents(); - debug(format("got %d keys, next marker '%s'") - % contents.size() % res.GetNextMarker()); + debug("got %d keys, next marker '%s'", + contents.size(), res.GetNextMarker()); for (auto object : contents) { auto & key = object.GetKey(); diff --git a/src/libstore/sqlite.cc b/src/libstore/sqlite.cc index 2090beabd..871f2f3be 100644 --- a/src/libstore/sqlite.cc +++ b/src/libstore/sqlite.cc @@ -8,12 +8,15 @@ namespace nix { -SQLiteError::SQLiteError(const char *path, int errNo, int extendedErrNo, hintformat && hf) - : Error(""), path(path), errNo(errNo), extendedErrNo(extendedErrNo) +SQLiteError::SQLiteError(const char *path, const char *errMsg, int errNo, int extendedErrNo, int offset, hintformat && hf) + : Error(""), path(path), errMsg(errMsg), errNo(errNo), extendedErrNo(extendedErrNo), offset(offset) { - err.msg = hintfmt("%s: %s (in '%s')", + auto offsetStr = (offset == -1) ? "" : "at offset " + std::to_string(offset) + ": "; + err.msg = hintfmt("%s: %s%s, %s (in '%s')", normaltxt(hf.str()), + offsetStr, sqlite3_errstr(extendedErrNo), + errMsg, path ? path : "(in-memory)"); } @@ -21,11 +24,13 @@ SQLiteError::SQLiteError(const char *path, int errNo, int extendedErrNo, hintfor { int err = sqlite3_errcode(db); int exterr = sqlite3_extended_errcode(db); + int offset = sqlite3_error_offset(db); auto path = sqlite3_db_filename(db, nullptr); + auto errMsg = sqlite3_errmsg(db); if (err == SQLITE_BUSY || err == SQLITE_PROTOCOL) { - auto exp = SQLiteBusy(path, err, exterr, std::move(hf)); + auto exp = SQLiteBusy(path, errMsg, err, exterr, offset, std::move(hf)); exp.err.msg = hintfmt( err == SQLITE_PROTOCOL ? "SQLite database '%s' is busy (SQLITE_PROTOCOL)" @@ -33,22 +38,40 @@ SQLiteError::SQLiteError(const char *path, int errNo, int extendedErrNo, hintfor path ? path : "(in-memory)"); throw exp; } else - throw SQLiteError(path, err, exterr, std::move(hf)); + throw SQLiteError(path, errMsg, err, exterr, offset, std::move(hf)); } +static void traceSQL(void * x, const char * sql) +{ + // wacky delimiters: + // so that we're quite unambiguous without escaping anything + // notice instead of trace: + // so that this can be enabled without getting the firehose in our face. + notice("SQL<[%1%]>", sql); +}; + SQLite::SQLite(const Path & path, bool create) { // useSQLiteWAL also indicates what virtual file system we need. Using // `unix-dotfile` is needed on NFS file systems and on Windows' Subsystem // for Linux (WSL) where useSQLiteWAL should be false by default. const char *vfs = settings.useSQLiteWAL ? 0 : "unix-dotfile"; - if (sqlite3_open_v2(path.c_str(), &db, - SQLITE_OPEN_READWRITE | (create ? SQLITE_OPEN_CREATE : 0), vfs) != SQLITE_OK) - throw Error("cannot open SQLite database '%s'", path); + int flags = SQLITE_OPEN_READWRITE; + if (create) flags |= SQLITE_OPEN_CREATE; + int ret = sqlite3_open_v2(path.c_str(), &db, flags, vfs); + if (ret != SQLITE_OK) { + const char * err = sqlite3_errstr(ret); + throw Error("cannot open SQLite database '%s': %s", path, err); + } if (sqlite3_busy_timeout(db, 60 * 60 * 1000) != SQLITE_OK) SQLiteError::throw_(db, "setting timeout"); + if (getEnv("NIX_DEBUG_SQLITE_TRACES") == "1") { + // To debug sqlite statements; trace all of them + sqlite3_trace(db, &traceSQL, nullptr); + } + exec("pragma foreign_keys = 1"); } diff --git a/src/libstore/sqlite.hh b/src/libstore/sqlite.hh index 1d1c553ea..1853731a2 100644 --- a/src/libstore/sqlite.hh +++ b/src/libstore/sqlite.hh @@ -98,21 +98,22 @@ struct SQLiteTxn struct SQLiteError : Error { - const char *path; - int errNo, extendedErrNo; + std::string path; + std::string errMsg; + int errNo, extendedErrNo, offset; template<typename... Args> [[noreturn]] static void throw_(sqlite3 * db, const std::string & fs, const Args & ... args) { throw_(db, hintfmt(fs, args...)); } - SQLiteError(const char *path, int errNo, int extendedErrNo, hintformat && hf); + SQLiteError(const char *path, const char *errMsg, int errNo, int extendedErrNo, int offset, hintformat && hf); protected: template<typename... Args> - SQLiteError(const char *path, int errNo, int extendedErrNo, const std::string & fs, const Args & ... args) - : SQLiteError(path, errNo, extendedErrNo, hintfmt(fs, args...)) + SQLiteError(const char *path, const char *errMsg, int errNo, int extendedErrNo, int offset, const std::string & fs, const Args & ... args) + : SQLiteError(path, errNo, extendedErrNo, offset, hintfmt(fs, args...)) { } [[noreturn]] static void throw_(sqlite3 * db, hintformat && hf); diff --git a/src/libstore/ssh-store.cc b/src/libstore/ssh-store.cc index 62daa838c..cfa063803 100644 --- a/src/libstore/ssh-store.cc +++ b/src/libstore/ssh-store.cc @@ -49,12 +49,9 @@ public: return *uriSchemes().begin() + "://" + host; } - bool sameMachine() override - { return false; } - // FIXME extend daemon protocol, move implementation to RemoteStore - std::optional<std::string> getBuildLog(const StorePath & path) override - { unsupported("getBuildLog"); } + std::optional<std::string> getBuildLogExact(const StorePath & path) override + { unsupported("getBuildLogExact"); } private: diff --git a/src/libstore/ssh.cc b/src/libstore/ssh.cc index 1bbad71f2..69bfe3418 100644 --- a/src/libstore/ssh.cc +++ b/src/libstore/ssh.cc @@ -67,7 +67,7 @@ std::unique_ptr<SSHMaster::Connection> SSHMaster::startCommand(const std::string if (fakeSSH) { args = { "bash", "-c" }; } else { - args = { "ssh", host.c_str(), "-x", "-a" }; + args = { "ssh", host.c_str(), "-x" }; addCommonSSHOpts(args); if (socketPath != "") args.insert(args.end(), {"-S", socketPath}); diff --git a/src/libstore/store-api.cc b/src/libstore/store-api.cc index 91dbd991e..19b0a7f5f 100644 --- a/src/libstore/store-api.cc +++ b/src/libstore/store-api.cc @@ -6,32 +6,34 @@ #include "util.hh" #include "nar-info-disk-cache.hh" #include "thread-pool.hh" -#include "json.hh" #include "url.hh" #include "archive.hh" #include "callback.hh" #include "remote-store.hh" +#include <nlohmann/json.hpp> #include <regex> +using json = nlohmann::json; + namespace nix { -bool Store::isInStore(const Path & path) const +bool Store::isInStore(PathView path) const { return isInDir(path, storeDir); } -std::pair<StorePath, Path> Store::toStorePath(const Path & path) const +std::pair<StorePath, Path> Store::toStorePath(PathView path) const { if (!isInStore(path)) throw Error("path '%1%' is not in the Nix store", path); - Path::size_type slash = path.find('/', storeDir.size() + 1); + auto slash = path.find('/', storeDir.size() + 1); if (slash == Path::npos) return {parseStorePath(path), ""}; else - return {parseStorePath(std::string_view(path).substr(0, slash)), path.substr(slash)}; + return {parseStorePath(path.substr(0, slash)), (Path) path.substr(slash)}; } @@ -258,6 +260,84 @@ StorePath Store::addToStore( return addToStoreFromDump(*source, name, method, hashAlgo, repair, references); } +void Store::addMultipleToStore( + PathsSource & pathsToCopy, + Activity & act, + RepairFlag repair, + CheckSigsFlag checkSigs) +{ + std::atomic<size_t> nrDone{0}; + std::atomic<size_t> nrFailed{0}; + std::atomic<uint64_t> bytesExpected{0}; + std::atomic<uint64_t> nrRunning{0}; + + using PathWithInfo = std::pair<ValidPathInfo, std::unique_ptr<Source>>; + + std::map<StorePath, PathWithInfo *> infosMap; + StorePathSet storePathsToAdd; + for (auto & thingToAdd : pathsToCopy) { + infosMap.insert_or_assign(thingToAdd.first.path, &thingToAdd); + storePathsToAdd.insert(thingToAdd.first.path); + } + + auto showProgress = [&]() { + act.progress(nrDone, pathsToCopy.size(), nrRunning, nrFailed); + }; + + ThreadPool pool; + + processGraph<StorePath>(pool, + storePathsToAdd, + + [&](const StorePath & path) { + + auto & [info, _] = *infosMap.at(path); + + if (isValidPath(info.path)) { + nrDone++; + showProgress(); + return StorePathSet(); + } + + bytesExpected += info.narSize; + act.setExpected(actCopyPath, bytesExpected); + + return info.references; + }, + + [&](const StorePath & path) { + checkInterrupt(); + + auto & [info_, source_] = *infosMap.at(path); + auto info = info_; + info.ultimate = false; + + /* Make sure that the Source object is destroyed when + we're done. In particular, a SinkToSource object must + be destroyed to ensure that the destructors on its + stack frame are run; this includes + LegacySSHStore::narFromPath()'s connection lock. */ + auto source = std::move(source_); + + if (!isValidPath(info.path)) { + MaintainCount<decltype(nrRunning)> mc(nrRunning); + showProgress(); + try { + addToStore(info, *source, repair, checkSigs); + } catch (Error & e) { + nrFailed++; + if (!settings.keepGoing) + throw e; + printMsg(lvlError, "could not copy %s: %s", printStorePath(path), e.what()); + showProgress(); + return; + } + } + + nrDone++; + showProgress(); + }); +} void Store::addMultipleToStore( Source & source, @@ -378,6 +458,7 @@ Store::Store(const Params & params) : StoreConfig(params) , state({(size_t) pathInfoCacheSize}) { + assertLibStoreInitialized(); } @@ -661,13 +742,13 @@ StorePathSet Store::queryValidPaths(const StorePathSet & paths, SubstituteFlag m std::condition_variable wakeup; ThreadPool pool; - auto doQuery = [&](const Path & path) { + auto doQuery = [&](const StorePath & path) { checkInterrupt(); - queryPathInfo(parseStorePath(path), {[path, this, &state_, &wakeup](std::future<ref<const ValidPathInfo>> fut) { + queryPathInfo(path, {[path, &state_, &wakeup](std::future<ref<const ValidPathInfo>> fut) { auto state(state_.lock()); try { auto info = fut.get(); - state->valid.insert(parseStorePath(path)); + state->valid.insert(path); } catch (InvalidPath &) { } catch (...) { state->exc = std::current_exception(); @@ -679,7 +760,7 @@ StorePathSet Store::queryValidPaths(const StorePathSet & paths, SubstituteFlag m }; for (auto & path : paths) - pool.enqueue(std::bind(doQuery, printStorePath(path))); // FIXME + pool.enqueue(std::bind(doQuery, path)); pool.process(); @@ -709,13 +790,13 @@ std::string Store::makeValidityRegistration(const StorePathSet & paths, if (showHash) { s += info->narHash.to_string(Base16, false) + "\n"; - s += (format("%1%\n") % info->narSize).str(); + s += fmt("%1%\n", info->narSize); } auto deriver = showDerivers && info->deriver ? printStorePath(*info->deriver) : ""; s += deriver + "\n"; - s += (format("%1%\n") % info->references.size()).str(); + s += fmt("%1%\n", info->references.size()); for (auto & j : info->references) s += printStorePath(j) + "\n"; @@ -760,56 +841,54 @@ StorePathSet Store::exportReferences(const StorePathSet & storePaths, const Stor return paths; } - -void Store::pathInfoToJSON(JSONPlaceholder & jsonOut, const StorePathSet & storePaths, +json Store::pathInfoToJSON(const StorePathSet & storePaths, bool includeImpureInfo, bool showClosureSize, Base hashBase, AllowInvalidFlag allowInvalid) { - auto jsonList = jsonOut.list(); + json::array_t jsonList = json::array(); for (auto & storePath : storePaths) { - auto jsonPath = jsonList.object(); + auto& jsonPath = jsonList.emplace_back(json::object()); try { auto info = queryPathInfo(storePath); - jsonPath.attr("path", printStorePath(info->path)); - jsonPath - .attr("narHash", info->narHash.to_string(hashBase, true)) - .attr("narSize", info->narSize); + jsonPath["path"] = printStorePath(info->path); + jsonPath["valid"] = true; + jsonPath["narHash"] = info->narHash.to_string(hashBase, true); + jsonPath["narSize"] = info->narSize; { - auto jsonRefs = jsonPath.list("references"); + auto& jsonRefs = (jsonPath["references"] = json::array()); for (auto & ref : info->references) - jsonRefs.elem(printStorePath(ref)); + jsonRefs.emplace_back(printStorePath(ref)); } if (info->ca) - jsonPath.attr("ca", renderContentAddress(info->ca)); + jsonPath["ca"] = renderContentAddress(info->ca); std::pair<uint64_t, uint64_t> closureSizes; if (showClosureSize) { closureSizes = getClosureSize(info->path); - jsonPath.attr("closureSize", closureSizes.first); + jsonPath["closureSize"] = closureSizes.first; } if (includeImpureInfo) { if (info->deriver) - jsonPath.attr("deriver", printStorePath(*info->deriver)); + jsonPath["deriver"] = printStorePath(*info->deriver); if (info->registrationTime) - jsonPath.attr("registrationTime", info->registrationTime); + jsonPath["registrationTime"] = info->registrationTime; if (info->ultimate) - jsonPath.attr("ultimate", info->ultimate); + jsonPath["ultimate"] = info->ultimate; if (!info->sigs.empty()) { - auto jsonSigs = jsonPath.list("signatures"); for (auto & sig : info->sigs) - jsonSigs.elem(sig); + jsonPath["signatures"].push_back(sig); } auto narInfo = std::dynamic_pointer_cast<const NarInfo>( @@ -817,21 +896,22 @@ void Store::pathInfoToJSON(JSONPlaceholder & jsonOut, const StorePathSet & store if (narInfo) { if (!narInfo->url.empty()) - jsonPath.attr("url", narInfo->url); + jsonPath["url"] = narInfo->url; if (narInfo->fileHash) - jsonPath.attr("downloadHash", narInfo->fileHash->to_string(hashBase, true)); + jsonPath["downloadHash"] = narInfo->fileHash->to_string(hashBase, true); if (narInfo->fileSize) - jsonPath.attr("downloadSize", narInfo->fileSize); + jsonPath["downloadSize"] = narInfo->fileSize; if (showClosureSize) - jsonPath.attr("closureDownloadSize", closureSizes.second); + jsonPath["closureDownloadSize"] = closureSizes.second; } } } catch (InvalidPath &) { - jsonPath.attr("path", printStorePath(storePath)); - jsonPath.attr("valid", false); + jsonPath["path"] = printStorePath(storePath); + jsonPath["valid"] = false; } } + return jsonList; } @@ -992,113 +1072,69 @@ std::map<StorePath, StorePath> copyPaths( for (auto & path : storePaths) if (!valid.count(path)) missing.insert(path); + Activity act(*logger, lvlInfo, actCopyPaths, fmt("copying %d paths", missing.size())); + + // In the general case, `addMultipleToStore` requires a sorted list of + // store paths to add, so sort them right now + auto sortedMissing = srcStore.topoSortPaths(missing); + std::reverse(sortedMissing.begin(), sortedMissing.end()); + std::map<StorePath, StorePath> pathsMap; for (auto & path : storePaths) pathsMap.insert_or_assign(path, path); - Activity act(*logger, lvlInfo, actCopyPaths, fmt("copying %d paths", missing.size())); + Store::PathsSource pathsToCopy; + + auto computeStorePathForDst = [&](const ValidPathInfo & currentPathInfo) -> StorePath { + auto storePathForSrc = currentPathInfo.path; + auto storePathForDst = storePathForSrc; + if (currentPathInfo.ca && currentPathInfo.references.empty()) { + storePathForDst = dstStore.makeFixedOutputPathFromCA(storePathForSrc.name(), *currentPathInfo.ca); + if (dstStore.storeDir == srcStore.storeDir) + assert(storePathForDst == storePathForSrc); + if (storePathForDst != storePathForSrc) + debug("replaced path '%s' to '%s' for substituter '%s'", + srcStore.printStorePath(storePathForSrc), + dstStore.printStorePath(storePathForDst), + dstStore.getUri()); + } + return storePathForDst; + }; - auto sorted = srcStore.topoSortPaths(missing); - std::reverse(sorted.begin(), sorted.end()); + uint64_t total = 0; - auto source = sinkToSource([&](Sink & sink) { - sink << sorted.size(); - for (auto & storePath : sorted) { + for (auto & missingPath : sortedMissing) { + auto info = srcStore.queryPathInfo(missingPath); + + auto storePathForDst = computeStorePathForDst(*info); + pathsMap.insert_or_assign(missingPath, storePathForDst); + + ValidPathInfo infoForDst = *info; + infoForDst.path = storePathForDst; + + auto source = sinkToSource([&](Sink & sink) { + // We can reasonably assume that the copy will happen whenever we + // read the path, so log something about that at that point auto srcUri = srcStore.getUri(); auto dstUri = dstStore.getUri(); - auto storePathS = srcStore.printStorePath(storePath); + auto storePathS = srcStore.printStorePath(missingPath); Activity act(*logger, lvlInfo, actCopyPath, makeCopyPathMessage(srcUri, dstUri, storePathS), {storePathS, srcUri, dstUri}); PushActivity pact(act.id); - auto info = srcStore.queryPathInfo(storePath); - info->write(sink, srcStore, 16); - srcStore.narFromPath(storePath, sink); - } - }); - - dstStore.addMultipleToStore(*source, repair, checkSigs); - - #if 0 - std::atomic<size_t> nrDone{0}; - std::atomic<size_t> nrFailed{0}; - std::atomic<uint64_t> bytesExpected{0}; - std::atomic<uint64_t> nrRunning{0}; - - auto showProgress = [&]() { - act.progress(nrDone, missing.size(), nrRunning, nrFailed); - }; - - ThreadPool pool; - - processGraph<StorePath>(pool, - StorePathSet(missing.begin(), missing.end()), - - [&](const StorePath & storePath) { - auto info = srcStore.queryPathInfo(storePath); - auto storePathForDst = storePath; - if (info->ca && info->references.empty()) { - storePathForDst = dstStore.makeFixedOutputPathFromCA(storePath.name(), *info->ca); - if (dstStore.storeDir == srcStore.storeDir) - assert(storePathForDst == storePath); - if (storePathForDst != storePath) - debug("replaced path '%s' to '%s' for substituter '%s'", - srcStore.printStorePath(storePath), - dstStore.printStorePath(storePathForDst), - dstStore.getUri()); - } - pathsMap.insert_or_assign(storePath, storePathForDst); - - if (dstStore.isValidPath(storePath)) { - nrDone++; - showProgress(); - return StorePathSet(); - } - - bytesExpected += info->narSize; - act.setExpected(actCopyPath, bytesExpected); - - return info->references; - }, - - [&](const StorePath & storePath) { - checkInterrupt(); - - auto info = srcStore.queryPathInfo(storePath); - - auto storePathForDst = storePath; - if (info->ca && info->references.empty()) { - storePathForDst = dstStore.makeFixedOutputPathFromCA(storePath.name(), *info->ca); - if (dstStore.storeDir == srcStore.storeDir) - assert(storePathForDst == storePath); - if (storePathForDst != storePath) - debug("replaced path '%s' to '%s' for substituter '%s'", - srcStore.printStorePath(storePath), - dstStore.printStorePath(storePathForDst), - dstStore.getUri()); - } - pathsMap.insert_or_assign(storePath, storePathForDst); - - if (!dstStore.isValidPath(storePathForDst)) { - MaintainCount<decltype(nrRunning)> mc(nrRunning); - showProgress(); - try { - copyStorePath(srcStore, dstStore, storePath, repair, checkSigs); - } catch (Error &e) { - nrFailed++; - if (!settings.keepGoing) - throw e; - printMsg(lvlError, "could not copy %s: %s", dstStore.printStorePath(storePath), e.what()); - showProgress(); - return; - } - } + LambdaSink progressSink([&](std::string_view data) { + total += data.size(); + act.progress(total, info->narSize); + }); + TeeSink tee { sink, progressSink }; - nrDone++; - showProgress(); + srcStore.narFromPath(missingPath, tee); }); - #endif + pathsToCopy.push_back(std::pair{infoForDst, std::move(source)}); + } + + dstStore.addMultipleToStore(pathsToCopy, act, repair, checkSigs); return pathsMap; } @@ -1183,79 +1219,6 @@ std::string showPaths(const PathSet & paths) } -std::string ValidPathInfo::fingerprint(const Store & store) const -{ - if (narSize == 0) - throw Error("cannot calculate fingerprint of path '%s' because its size is not known", - store.printStorePath(path)); - return - "1;" + store.printStorePath(path) + ";" - + narHash.to_string(Base32, true) + ";" - + std::to_string(narSize) + ";" - + concatStringsSep(",", store.printStorePathSet(references)); -} - - -void ValidPathInfo::sign(const Store & store, const SecretKey & secretKey) -{ - sigs.insert(secretKey.signDetached(fingerprint(store))); -} - -bool ValidPathInfo::isContentAddressed(const Store & store) const -{ - if (! ca) return false; - - auto caPath = std::visit(overloaded { - [&](const TextHash & th) { - return store.makeTextPath(path.name(), th.hash, references); - }, - [&](const FixedOutputHash & fsh) { - auto refs = references; - bool hasSelfReference = false; - if (refs.count(path)) { - hasSelfReference = true; - refs.erase(path); - } - return store.makeFixedOutputPath(fsh.method, fsh.hash, path.name(), refs, hasSelfReference); - } - }, *ca); - - bool res = caPath == path; - - if (!res) - printError("warning: path '%s' claims to be content-addressed but isn't", store.printStorePath(path)); - - return res; -} - - -size_t ValidPathInfo::checkSignatures(const Store & store, const PublicKeys & publicKeys) const -{ - if (isContentAddressed(store)) return maxSigs; - - size_t good = 0; - for (auto & sig : sigs) - if (checkSignature(store, publicKeys, sig)) - good++; - return good; -} - - -bool ValidPathInfo::checkSignature(const Store & store, const PublicKeys & publicKeys, const std::string & sig) const -{ - return verifyDetached(fingerprint(store), sig, publicKeys); -} - - -Strings ValidPathInfo::shortRefs() const -{ - Strings refs; - for (auto & r : references) - refs.push_back(std::string(r.to_string())); - return refs; -} - - Derivation Store::derivationFromPath(const StorePath & drvPath) { ensurePath(drvPath); @@ -1274,6 +1237,34 @@ Derivation readDerivationCommon(Store& store, const StorePath& drvPath, bool req } } +std::optional<StorePath> Store::getBuildDerivationPath(const StorePath & path) +{ + + if (!path.isDerivation()) { + try { + auto info = queryPathInfo(path); + if (!info->deriver) return std::nullopt; + return *info->deriver; + } catch (InvalidPath &) { + return std::nullopt; + } + } + + if (!settings.isExperimentalFeatureEnabled(Xp::CaDerivations) || !isValidPath(path)) + return path; + + auto drv = readDerivation(path); + if (!drv.type().hasKnownOutputPaths()) { + // The build log is actually attached to the corresponding + // resolved derivation, so we need to get it first + auto resolvedDrv = drv.tryResolve(*this); + if (resolvedDrv) + return writeDerivation(*this, *resolvedDrv, NoRepair, true); + } + + return path; +} + Derivation Store::readDerivation(const StorePath & drvPath) { return readDerivationCommon(*this, drvPath, true); } @@ -1337,9 +1328,9 @@ std::shared_ptr<Store> openFromNonUri(const std::string & uri, const Store::Para } catch (Error & e) { return std::make_shared<LocalStore>(params); } - warn("'/nix' does not exist, so Nix will use '%s' as a chroot store", chrootStore); + warn("'%s' does not exist, so Nix will use '%s' as a chroot store", stateDir, chrootStore); } else - debug("'/nix' does not exist, so Nix will use '%s' as a chroot store", chrootStore); + debug("'%s' does not exist, so Nix will use '%s' as a chroot store", stateDir, chrootStore); Store::Params params2; params2["root"] = chrootStore; return std::make_shared<LocalStore>(params2); diff --git a/src/libstore/store-api.hh b/src/libstore/store-api.hh index 0c8a4db56..4d8db3596 100644 --- a/src/libstore/store-api.hh +++ b/src/libstore/store-api.hh @@ -1,5 +1,6 @@ #pragma once +#include "nar-info.hh" #include "realisation.hh" #include "path.hh" #include "derived-path.hh" @@ -13,6 +14,7 @@ #include "path-info.hh" #include "repair-flag.hh" +#include <nlohmann/json_fwd.hpp> #include <atomic> #include <limits> #include <map> @@ -67,7 +69,9 @@ struct Derivation; class FSAccessor; class NarInfoDiskCache; class Store; -class JSONPlaceholder; + + +typedef std::map<std::string, StorePath> OutputPathMap; enum CheckSigsFlag : bool { NoCheckSigs = false, CheckSigs = true }; @@ -83,6 +87,8 @@ enum BuildMode { bmNormal, bmRepair, bmCheck }; struct BuildResult; +typedef std::map<StorePath, std::optional<ContentAddress>> StorePathCAMap; + struct StoreConfig : public Config { using Config::Config; @@ -119,6 +125,8 @@ public: typedef std::map<std::string, std::string> Params; + + protected: struct PathInfoCacheValue { @@ -178,7 +186,7 @@ public: /* Return true if ‘path’ is in the Nix store (but not the Nix store itself). */ - bool isInStore(const Path & path) const; + bool isInStore(PathView path) const; /* Return true if ‘path’ is a store path, i.e. a direct child of the Nix store. */ @@ -186,7 +194,7 @@ public: /* Split a path like /nix/store/<hash>-<name>/<bla> into /nix/store/<hash>-<name> and /<bla>. */ - std::pair<StorePath, Path> toStorePath(const Path & path) const; + std::pair<StorePath, Path> toStorePath(PathView path) const; /* Follow symlinks until we end up with a path in the Nix store. */ Path followLinksToStore(std::string_view path) const; @@ -359,12 +367,22 @@ public: virtual void addToStore(const ValidPathInfo & info, Source & narSource, RepairFlag repair = NoRepair, CheckSigsFlag checkSigs = CheckSigs) = 0; + // A list of paths infos along with a source providing the content of the + // associated store path + using PathsSource = std::vector<std::pair<ValidPathInfo, std::unique_ptr<Source>>>; + /* Import multiple paths into the store. */ virtual void addMultipleToStore( Source & source, RepairFlag repair = NoRepair, CheckSigsFlag checkSigs = CheckSigs); + virtual void addMultipleToStore( + PathsSource & pathsToCopy, + Activity & act, + RepairFlag repair = NoRepair, + CheckSigsFlag checkSigs = CheckSigs); + /* Copy the contents of a path to the store and register the validity the resulting path. The resulting path is returned. The function object `filter' can be used to exclude files (see @@ -501,7 +519,7 @@ public: variable elements such as the registration time are included. If ‘showClosureSize’ is true, the closure size of each path is included. */ - void pathInfoToJSON(JSONPlaceholder & jsonOut, const StorePathSet & storePaths, + nlohmann::json pathInfoToJSON(const StorePathSet & storePaths, bool includeImpureInfo, bool showClosureSize, Base hashBase = Base32, AllowInvalidFlag allowInvalid = DisallowInvalid); @@ -607,6 +625,13 @@ public: */ StorePathSet exportReferences(const StorePathSet & storePaths, const StorePathSet & inputPaths); + /** + * Given a store path, return the realisation actually used in the realisation of this path: + * - If the path is a content-addressed derivation, try to resolve it + * - Otherwise, find one of its derivers + */ + std::optional<StorePath> getBuildDerivationPath(const StorePath &); + /* Hack to allow long-running processes like hydra-queue-runner to occasionally flush their path info cache. */ void clearPathInfoCache() @@ -634,9 +659,6 @@ public: return toRealPath(printStorePath(storePath)); } - virtual void createUser(const std::string & userName, uid_t userId) - { } - /* * Synchronises the options of the client with those of the daemon * (a no-op when there’s no daemon) @@ -708,6 +730,11 @@ void copyClosure( void removeTempRoots(); +/* Resolve the derived path completely, failing if any derivation output + is unknown. */ +OutputPathMap resolveDerivedPath(Store &, const DerivedPath::Built &, Store * evalStore = nullptr); + + /* Return a Store object to access the Nix store denoted by ‘uri’ (slight misnomer...). Supported values are: diff --git a/src/libstore/tests/derivation.cc b/src/libstore/tests/derivation.cc new file mode 100644 index 000000000..12be8504d --- /dev/null +++ b/src/libstore/tests/derivation.cc @@ -0,0 +1,124 @@ +#include <nlohmann/json.hpp> +#include <gtest/gtest.h> + +#include "derivations.hh" + +#include "tests/libstore.hh" + +namespace nix { + +class DerivationTest : public LibStoreTest +{ +}; + +#define TEST_JSON(TYPE, NAME, STR, VAL, ...) \ + TEST_F(DerivationTest, TYPE ## _ ## NAME ## _to_json) { \ + using nlohmann::literals::operator "" _json; \ + ASSERT_EQ( \ + STR ## _json, \ + (TYPE { VAL }).toJSON(*store __VA_OPT__(,) __VA_ARGS__)); \ + } + +TEST_JSON(DerivationOutput, inputAddressed, + R"({ + "path": "/nix/store/c015dhfh5l0lp6wxyvdn7bmwhbbr6hr9-drv-name-output-name" + })", + (DerivationOutput::InputAddressed { + .path = store->parseStorePath("/nix/store/c015dhfh5l0lp6wxyvdn7bmwhbbr6hr9-drv-name-output-name"), + }), + "drv-name", "output-name") + +TEST_JSON(DerivationOutput, caFixed, + R"({ + "hashAlgo": "r:sha256", + "hash": "894517c9163c896ec31a2adbd33c0681fd5f45b2c0ef08a64c92a03fb97f390f", + "path": "/nix/store/c015dhfh5l0lp6wxyvdn7bmwhbbr6hr9-drv-name-output-name" + })", + (DerivationOutput::CAFixed { + .hash = { + .method = FileIngestionMethod::Recursive, + .hash = Hash::parseAnyPrefixed("sha256-iUUXyRY8iW7DGirb0zwGgf1fRbLA7wimTJKgP7l/OQ8="), + }, + }), + "drv-name", "output-name") + +TEST_JSON(DerivationOutput, caFloating, + R"({ + "hashAlgo": "r:sha256" + })", + (DerivationOutput::CAFloating { + .method = FileIngestionMethod::Recursive, + .hashType = htSHA256, + }), + "drv-name", "output-name") + +TEST_JSON(DerivationOutput, deferred, + R"({ })", + DerivationOutput::Deferred { }, + "drv-name", "output-name") + +TEST_JSON(DerivationOutput, impure, + R"({ + "hashAlgo": "r:sha256", + "impure": true + })", + (DerivationOutput::Impure { + .method = FileIngestionMethod::Recursive, + .hashType = htSHA256, + }), + "drv-name", "output-name") + +TEST_JSON(Derivation, impure, + R"({ + "inputSrcs": [ + "/nix/store/c015dhfh5l0lp6wxyvdn7bmwhbbr6hr9-dep1" + ], + "inputDrvs": { + "/nix/store/c015dhfh5l0lp6wxyvdn7bmwhbbr6hr9-dep2.drv": [ + "cat", + "dog" + ] + }, + "system": "wasm-sel4", + "builder": "foo", + "args": [ + "bar", + "baz" + ], + "env": { + "BIG_BAD": "WOLF" + }, + "outputs": {} + })", + ({ + Derivation drv; + drv.inputSrcs = { + store->parseStorePath("/nix/store/c015dhfh5l0lp6wxyvdn7bmwhbbr6hr9-dep1"), + }; + drv.inputDrvs = { + { + store->parseStorePath("/nix/store/c015dhfh5l0lp6wxyvdn7bmwhbbr6hr9-dep2.drv"), + { + "cat", + "dog", + }, + } + }; + drv.platform = "wasm-sel4"; + drv.builder = "foo"; + drv.args = { + "bar", + "baz", + }; + drv.env = { + { + "BIG_BAD", + "WOLF", + }, + }; + drv; + })) + +#undef TEST_JSON + +} diff --git a/src/libstore/tests/derived-path.cc b/src/libstore/tests/derived-path.cc new file mode 100644 index 000000000..d1ac2c5e7 --- /dev/null +++ b/src/libstore/tests/derived-path.cc @@ -0,0 +1,62 @@ +#include <regex> + +#include <nlohmann/json.hpp> +#include <gtest/gtest.h> +#include <rapidcheck/gtest.h> + +#include "tests/derived-path.hh" +#include "tests/libstore.hh" + +namespace rc { +using namespace nix; + +Gen<DerivedPath::Opaque> Arbitrary<DerivedPath::Opaque>::arbitrary() +{ + return gen::just(DerivedPath::Opaque { + .path = *gen::arbitrary<StorePath>(), + }); +} + +Gen<DerivedPath::Built> Arbitrary<DerivedPath::Built>::arbitrary() +{ + return gen::just(DerivedPath::Built { + .drvPath = *gen::arbitrary<StorePath>(), + .outputs = *gen::arbitrary<OutputsSpec>(), + }); +} + +Gen<DerivedPath> Arbitrary<DerivedPath>::arbitrary() +{ + switch (*gen::inRange<uint8_t>(0, 1)) { + case 0: + return gen::just<DerivedPath>(*gen::arbitrary<DerivedPath::Opaque>()); + default: + return gen::just<DerivedPath>(*gen::arbitrary<DerivedPath::Built>()); + } +} + +} + +namespace nix { + +class DerivedPathTest : public LibStoreTest +{ +}; + +// FIXME: `RC_GTEST_FIXTURE_PROP` isn't calling `SetUpTestSuite` because it is +// no a real fixture. +// +// See https://github.com/emil-e/rapidcheck/blob/master/doc/gtest.md#rc_gtest_fixture_propfixture-name-args +TEST_F(DerivedPathTest, force_init) +{ +} + +RC_GTEST_FIXTURE_PROP( + DerivedPathTest, + prop_round_rip, + (const DerivedPath & o)) +{ + RC_ASSERT(o == DerivedPath::parse(*store, o.to_string(*store))); +} + +} diff --git a/src/libstore/tests/derived-path.hh b/src/libstore/tests/derived-path.hh new file mode 100644 index 000000000..3bc812440 --- /dev/null +++ b/src/libstore/tests/derived-path.hh @@ -0,0 +1,28 @@ +#pragma once + +#include <rapidcheck/gen/Arbitrary.h> + +#include <derived-path.hh> + +#include "tests/path.hh" +#include "tests/outputs-spec.hh" + +namespace rc { +using namespace nix; + +template<> +struct Arbitrary<DerivedPath::Opaque> { + static Gen<DerivedPath::Opaque> arbitrary(); +}; + +template<> +struct Arbitrary<DerivedPath::Built> { + static Gen<DerivedPath::Built> arbitrary(); +}; + +template<> +struct Arbitrary<DerivedPath> { + static Gen<DerivedPath> arbitrary(); +}; + +} diff --git a/src/libstore/tests/libstore.hh b/src/libstore/tests/libstore.hh new file mode 100644 index 000000000..05397659b --- /dev/null +++ b/src/libstore/tests/libstore.hh @@ -0,0 +1,23 @@ +#include <gtest/gtest.h> +#include <gmock/gmock.h> + +#include "store-api.hh" + +namespace nix { + +class LibStoreTest : public ::testing::Test { + public: + static void SetUpTestSuite() { + initLibStore(); + } + + protected: + LibStoreTest() + : store(openStore("dummy://")) + { } + + ref<Store> store; +}; + + +} /* namespace nix */ diff --git a/src/libstore/tests/local.mk b/src/libstore/tests/local.mk index f74295d97..03becc7d1 100644 --- a/src/libstore/tests/local.mk +++ b/src/libstore/tests/local.mk @@ -1,6 +1,20 @@ -check: libstore-tests_RUN +check: libstore-tests-exe_RUN -programs += libstore-tests +programs += libstore-tests-exe + +libstore-tests-exe_NAME = libnixstore-tests + +libstore-tests-exe_DIR := $(d) + +libstore-tests-exe_INSTALL_DIR := + +libstore-tests-exe_LIBS = libstore-tests + +libstore-tests-exe_LDFLAGS := $(GTEST_LIBS) + +libraries += libstore-tests + +libstore-tests_NAME = libnixstore-tests libstore-tests_DIR := $(d) @@ -10,6 +24,6 @@ libstore-tests_SOURCES := $(wildcard $(d)/*.cc) libstore-tests_CXXFLAGS += -I src/libstore -I src/libutil -libstore-tests_LIBS = libstore libutil +libstore-tests_LIBS = libutil-tests libstore libutil -libstore-tests_LDFLAGS := $(GTEST_LIBS) +libstore-tests_LDFLAGS := -lrapidcheck $(GTEST_LIBS) diff --git a/src/libstore/tests/nar-info-disk-cache.cc b/src/libstore/tests/nar-info-disk-cache.cc new file mode 100644 index 000000000..b4bdb8329 --- /dev/null +++ b/src/libstore/tests/nar-info-disk-cache.cc @@ -0,0 +1,123 @@ +#include "nar-info-disk-cache.hh" + +#include <gtest/gtest.h> +#include <rapidcheck/gtest.h> +#include "sqlite.hh" +#include <sqlite3.h> + + +namespace nix { + +TEST(NarInfoDiskCacheImpl, create_and_read) { + // This is a large single test to avoid some setup overhead. + + int prio = 12345; + bool wantMassQuery = true; + + Path tmpDir = createTempDir(); + AutoDelete delTmpDir(tmpDir); + Path dbPath(tmpDir + "/test-narinfo-disk-cache.sqlite"); + + int savedId; + int barId; + SQLite db; + SQLiteStmt getIds; + + { + auto cache = getTestNarInfoDiskCache(dbPath); + + // Set up "background noise" and check that different caches receive different ids + { + auto bc1 = cache->createCache("https://bar", "/nix/storedir", wantMassQuery, prio); + auto bc2 = cache->createCache("https://xyz", "/nix/storedir", false, 12); + ASSERT_NE(bc1, bc2); + barId = bc1; + } + + // Check that the fields are saved and returned correctly. This does not test + // the select statement yet, because of in-memory caching. + savedId = cache->createCache("http://foo", "/nix/storedir", wantMassQuery, prio);; + { + auto r = cache->upToDateCacheExists("http://foo"); + ASSERT_TRUE(r); + ASSERT_EQ(r->priority, prio); + ASSERT_EQ(r->wantMassQuery, wantMassQuery); + ASSERT_EQ(savedId, r->id); + } + + // We're going to pay special attention to the id field because we had a bug + // that changed it. + db = SQLite(dbPath); + getIds.create(db, "select id from BinaryCaches where url = 'http://foo'"); + + { + auto q(getIds.use()); + ASSERT_TRUE(q.next()); + ASSERT_EQ(savedId, q.getInt(0)); + ASSERT_FALSE(q.next()); + } + + // Pretend that the caches are older, but keep one up to date, as "background noise" + db.exec("update BinaryCaches set timestamp = timestamp - 1 - 7 * 24 * 3600 where url <> 'https://xyz';"); + + // This shows that the in-memory cache works + { + auto r = cache->upToDateCacheExists("http://foo"); + ASSERT_TRUE(r); + ASSERT_EQ(r->priority, prio); + ASSERT_EQ(r->wantMassQuery, wantMassQuery); + } + } + + { + // We can't clear the in-memory cache, so we use a new cache object. This is + // more realistic anyway. + auto cache2 = getTestNarInfoDiskCache(dbPath); + + { + auto r = cache2->upToDateCacheExists("http://foo"); + ASSERT_FALSE(r); + } + + // "Update", same data, check that the id number is reused + cache2->createCache("http://foo", "/nix/storedir", wantMassQuery, prio); + + { + auto r = cache2->upToDateCacheExists("http://foo"); + ASSERT_TRUE(r); + ASSERT_EQ(r->priority, prio); + ASSERT_EQ(r->wantMassQuery, wantMassQuery); + ASSERT_EQ(r->id, savedId); + } + + { + auto q(getIds.use()); + ASSERT_TRUE(q.next()); + auto currentId = q.getInt(0); + ASSERT_FALSE(q.next()); + ASSERT_EQ(currentId, savedId); + } + + // Check that the fields can be modified, and the id remains the same + { + auto r0 = cache2->upToDateCacheExists("https://bar"); + ASSERT_FALSE(r0); + + cache2->createCache("https://bar", "/nix/storedir", !wantMassQuery, prio + 10); + auto r = cache2->upToDateCacheExists("https://bar"); + ASSERT_EQ(r->wantMassQuery, !wantMassQuery); + ASSERT_EQ(r->priority, prio + 10); + ASSERT_EQ(r->id, barId); + } + + // // Force update (no use case yet; we only retrieve cache metadata when stale based on timestamp) + // { + // cache2->createCache("https://bar", "/nix/storedir", wantMassQuery, prio + 20); + // auto r = cache2->upToDateCacheExists("https://bar"); + // ASSERT_EQ(r->wantMassQuery, wantMassQuery); + // ASSERT_EQ(r->priority, prio + 20); + // } + } +} + +} diff --git a/src/libstore/tests/outputs-spec.cc b/src/libstore/tests/outputs-spec.cc new file mode 100644 index 000000000..984d1d963 --- /dev/null +++ b/src/libstore/tests/outputs-spec.cc @@ -0,0 +1,233 @@ +#include "outputs-spec.hh" + +#include <nlohmann/json.hpp> +#include <gtest/gtest.h> +#include <rapidcheck/gtest.h> + +namespace nix { + +#ifndef NDEBUG +TEST(OutputsSpec, no_empty_names) { + ASSERT_DEATH(OutputsSpec::Names { std::set<std::string> { } }, ""); +} +#endif + +#define TEST_DONT_PARSE(NAME, STR) \ + TEST(OutputsSpec, bad_ ## NAME) { \ + std::optional OutputsSpecOpt = \ + OutputsSpec::parseOpt(STR); \ + ASSERT_FALSE(OutputsSpecOpt); \ + } + +TEST_DONT_PARSE(empty, "") +TEST_DONT_PARSE(garbage, "&*()") +TEST_DONT_PARSE(double_star, "**") +TEST_DONT_PARSE(star_first, "*,foo") +TEST_DONT_PARSE(star_second, "foo,*") + +#undef TEST_DONT_PARSE + +TEST(OutputsSpec, all) { + std::string_view str = "*"; + OutputsSpec expected = OutputsSpec::All { }; + ASSERT_EQ(OutputsSpec::parse(str), expected); + ASSERT_EQ(expected.to_string(), str); +} + +TEST(OutputsSpec, names_out) { + std::string_view str = "out"; + OutputsSpec expected = OutputsSpec::Names { "out" }; + ASSERT_EQ(OutputsSpec::parse(str), expected); + ASSERT_EQ(expected.to_string(), str); +} + +TEST(OutputsSpec, names_underscore) { + std::string_view str = "a_b"; + OutputsSpec expected = OutputsSpec::Names { "a_b" }; + ASSERT_EQ(OutputsSpec::parse(str), expected); + ASSERT_EQ(expected.to_string(), str); +} + +TEST(OutputsSpec, names_numberic) { + std::string_view str = "01"; + OutputsSpec expected = OutputsSpec::Names { "01" }; + ASSERT_EQ(OutputsSpec::parse(str), expected); + ASSERT_EQ(expected.to_string(), str); +} + +TEST(OutputsSpec, names_out_bin) { + OutputsSpec expected = OutputsSpec::Names { "out", "bin" }; + ASSERT_EQ(OutputsSpec::parse("out,bin"), expected); + // N.B. This normalization is OK. + ASSERT_EQ(expected.to_string(), "bin,out"); +} + +#define TEST_SUBSET(X, THIS, THAT) \ + X((OutputsSpec { THIS }).isSubsetOf(THAT)); + +TEST(OutputsSpec, subsets_all_all) { + TEST_SUBSET(ASSERT_TRUE, OutputsSpec::All { }, OutputsSpec::All { }); +} + +TEST(OutputsSpec, subsets_names_all) { + TEST_SUBSET(ASSERT_TRUE, OutputsSpec::Names { "a" }, OutputsSpec::All { }); +} + +TEST(OutputsSpec, subsets_names_names_eq) { + TEST_SUBSET(ASSERT_TRUE, OutputsSpec::Names { "a" }, OutputsSpec::Names { "a" }); +} + +TEST(OutputsSpec, subsets_names_names_noneq) { + TEST_SUBSET(ASSERT_TRUE, OutputsSpec::Names { "a" }, (OutputsSpec::Names { "a", "b" })); +} + +TEST(OutputsSpec, not_subsets_all_names) { + TEST_SUBSET(ASSERT_FALSE, OutputsSpec::All { }, OutputsSpec::Names { "a" }); +} + +TEST(OutputsSpec, not_subsets_names_names) { + TEST_SUBSET(ASSERT_FALSE, (OutputsSpec::Names { "a", "b" }), (OutputsSpec::Names { "a" })); +} + +#undef TEST_SUBSET + +#define TEST_UNION(RES, THIS, THAT) \ + ASSERT_EQ(OutputsSpec { RES }, (OutputsSpec { THIS }).union_(THAT)); + +TEST(OutputsSpec, union_all_all) { + TEST_UNION(OutputsSpec::All { }, OutputsSpec::All { }, OutputsSpec::All { }); +} + +TEST(OutputsSpec, union_all_names) { + TEST_UNION(OutputsSpec::All { }, OutputsSpec::All { }, OutputsSpec::Names { "a" }); +} + +TEST(OutputsSpec, union_names_all) { + TEST_UNION(OutputsSpec::All { }, OutputsSpec::Names { "a" }, OutputsSpec::All { }); +} + +TEST(OutputsSpec, union_names_names) { + TEST_UNION((OutputsSpec::Names { "a", "b" }), OutputsSpec::Names { "a" }, OutputsSpec::Names { "b" }); +} + +#undef TEST_UNION + +#define TEST_DONT_PARSE(NAME, STR) \ + TEST(ExtendedOutputsSpec, bad_ ## NAME) { \ + std::optional extendedOutputsSpecOpt = \ + ExtendedOutputsSpec::parseOpt(STR); \ + ASSERT_FALSE(extendedOutputsSpecOpt); \ + } + +TEST_DONT_PARSE(carot_empty, "^") +TEST_DONT_PARSE(prefix_carot_empty, "foo^") +TEST_DONT_PARSE(garbage, "^&*()") +TEST_DONT_PARSE(double_star, "^**") +TEST_DONT_PARSE(star_first, "^*,foo") +TEST_DONT_PARSE(star_second, "^foo,*") + +#undef TEST_DONT_PARSE + +TEST(ExtendedOutputsSpec, defeault) { + std::string_view str = "foo"; + auto [prefix, extendedOutputsSpec] = ExtendedOutputsSpec::parse(str); + ASSERT_EQ(prefix, "foo"); + ExtendedOutputsSpec expected = ExtendedOutputsSpec::Default { }; + ASSERT_EQ(extendedOutputsSpec, expected); + ASSERT_EQ(std::string { prefix } + expected.to_string(), str); +} + +TEST(ExtendedOutputsSpec, all) { + std::string_view str = "foo^*"; + auto [prefix, extendedOutputsSpec] = ExtendedOutputsSpec::parse(str); + ASSERT_EQ(prefix, "foo"); + ExtendedOutputsSpec expected = OutputsSpec::All { }; + ASSERT_EQ(extendedOutputsSpec, expected); + ASSERT_EQ(std::string { prefix } + expected.to_string(), str); +} + +TEST(ExtendedOutputsSpec, out) { + std::string_view str = "foo^out"; + auto [prefix, extendedOutputsSpec] = ExtendedOutputsSpec::parse(str); + ASSERT_EQ(prefix, "foo"); + ExtendedOutputsSpec expected = OutputsSpec::Names { "out" }; + ASSERT_EQ(extendedOutputsSpec, expected); + ASSERT_EQ(std::string { prefix } + expected.to_string(), str); +} + +TEST(ExtendedOutputsSpec, out_bin) { + auto [prefix, extendedOutputsSpec] = ExtendedOutputsSpec::parse("foo^out,bin"); + ASSERT_EQ(prefix, "foo"); + ExtendedOutputsSpec expected = OutputsSpec::Names { "out", "bin" }; + ASSERT_EQ(extendedOutputsSpec, expected); + ASSERT_EQ(std::string { prefix } + expected.to_string(), "foo^bin,out"); +} + +TEST(ExtendedOutputsSpec, many_carrot) { + auto [prefix, extendedOutputsSpec] = ExtendedOutputsSpec::parse("foo^bar^out,bin"); + ASSERT_EQ(prefix, "foo^bar"); + ExtendedOutputsSpec expected = OutputsSpec::Names { "out", "bin" }; + ASSERT_EQ(extendedOutputsSpec, expected); + ASSERT_EQ(std::string { prefix } + expected.to_string(), "foo^bar^bin,out"); +} + + +#define TEST_JSON(TYPE, NAME, STR, VAL) \ + \ + TEST(TYPE, NAME ## _to_json) { \ + using nlohmann::literals::operator "" _json; \ + ASSERT_EQ( \ + STR ## _json, \ + ((nlohmann::json) TYPE { VAL })); \ + } \ + \ + TEST(TYPE, NAME ## _from_json) { \ + using nlohmann::literals::operator "" _json; \ + ASSERT_EQ( \ + TYPE { VAL }, \ + (STR ## _json).get<TYPE>()); \ + } + +TEST_JSON(OutputsSpec, all, R"(["*"])", OutputsSpec::All { }) +TEST_JSON(OutputsSpec, name, R"(["a"])", OutputsSpec::Names { "a" }) +TEST_JSON(OutputsSpec, names, R"(["a","b"])", (OutputsSpec::Names { "a", "b" })) + +TEST_JSON(ExtendedOutputsSpec, def, R"(null)", ExtendedOutputsSpec::Default { }) +TEST_JSON(ExtendedOutputsSpec, all, R"(["*"])", ExtendedOutputsSpec::Explicit { OutputsSpec::All { } }) +TEST_JSON(ExtendedOutputsSpec, name, R"(["a"])", ExtendedOutputsSpec::Explicit { OutputsSpec::Names { "a" } }) +TEST_JSON(ExtendedOutputsSpec, names, R"(["a","b"])", (ExtendedOutputsSpec::Explicit { OutputsSpec::Names { "a", "b" } })) + +#undef TEST_JSON + +} + +namespace rc { +using namespace nix; + +Gen<OutputsSpec> Arbitrary<OutputsSpec>::arbitrary() +{ + switch (*gen::inRange<uint8_t>(0, 1)) { + case 0: + return gen::just((OutputsSpec) OutputsSpec::All { }); + default: + return gen::just((OutputsSpec) OutputsSpec::Names { + *gen::nonEmpty(gen::container<StringSet>(gen::map( + gen::arbitrary<StorePathName>(), + [](StorePathName n) { return n.name; }))), + }); + } +} + +} + +namespace nix { + +RC_GTEST_PROP( + OutputsSpec, + prop_round_rip, + (const OutputsSpec & o)) +{ + RC_ASSERT(o == OutputsSpec::parse(o.to_string())); +} + +} diff --git a/src/libstore/tests/outputs-spec.hh b/src/libstore/tests/outputs-spec.hh new file mode 100644 index 000000000..2d455c817 --- /dev/null +++ b/src/libstore/tests/outputs-spec.hh @@ -0,0 +1,17 @@ +#pragma once + +#include <rapidcheck/gen/Arbitrary.h> + +#include <outputs-spec.hh> + +#include <tests/path.hh> + +namespace rc { +using namespace nix; + +template<> +struct Arbitrary<OutputsSpec> { + static Gen<OutputsSpec> arbitrary(); +}; + +} diff --git a/src/libstore/tests/path-with-outputs.cc b/src/libstore/tests/path-with-outputs.cc deleted file mode 100644 index 350ea7ffd..000000000 --- a/src/libstore/tests/path-with-outputs.cc +++ /dev/null @@ -1,46 +0,0 @@ -#include "path-with-outputs.hh" - -#include <gtest/gtest.h> - -namespace nix { - -TEST(parseOutputsSpec, basic) -{ - { - auto [prefix, outputsSpec] = parseOutputsSpec("foo"); - ASSERT_EQ(prefix, "foo"); - ASSERT_TRUE(std::get_if<DefaultOutputs>(&outputsSpec)); - } - - { - auto [prefix, outputsSpec] = parseOutputsSpec("foo^*"); - ASSERT_EQ(prefix, "foo"); - ASSERT_TRUE(std::get_if<AllOutputs>(&outputsSpec)); - } - - { - auto [prefix, outputsSpec] = parseOutputsSpec("foo^out"); - ASSERT_EQ(prefix, "foo"); - ASSERT_TRUE(std::get<OutputNames>(outputsSpec) == OutputNames({"out"})); - } - - { - auto [prefix, outputsSpec] = parseOutputsSpec("foo^out,bin"); - ASSERT_EQ(prefix, "foo"); - ASSERT_TRUE(std::get<OutputNames>(outputsSpec) == OutputNames({"out", "bin"})); - } - - { - auto [prefix, outputsSpec] = parseOutputsSpec("foo^bar^out,bin"); - ASSERT_EQ(prefix, "foo^bar"); - ASSERT_TRUE(std::get<OutputNames>(outputsSpec) == OutputNames({"out", "bin"})); - } - - { - auto [prefix, outputsSpec] = parseOutputsSpec("foo^&*()"); - ASSERT_EQ(prefix, "foo^&*()"); - ASSERT_TRUE(std::get_if<DefaultOutputs>(&outputsSpec)); - } -} - -} diff --git a/src/libstore/tests/path.cc b/src/libstore/tests/path.cc new file mode 100644 index 000000000..430aa0099 --- /dev/null +++ b/src/libstore/tests/path.cc @@ -0,0 +1,153 @@ +#include <regex> + +#include <nlohmann/json.hpp> +#include <gtest/gtest.h> +#include <rapidcheck/gtest.h> + +#include "path-regex.hh" +#include "store-api.hh" + +#include "tests/hash.hh" +#include "tests/libstore.hh" +#include "tests/path.hh" + +namespace nix { + +#define STORE_DIR "/nix/store/" +#define HASH_PART "g1w7hy3qg1w7hy3qg1w7hy3qg1w7hy3q" + +class StorePathTest : public LibStoreTest +{ +}; + +static std::regex nameRegex { std::string { nameRegexStr } }; + +#define TEST_DONT_PARSE(NAME, STR) \ + TEST_F(StorePathTest, bad_ ## NAME) { \ + std::string_view str = \ + STORE_DIR HASH_PART "-" STR; \ + ASSERT_THROW( \ + store->parseStorePath(str), \ + BadStorePath); \ + std::string name { STR }; \ + EXPECT_FALSE(std::regex_match(name, nameRegex)); \ + } + +TEST_DONT_PARSE(empty, "") +TEST_DONT_PARSE(garbage, "&*()") +TEST_DONT_PARSE(double_star, "**") +TEST_DONT_PARSE(star_first, "*,foo") +TEST_DONT_PARSE(star_second, "foo,*") +TEST_DONT_PARSE(bang, "foo!o") + +#undef TEST_DONT_PARSE + +#define TEST_DO_PARSE(NAME, STR) \ + TEST_F(StorePathTest, good_ ## NAME) { \ + std::string_view str = \ + STORE_DIR HASH_PART "-" STR; \ + auto p = store->parseStorePath(str); \ + std::string name { p.name() }; \ + EXPECT_TRUE(std::regex_match(name, nameRegex)); \ + } + +// 0-9 a-z A-Z + - . _ ? = + +TEST_DO_PARSE(numbers, "02345") +TEST_DO_PARSE(lower_case, "foo") +TEST_DO_PARSE(upper_case, "FOO") +TEST_DO_PARSE(plus, "foo+bar") +TEST_DO_PARSE(dash, "foo-dev") +TEST_DO_PARSE(underscore, "foo_bar") +TEST_DO_PARSE(period, "foo.txt") +TEST_DO_PARSE(question_mark, "foo?why") +TEST_DO_PARSE(equals_sign, "foo=foo") + +#undef TEST_DO_PARSE + +// For rapidcheck +void showValue(const StorePath & p, std::ostream & os) { + os << p.to_string(); +} + +} + +namespace rc { +using namespace nix; + +Gen<StorePathName> Arbitrary<StorePathName>::arbitrary() +{ + auto len = *gen::inRange<size_t>( + 1, + StorePath::MaxPathLen - std::string_view { HASH_PART }.size()); + + std::string pre; + pre.reserve(len); + + for (size_t c = 0; c < len; ++c) { + switch (auto i = *gen::inRange<uint8_t>(0, 10 + 2 * 26 + 6)) { + case 0 ... 9: + pre += '0' + i; + case 10 ... 35: + pre += 'A' + (i - 10); + break; + case 36 ... 61: + pre += 'a' + (i - 36); + break; + case 62: + pre += '+'; + break; + case 63: + pre += '-'; + break; + case 64: + pre += '.'; + break; + case 65: + pre += '_'; + break; + case 66: + pre += '?'; + break; + case 67: + pre += '='; + break; + default: + assert(false); + } + } + + return gen::just(StorePathName { + .name = std::move(pre), + }); +} + +Gen<StorePath> Arbitrary<StorePath>::arbitrary() +{ + return gen::just(StorePath { + *gen::arbitrary<Hash>(), + (*gen::arbitrary<StorePathName>()).name, + }); +} + +} // namespace rc + +namespace nix { + +RC_GTEST_FIXTURE_PROP( + StorePathTest, + prop_regex_accept, + (const StorePath & p)) +{ + RC_ASSERT(std::regex_match(std::string { p.name() }, nameRegex)); +} + +RC_GTEST_FIXTURE_PROP( + StorePathTest, + prop_round_rip, + (const StorePath & p)) +{ + RC_ASSERT(p == store->parseStorePath(store->printStorePath(p))); +} + +} diff --git a/src/libstore/tests/path.hh b/src/libstore/tests/path.hh new file mode 100644 index 000000000..d7f1a8988 --- /dev/null +++ b/src/libstore/tests/path.hh @@ -0,0 +1,28 @@ +#pragma once + +#include <rapidcheck/gen/Arbitrary.h> + +#include <path.hh> + +namespace nix { + +struct StorePathName { + std::string name; +}; + +} + +namespace rc { +using namespace nix; + +template<> +struct Arbitrary<StorePathName> { + static Gen<StorePathName> arbitrary(); +}; + +template<> +struct Arbitrary<StorePath> { + static Gen<StorePath> arbitrary(); +}; + +} diff --git a/src/libstore/uds-remote-store.hh b/src/libstore/uds-remote-store.hh index f8dfcca70..d31a4d592 100644 --- a/src/libstore/uds-remote-store.hh +++ b/src/libstore/uds-remote-store.hh @@ -29,9 +29,6 @@ public: static std::set<std::string> uriSchemes() { return {"unix"}; } - bool sameMachine() override - { return true; } - ref<FSAccessor> getFSAccessor() override { return LocalFSStore::getFSAccessor(); } diff --git a/src/libutil/archive.cc b/src/libutil/archive.cc index 30b471af5..268a798d9 100644 --- a/src/libutil/archive.cc +++ b/src/libutil/archive.cc @@ -35,10 +35,6 @@ static ArchiveSettings archiveSettings; static GlobalConfig::Register rArchiveSettings(&archiveSettings); -const std::string narVersionMagic1 = "nix-archive-1"; - -static std::string caseHackSuffix = "~nix~case~hack~"; - PathFilter defaultPathFilter = [](const Path &) { return true; }; @@ -91,7 +87,7 @@ static time_t dump(const Path & path, Sink & sink, PathFilter & filter) std::string name(i.name); size_t pos = i.name.find(caseHackSuffix); if (pos != std::string::npos) { - debug(format("removing case hack suffix from '%1%'") % (path + "/" + i.name)); + debug("removing case hack suffix from '%1%'", path + "/" + i.name); name.erase(pos); } if (!unhacked.emplace(name, i.name).second) @@ -234,6 +230,7 @@ static void parse(ParseSink & sink, Source & source, const Path & path) else if (s == "contents" && type == tpRegular) { parseContents(sink, source, path); + sink.closeRegularFile(); } else if (s == "executable" && type == tpRegular) { @@ -265,7 +262,7 @@ static void parse(ParseSink & sink, Source & source, const Path & path) if (archiveSettings.useCaseHack) { auto i = names.find(name); if (i != names.end()) { - debug(format("case collision between '%1%' and '%2%'") % i->first % name); + debug("case collision between '%1%' and '%2%'", i->first, name); name += caseHackSuffix; name += std::to_string(++i->second); } else @@ -324,6 +321,12 @@ struct RestoreSink : ParseSink if (!fd) throw SysError("creating file '%1%'", p); } + void closeRegularFile() override + { + /* Call close explicitly to make sure the error is checked */ + fd.close(); + } + void isExecutable() override { struct stat st; diff --git a/src/libutil/archive.hh b/src/libutil/archive.hh index 79ce08df0..e42dea540 100644 --- a/src/libutil/archive.hh +++ b/src/libutil/archive.hh @@ -60,6 +60,7 @@ struct ParseSink virtual void createDirectory(const Path & path) { }; virtual void createRegularFile(const Path & path) { }; + virtual void closeRegularFile() { }; virtual void isExecutable() { }; virtual void preallocateContents(uint64_t size) { }; virtual void receiveContents(std::string_view data) { }; @@ -102,7 +103,9 @@ void copyNAR(Source & source, Sink & sink); void copyPath(const Path & from, const Path & to); -extern const std::string narVersionMagic1; +inline constexpr std::string_view narVersionMagic1 = "nix-archive-1"; + +inline constexpr std::string_view caseHackSuffix = "~nix~case~hack~"; } diff --git a/src/libutil/args.cc b/src/libutil/args.cc index 44b63f0f6..35686a8aa 100644 --- a/src/libutil/args.cc +++ b/src/libutil/args.cc @@ -29,7 +29,15 @@ void Args::removeFlag(const std::string & longName) void Completions::add(std::string completion, std::string description) { - assert(description.find('\n') == std::string::npos); + description = trim(description); + // ellipsize overflowing content on the back of the description + auto end_index = description.find_first_of(".\n"); + if (end_index != std::string::npos) { + auto needs_ellipsis = end_index != description.size() - 1; + description.resize(end_index); + if (needs_ellipsis) + description.append(" [...]"); + } insert(Completion { .completion = completion, .description = description @@ -216,7 +224,7 @@ nlohmann::json Args::toJSON() if (flag->shortName) j["shortName"] = std::string(1, flag->shortName); if (flag->description != "") - j["description"] = flag->description; + j["description"] = trim(flag->description); j["category"] = flag->category; if (flag->handler.arity != ArityAny) j["arity"] = flag->handler.arity; @@ -237,7 +245,7 @@ nlohmann::json Args::toJSON() } auto res = nlohmann::json::object(); - res["description"] = description(); + res["description"] = trim(description()); res["flags"] = std::move(flags); res["args"] = std::move(args); auto s = doc(); @@ -324,7 +332,7 @@ MultiCommand::MultiCommand(const Commands & commands_) expectArgs({ .label = "subcommand", .optional = true, - .handler = {[=](std::string s) { + .handler = {[=,this](std::string s) { assert(!command); auto i = commands.find(s); if (i == commands.end()) { @@ -379,7 +387,7 @@ nlohmann::json MultiCommand::toJSON() auto j = command->toJSON(); auto cat = nlohmann::json::object(); cat["id"] = command->category(); - cat["description"] = categories[command->category()]; + cat["description"] = trim(categories[command->category()]); j["category"] = std::move(cat); cmds[name] = std::move(j); } diff --git a/src/libutil/args.hh b/src/libutil/args.hh index 84866f12b..7211ee307 100644 --- a/src/libutil/args.hh +++ b/src/libutil/args.hh @@ -198,7 +198,6 @@ struct Command : virtual public Args virtual ~Command() { } - virtual void prepare() { }; virtual void run() = 0; typedef int Category; diff --git a/src/libutil/canon-path.cc b/src/libutil/canon-path.cc new file mode 100644 index 000000000..b132b4262 --- /dev/null +++ b/src/libutil/canon-path.cc @@ -0,0 +1,103 @@ +#include "canon-path.hh" +#include "util.hh" + +namespace nix { + +CanonPath CanonPath::root = CanonPath("/"); + +CanonPath::CanonPath(std::string_view raw) + : path(absPath((Path) raw, "/")) +{ } + +CanonPath::CanonPath(std::string_view raw, const CanonPath & root) + : path(absPath((Path) raw, root.abs())) +{ } + +std::optional<CanonPath> CanonPath::parent() const +{ + if (isRoot()) return std::nullopt; + return CanonPath(unchecked_t(), path.substr(0, std::max((size_t) 1, path.rfind('/')))); +} + +void CanonPath::pop() +{ + assert(!isRoot()); + path.resize(std::max((size_t) 1, path.rfind('/'))); +} + +bool CanonPath::isWithin(const CanonPath & parent) const +{ + return !( + path.size() < parent.path.size() + || path.substr(0, parent.path.size()) != parent.path + || (parent.path.size() > 1 && path.size() > parent.path.size() + && path[parent.path.size()] != '/')); +} + +CanonPath CanonPath::removePrefix(const CanonPath & prefix) const +{ + assert(isWithin(prefix)); + if (prefix.isRoot()) return *this; + if (path.size() == prefix.path.size()) return root; + return CanonPath(unchecked_t(), path.substr(prefix.path.size())); +} + +void CanonPath::extend(const CanonPath & x) +{ + if (x.isRoot()) return; + if (isRoot()) + path += x.rel(); + else + path += x.abs(); +} + +CanonPath CanonPath::operator + (const CanonPath & x) const +{ + auto res = *this; + res.extend(x); + return res; +} + +void CanonPath::push(std::string_view c) +{ + assert(c.find('/') == c.npos); + assert(c != "." && c != ".."); + if (!isRoot()) path += '/'; + path += c; +} + +CanonPath CanonPath::operator + (std::string_view c) const +{ + auto res = *this; + res.push(c); + return res; +} + +bool CanonPath::isAllowed(const std::set<CanonPath> & allowed) const +{ + /* Check if `this` is an exact match or the parent of an + allowed path. */ + auto lb = allowed.lower_bound(*this); + if (lb != allowed.end()) { + if (lb->isWithin(*this)) + return true; + } + + /* Check if a parent of `this` is allowed. */ + auto path = *this; + while (!path.isRoot()) { + path.pop(); + if (allowed.count(path)) + return true; + } + + return false; +} + +std::ostream & operator << (std::ostream & stream, const CanonPath & path) +{ + stream << path.abs(); + return stream; +} + +} diff --git a/src/libutil/canon-path.hh b/src/libutil/canon-path.hh new file mode 100644 index 000000000..9d5984584 --- /dev/null +++ b/src/libutil/canon-path.hh @@ -0,0 +1,173 @@ +#pragma once + +#include <string> +#include <optional> +#include <cassert> +#include <iostream> +#include <set> + +namespace nix { + +/* A canonical representation of a path. It ensures the following: + + - It always starts with a slash. + + - It never ends with a slash, except if the path is "/". + + - A slash is never followed by a slash (i.e. no empty components). + + - There are no components equal to '.' or '..'. + + Note that the path does not need to correspond to an actually + existing path, and there is no guarantee that symlinks are + resolved. +*/ +class CanonPath +{ + std::string path; + +public: + + /* Construct a canon path from a non-canonical path. Any '.', '..' + or empty components are removed. */ + CanonPath(std::string_view raw); + + explicit CanonPath(const char * raw) + : CanonPath(std::string_view(raw)) + { } + + struct unchecked_t { }; + + CanonPath(unchecked_t _, std::string path) + : path(std::move(path)) + { } + + static CanonPath root; + + /* If `raw` starts with a slash, return + `CanonPath(raw)`. Otherwise return a `CanonPath` representing + `root + "/" + raw`. */ + CanonPath(std::string_view raw, const CanonPath & root); + + bool isRoot() const + { return path.size() <= 1; } + + explicit operator std::string_view() const + { return path; } + + const std::string & abs() const + { return path; } + + /* Like abs(), but return an empty string if this path is + '/'. Thus the returned string never ends in a slash. */ + const std::string & absOrEmpty() const + { + const static std::string epsilon; + return isRoot() ? epsilon : path; + } + + const char * c_str() const + { return path.c_str(); } + + std::string_view rel() const + { return ((std::string_view) path).substr(1); } + + struct Iterator + { + std::string_view remaining; + size_t slash; + + Iterator(std::string_view remaining) + : remaining(remaining) + , slash(remaining.find('/')) + { } + + bool operator != (const Iterator & x) const + { return remaining.data() != x.remaining.data(); } + + const std::string_view operator * () const + { return remaining.substr(0, slash); } + + void operator ++ () + { + if (slash == remaining.npos) + remaining = remaining.substr(remaining.size()); + else { + remaining = remaining.substr(slash + 1); + slash = remaining.find('/'); + } + } + }; + + Iterator begin() const { return Iterator(rel()); } + Iterator end() const { return Iterator(rel().substr(path.size() - 1)); } + + std::optional<CanonPath> parent() const; + + /* Remove the last component. Panics if this path is the root. */ + void pop(); + + std::optional<std::string_view> dirOf() const + { + if (isRoot()) return std::nullopt; + return ((std::string_view) path).substr(0, path.rfind('/')); + } + + std::optional<std::string_view> baseName() const + { + if (isRoot()) return std::nullopt; + return ((std::string_view) path).substr(path.rfind('/') + 1); + } + + bool operator == (const CanonPath & x) const + { return path == x.path; } + + bool operator != (const CanonPath & x) const + { return path != x.path; } + + /* Compare paths lexicographically except that path separators + are sorted before any other character. That is, in the sorted order + a directory is always followed directly by its children. For + instance, 'foo' < 'foo/bar' < 'foo!'. */ + bool operator < (const CanonPath & x) const + { + auto i = path.begin(); + auto j = x.path.begin(); + for ( ; i != path.end() && j != x.path.end(); ++i, ++j) { + auto c_i = *i; + if (c_i == '/') c_i = 0; + auto c_j = *j; + if (c_j == '/') c_j = 0; + if (c_i < c_j) return true; + if (c_i > c_j) return false; + } + return i == path.end() && j != x.path.end(); + } + + /* Return true if `this` is equal to `parent` or a child of + `parent`. */ + bool isWithin(const CanonPath & parent) const; + + CanonPath removePrefix(const CanonPath & prefix) const; + + /* Append another path to this one. */ + void extend(const CanonPath & x); + + /* Concatenate two paths. */ + CanonPath operator + (const CanonPath & x) const; + + /* Add a path component to this one. It must not contain any slashes. */ + void push(std::string_view c); + + CanonPath operator + (std::string_view c) const; + + /* Check whether access to this path is allowed, which is the case + if 1) `this` is within any of the `allowed` paths; or 2) any of + the `allowed` paths are within `this`. (The latter condition + ensures access to the parents of allowed paths.) */ + bool isAllowed(const std::set<CanonPath> & allowed) const; +}; + +std::ostream & operator << (std::ostream & stream, const CanonPath & path); + +} diff --git a/src/libutil/cgroup.cc b/src/libutil/cgroup.cc new file mode 100644 index 000000000..a008481ca --- /dev/null +++ b/src/libutil/cgroup.cc @@ -0,0 +1,148 @@ +#if __linux__ + +#include "cgroup.hh" +#include "util.hh" +#include "finally.hh" + +#include <chrono> +#include <cmath> +#include <regex> +#include <unordered_set> +#include <thread> + +#include <dirent.h> +#include <mntent.h> + +namespace nix { + +std::optional<Path> getCgroupFS() +{ + static auto res = [&]() -> std::optional<Path> { + auto fp = fopen("/proc/mounts", "r"); + if (!fp) return std::nullopt; + Finally delFP = [&]() { fclose(fp); }; + while (auto ent = getmntent(fp)) + if (std::string_view(ent->mnt_type) == "cgroup2") + return ent->mnt_dir; + + return std::nullopt; + }(); + return res; +} + +// FIXME: obsolete, check for cgroup2 +std::map<std::string, std::string> getCgroups(const Path & cgroupFile) +{ + std::map<std::string, std::string> cgroups; + + for (auto & line : tokenizeString<std::vector<std::string>>(readFile(cgroupFile), "\n")) { + static std::regex regex("([0-9]+):([^:]*):(.*)"); + std::smatch match; + if (!std::regex_match(line, match, regex)) + throw Error("invalid line '%s' in '%s'", line, cgroupFile); + + std::string name = hasPrefix(std::string(match[2]), "name=") ? std::string(match[2], 5) : match[2]; + cgroups.insert_or_assign(name, match[3]); + } + + return cgroups; +} + +static CgroupStats destroyCgroup(const Path & cgroup, bool returnStats) +{ + if (!pathExists(cgroup)) return {}; + + auto procsFile = cgroup + "/cgroup.procs"; + + if (!pathExists(procsFile)) + throw Error("'%s' is not a cgroup", cgroup); + + /* Use the fast way to kill every process in a cgroup, if + available. */ + auto killFile = cgroup + "/cgroup.kill"; + if (pathExists(killFile)) + writeFile(killFile, "1"); + + /* Otherwise, manually kill every process in the subcgroups and + this cgroup. */ + for (auto & entry : readDirectory(cgroup)) { + if (entry.type != DT_DIR) continue; + destroyCgroup(cgroup + "/" + entry.name, false); + } + + int round = 1; + + std::unordered_set<pid_t> pidsShown; + + while (true) { + auto pids = tokenizeString<std::vector<std::string>>(readFile(procsFile)); + + if (pids.empty()) break; + + if (round > 20) + throw Error("cannot kill cgroup '%s'", cgroup); + + for (auto & pid_s : pids) { + pid_t pid; + if (auto o = string2Int<pid_t>(pid_s)) + pid = *o; + else + throw Error("invalid pid '%s'", pid); + if (pidsShown.insert(pid).second) { + try { + auto cmdline = readFile(fmt("/proc/%d/cmdline", pid)); + using namespace std::string_literals; + warn("killing stray builder process %d (%s)...", + pid, trim(replaceStrings(cmdline, "\0"s, " "))); + } catch (SysError &) { + } + } + // FIXME: pid wraparound + if (kill(pid, SIGKILL) == -1 && errno != ESRCH) + throw SysError("killing member %d of cgroup '%s'", pid, cgroup); + } + + auto sleep = std::chrono::milliseconds((int) std::pow(2.0, std::min(round, 10))); + if (sleep.count() > 100) + printError("waiting for %d ms for cgroup '%s' to become empty", sleep.count(), cgroup); + std::this_thread::sleep_for(sleep); + round++; + } + + CgroupStats stats; + + if (returnStats) { + auto cpustatPath = cgroup + "/cpu.stat"; + + if (pathExists(cpustatPath)) { + for (auto & line : tokenizeString<std::vector<std::string>>(readFile(cpustatPath), "\n")) { + std::string_view userPrefix = "user_usec "; + if (hasPrefix(line, userPrefix)) { + auto n = string2Int<uint64_t>(line.substr(userPrefix.size())); + if (n) stats.cpuUser = std::chrono::microseconds(*n); + } + + std::string_view systemPrefix = "system_usec "; + if (hasPrefix(line, systemPrefix)) { + auto n = string2Int<uint64_t>(line.substr(systemPrefix.size())); + if (n) stats.cpuSystem = std::chrono::microseconds(*n); + } + } + } + + } + + if (rmdir(cgroup.c_str()) == -1) + throw SysError("deleting cgroup '%s'", cgroup); + + return stats; +} + +CgroupStats destroyCgroup(const Path & cgroup) +{ + return destroyCgroup(cgroup, true); +} + +} + +#endif diff --git a/src/libutil/cgroup.hh b/src/libutil/cgroup.hh new file mode 100644 index 000000000..d08c8ad29 --- /dev/null +++ b/src/libutil/cgroup.hh @@ -0,0 +1,29 @@ +#pragma once + +#if __linux__ + +#include <chrono> +#include <optional> + +#include "types.hh" + +namespace nix { + +std::optional<Path> getCgroupFS(); + +std::map<std::string, std::string> getCgroups(const Path & cgroupFile); + +struct CgroupStats +{ + std::optional<std::chrono::microseconds> cpuUser, cpuSystem; +}; + +/* Destroy the cgroup denoted by 'path'. The postcondition is that + 'path' does not exist, and thus any processes in the cgroup have + been killed. Also return statistics from the cgroup just before + destruction. */ +CgroupStats destroyCgroup(const Path & cgroup); + +} + +#endif diff --git a/src/libutil/config.cc b/src/libutil/config.cc index 9bb412b4f..b349f2d80 100644 --- a/src/libutil/config.cc +++ b/src/libutil/config.cc @@ -209,7 +209,7 @@ void BaseSetting<T>::convertToArg(Args & args, const std::string & category) .description = fmt("Set the `%s` setting.", name), .category = category, .labels = {"value"}, - .handler = {[=](std::string s) { overridden = true; set(s); }}, + .handler = {[this](std::string s) { overridden = true; set(s); }}, }); if (isAppendable()) @@ -218,7 +218,7 @@ void BaseSetting<T>::convertToArg(Args & args, const std::string & category) .description = fmt("Append to the `%s` setting.", name), .category = category, .labels = {"value"}, - .handler = {[=](std::string s) { overridden = true; set(s, true); }}, + .handler = {[this](std::string s) { overridden = true; set(s, true); }}, }); } @@ -270,13 +270,13 @@ template<> void BaseSetting<bool>::convertToArg(Args & args, const std::string & .longName = name, .description = fmt("Enable the `%s` setting.", name), .category = category, - .handler = {[=]() { override(true); }} + .handler = {[this]() { override(true); }} }); args.addFlag({ .longName = "no-" + name, .description = fmt("Disable the `%s` setting.", name), .category = category, - .handler = {[=]() { override(false); }} + .handler = {[this]() { override(false); }} }); } diff --git a/src/libutil/config.hh b/src/libutil/config.hh index 79ec0f9cf..7ac43c854 100644 --- a/src/libutil/config.hh +++ b/src/libutil/config.hh @@ -250,11 +250,15 @@ public: operator const T &() const { return value; } operator T &() { return value; } const T & get() const { return value; } - bool operator ==(const T & v2) const { return value == v2; } - bool operator !=(const T & v2) const { return value != v2; } - void operator =(const T & v) { assign(v); } + template<typename U> + bool operator ==(const U & v2) const { return value == v2; } + template<typename U> + bool operator !=(const U & v2) const { return value != v2; } + template<typename U> + void operator =(const U & v) { assign(v); } virtual void assign(const T & v) { value = v; } - void setDefault(const T & v) { if (!overridden) value = v; } + template<typename U> + void setDefault(const U & v) { if (!overridden) value = v; } void set(const std::string & str, bool append = false) override; diff --git a/src/libutil/error.cc b/src/libutil/error.cc index 9172f67a6..e4f0d4677 100644 --- a/src/libutil/error.cc +++ b/src/libutil/error.cc @@ -9,9 +9,9 @@ namespace nix { const std::string nativeSystem = SYSTEM; -void BaseError::addTrace(std::optional<ErrPos> e, hintformat hint) +void BaseError::addTrace(std::shared_ptr<AbstractPos> && e, hintformat hint, bool frame) { - err.traces.push_front(Trace { .pos = e, .hint = hint }); + err.traces.push_front(Trace { .pos = std::move(e), .hint = hint, .frame = frame }); } // c++ std::exception descendants must have a 'const char* what()' function. @@ -30,91 +30,46 @@ const std::string & BaseError::calcWhat() const std::optional<std::string> ErrorInfo::programName = std::nullopt; -std::ostream & operator<<(std::ostream & os, const hintformat & hf) +std::ostream & operator <<(std::ostream & os, const hintformat & hf) { return os << hf.str(); } -std::string showErrPos(const ErrPos & errPos) +std::ostream & operator <<(std::ostream & str, const AbstractPos & pos) { - if (errPos.line > 0) { - if (errPos.column > 0) { - return fmt("%d:%d", errPos.line, errPos.column); - } else { - return fmt("%d", errPos.line); - } - } - else { - return ""; - } + pos.print(str); + str << ":" << pos.line; + if (pos.column > 0) + str << ":" << pos.column; + return str; } -std::optional<LinesOfCode> getCodeLines(const ErrPos & errPos) +std::optional<LinesOfCode> AbstractPos::getCodeLines() const { - if (errPos.line <= 0) + if (line == 0) return std::nullopt; - if (errPos.origin == foFile) { - LinesOfCode loc; - try { - // FIXME: when running as the daemon, make sure we don't - // open a file to which the client doesn't have access. - AutoCloseFD fd = open(errPos.file.c_str(), O_RDONLY | O_CLOEXEC); - if (!fd) return {}; - - // count the newlines. - int count = 0; - std::string line; - int pl = errPos.line - 1; - do - { - line = readLine(fd.get()); - ++count; - if (count < pl) - ; - else if (count == pl) - loc.prevLineOfCode = line; - else if (count == pl + 1) - loc.errLineOfCode = line; - else if (count == pl + 2) { - loc.nextLineOfCode = line; - break; - } - } while (true); - return loc; - } - catch (EndOfFile & eof) { - if (loc.errLineOfCode.has_value()) - return loc; - else - return std::nullopt; - } - catch (std::exception & e) { - return std::nullopt; - } - } else { - std::istringstream iss(errPos.file); + if (auto source = getSource()) { + + std::istringstream iss(*source); // count the newlines. int count = 0; - std::string line; - int pl = errPos.line - 1; + std::string curLine; + int pl = line - 1; LinesOfCode loc; - do - { - std::getline(iss, line); + do { + std::getline(iss, curLine); ++count; if (count < pl) - { ; - } else if (count == pl) { - loc.prevLineOfCode = line; + loc.prevLineOfCode = curLine; } else if (count == pl + 1) { - loc.errLineOfCode = line; + loc.errLineOfCode = curLine; } else if (count == pl + 2) { - loc.nextLineOfCode = line; + loc.nextLineOfCode = curLine; break; } @@ -124,12 +79,14 @@ std::optional<LinesOfCode> getCodeLines(const ErrPos & errPos) return loc; } + + return std::nullopt; } // print lines of code to the ostream, indicating the error column. void printCodeLines(std::ostream & out, const std::string & prefix, - const ErrPos & errPos, + const AbstractPos & errPos, const LinesOfCode & loc) { // previous line of code. @@ -176,28 +133,6 @@ void printCodeLines(std::ostream & out, } } -void printAtPos(const ErrPos & pos, std::ostream & out) -{ - if (pos) { - switch (pos.origin) { - case foFile: { - out << fmt(ANSI_BLUE "at " ANSI_WARNING "%s:%s" ANSI_NORMAL ":", pos.file, showErrPos(pos)); - break; - } - case foString: { - out << fmt(ANSI_BLUE "at " ANSI_WARNING "«string»:%s" ANSI_NORMAL ":", showErrPos(pos)); - break; - } - case foStdin: { - out << fmt(ANSI_BLUE "at " ANSI_WARNING "«stdin»:%s" ANSI_NORMAL ":", showErrPos(pos)); - break; - } - default: - throw Error("invalid FileOrigin in errPos"); - } - } -} - static std::string indent(std::string_view indentFirst, std::string_view indentRest, std::string_view s) { std::string res; @@ -262,49 +197,160 @@ std::ostream & showErrorInfo(std::ostream & out, const ErrorInfo & einfo, bool s prefix += ":" ANSI_NORMAL " "; std::ostringstream oss; - oss << einfo.msg << "\n"; - if (einfo.errPos.has_value() && *einfo.errPos) { - oss << "\n"; - printAtPos(*einfo.errPos, oss); + auto noSource = ANSI_ITALIC " (source not available)" ANSI_NORMAL "\n"; + + /* + * Traces + * ------ + * + * The semantics of traces is a bit weird. We have only one option to + * print them and to make them verbose (--show-trace). In the code they + * are always collected, but they are not printed by default. The code + * also collects more traces when the option is on. This means that there + * is no way to print the simplified traces at all. + * + * I (layus) designed the code to attach positions to a restricted set of + * messages. This means that we have a lot of traces with no position at + * all, including most of the base error messages. For example "type + * error: found a string while a set was expected" has no position, but + * will come with several traces detailing it's precise relation to the + * closest know position. This makes erroring without printing traces + * quite useless. + * + * This is why I introduced the idea to always print a few traces on + * error. The number 3 is quite arbitrary, and was selected so as not to + * clutter the console on error. For the same reason, a trace with an + * error position takes more space, and counts as two traces towards the + * limit. + * + * The rest is truncated, unless --show-trace is passed. This preserves + * the same bad semantics of --show-trace to both show the trace and + * augment it with new data. Not too sure what is the best course of + * action. + * + * The issue is that it is fundamentally hard to provide a trace for a + * lazy language. The trace will only cover the current spine of the + * evaluation, missing things that have been evaluated before. For + * example, most type errors are hard to inspect because there is not + * trace for the faulty value. These errors should really print the faulty + * value itself. + * + * In function calls, the --show-trace flag triggers extra traces for each + * function invocation. These work as scopes, allowing to follow the + * current spine of the evaluation graph. Without that flag, the error + * trace should restrict itself to a restricted prefix of that trace, + * until the first scope. If we ever get to such a precise error + * reporting, there would be no need to add an arbitrary limit here. We + * could always print the full trace, and it would just be small without + * the flag. + * + * One idea I had is for XxxError.addTrace() to perform nothing if one + * scope has already been traced. Alternatively, we could stop here when + * we encounter such a scope instead of after an arbitrary number of + * traces. This however requires to augment traces with the notion of + * "scope". + * + * This is particularly visible in code like evalAttrs(...) where we have + * to make a decision between the two following options. + * + * ``` long traces + * inline void EvalState::evalAttrs(Env & env, Expr * e, Value & v, const Pos & pos, std::string_view errorCtx) + * { + * try { + * e->eval(*this, env, v); + * if (v.type() != nAttrs) + * throwTypeError("value is %1% while a set was expected", v); + * } catch (Error & e) { + * e.addTrace(pos, errorCtx); + * throw; + * } + * } + * ``` + * + * ``` short traces + * inline void EvalState::evalAttrs(Env & env, Expr * e, Value & v, const Pos & pos, std::string_view errorCtx) + * { + * e->eval(*this, env, v); + * try { + * if (v.type() != nAttrs) + * throwTypeError("value is %1% while a set was expected", v); + * } catch (Error & e) { + * e.addTrace(pos, errorCtx); + * throw; + * } + * } + * ``` + * + * The second example can be rewritten more concisely, but kept in this + * form to highlight the symmetry. The first option adds more information, + * because whatever caused an error down the line, in the generic eval + * function, will get annotated with the code location that uses and + * required it. The second option is less verbose, but does not provide + * any context at all as to where and why a failing value was required. + * + * Scopes would fix that, by adding context only when --show-trace is + * passed, and keeping the trace terse otherwise. + * + */ + + // Enough indent to align with with the `... ` + // prepended to each element of the trace + auto ellipsisIndent = " "; + + bool frameOnly = false; + if (!einfo.traces.empty()) { + size_t count = 0; + for (const auto & trace : einfo.traces) { + if (!showTrace && count > 3) { + oss << "\n" << ANSI_WARNING "(stack trace truncated; use '--show-trace' to show the full trace)" ANSI_NORMAL << "\n"; + break; + } + + if (trace.hint.str().empty()) continue; + if (frameOnly && !trace.frame) continue; + + count++; + frameOnly = trace.frame; + + oss << "\n" << "… " << trace.hint.str() << "\n"; + + if (trace.pos) { + count++; + + oss << "\n" << ellipsisIndent << ANSI_BLUE << "at " ANSI_WARNING << *trace.pos << ANSI_NORMAL << ":"; + + if (auto loc = trace.pos->getCodeLines()) { + oss << "\n"; + printCodeLines(oss, "", *trace.pos, *loc); + oss << "\n"; + } else + oss << noSource; + } + } + oss << "\n" << prefix; + } + + oss << einfo.msg << "\n"; - auto loc = getCodeLines(*einfo.errPos); + if (einfo.errPos) { + oss << "\n" << ANSI_BLUE << "at " ANSI_WARNING << *einfo.errPos << ANSI_NORMAL << ":"; - // lines of code. - if (loc.has_value()) { + if (auto loc = einfo.errPos->getCodeLines()) { oss << "\n"; printCodeLines(oss, "", *einfo.errPos, *loc); oss << "\n"; - } + } else + oss << noSource; } auto suggestions = einfo.suggestions.trim(); - if (! suggestions.suggestions.empty()){ + if (!suggestions.suggestions.empty()) { oss << "Did you mean " << suggestions.trim() << "?" << std::endl; } - // traces - if (showTrace && !einfo.traces.empty()) { - for (auto iter = einfo.traces.rbegin(); iter != einfo.traces.rend(); ++iter) { - oss << "\n" << "… " << iter->hint.str() << "\n"; - - if (iter->pos.has_value() && (*iter->pos)) { - auto pos = iter->pos.value(); - oss << "\n"; - printAtPos(pos, oss); - - auto loc = getCodeLines(pos); - if (loc.has_value()) { - oss << "\n"; - printCodeLines(oss, "", pos, *loc); - oss << "\n"; - } - } - } - } - out << indent(prefix, std::string(filterANSIEscapes(prefix, true).size(), ' '), chomp(oss.str())); return out; diff --git a/src/libutil/error.hh b/src/libutil/error.hh index 3d1479c54..0ebeaba61 100644 --- a/src/libutil/error.hh +++ b/src/libutil/error.hh @@ -54,13 +54,6 @@ typedef enum { lvlVomit } Verbosity; -/* adjust Pos::origin bit width when adding stuff here */ -typedef enum { - foFile, - foStdin, - foString -} FileOrigin; - // the lines of code surrounding an error. struct LinesOfCode { std::optional<std::string> prevLineOfCode; @@ -68,54 +61,40 @@ struct LinesOfCode { std::optional<std::string> nextLineOfCode; }; -// ErrPos indicates the location of an error in a nix file. -struct ErrPos { - int line = 0; - int column = 0; - std::string file; - FileOrigin origin; +/* An abstract type that represents a location in a source file. */ +struct AbstractPos +{ + uint32_t line = 0; + uint32_t column = 0; - operator bool() const - { - return line != 0; - } + /* Return the contents of the source file. */ + virtual std::optional<std::string> getSource() const + { return std::nullopt; }; - // convert from the Pos struct, found in libexpr. - template <class P> - ErrPos & operator=(const P & pos) - { - origin = pos.origin; - line = pos.line; - column = pos.column; - file = pos.file; - return *this; - } + virtual void print(std::ostream & out) const = 0; - template <class P> - ErrPos(const P & p) - { - *this = p; - } + std::optional<LinesOfCode> getCodeLines() const; + + virtual ~AbstractPos() = default; }; -std::optional<LinesOfCode> getCodeLines(const ErrPos & errPos); +std::ostream & operator << (std::ostream & str, const AbstractPos & pos); void printCodeLines(std::ostream & out, const std::string & prefix, - const ErrPos & errPos, + const AbstractPos & errPos, const LinesOfCode & loc); -void printAtPos(const ErrPos & pos, std::ostream & out); - struct Trace { - std::optional<ErrPos> pos; + std::shared_ptr<AbstractPos> pos; hintformat hint; + bool frame; }; struct ErrorInfo { Verbosity level; hintformat msg; - std::optional<ErrPos> errPos; + std::shared_ptr<AbstractPos> errPos; std::list<Trace> traces; Suggestions suggestions; @@ -138,6 +117,8 @@ protected: public: unsigned int status = 1; // exit status + BaseError(const BaseError &) = default; + template<typename... Args> BaseError(unsigned int status, const Args & ... args) : err { .level = lvlError, .msg = hintfmt(args...) } @@ -176,15 +157,22 @@ public: const std::string & msg() const { return calcWhat(); } const ErrorInfo & info() const { calcWhat(); return err; } + void pushTrace(Trace trace) + { + err.traces.push_front(trace); + } + template<typename... Args> - void addTrace(std::optional<ErrPos> e, const std::string & fs, const Args & ... args) + void addTrace(std::shared_ptr<AbstractPos> && e, std::string_view fs, const Args & ... args) { - addTrace(e, hintfmt(fs, args...)); + addTrace(std::move(e), hintfmt(std::string(fs), args...)); } - void addTrace(std::optional<ErrPos> e, hintformat hint); + void addTrace(std::shared_ptr<AbstractPos> && e, hintformat hint, bool frame = false); bool hasTrace() const { return !err.traces.empty(); } + + const ErrorInfo & info() { return err; }; }; #define MakeError(newClass, superClass) \ diff --git a/src/libutil/experimental-features.cc b/src/libutil/experimental-features.cc index fa79cca6b..58d762ebb 100644 --- a/src/libutil/experimental-features.cc +++ b/src/libutil/experimental-features.cc @@ -14,6 +14,9 @@ std::map<ExperimentalFeature, std::string> stringifiedXpFeatures = { { Xp::NoUrlLiterals, "no-url-literals" }, { Xp::FetchClosure, "fetch-closure" }, { Xp::ReplFlake, "repl-flake" }, + { Xp::AutoAllocateUids, "auto-allocate-uids" }, + { Xp::Cgroups, "cgroups" }, + { Xp::DiscardReferences, "discard-references" }, }; const std::optional<ExperimentalFeature> parseExperimentalFeature(const std::string_view & name) diff --git a/src/libutil/experimental-features.hh b/src/libutil/experimental-features.hh index d09ab025c..ac372e03e 100644 --- a/src/libutil/experimental-features.hh +++ b/src/libutil/experimental-features.hh @@ -23,6 +23,9 @@ enum struct ExperimentalFeature NoUrlLiterals, FetchClosure, ReplFlake, + AutoAllocateUids, + Cgroups, + DiscardReferences, }; /** diff --git a/src/libutil/filesystem.cc b/src/libutil/filesystem.cc new file mode 100644 index 000000000..56be76ecc --- /dev/null +++ b/src/libutil/filesystem.cc @@ -0,0 +1,173 @@ +#include <sys/time.h> +#include <filesystem> +#include <atomic> + +#include "finally.hh" +#include "util.hh" +#include "types.hh" + +namespace fs = std::filesystem; + +namespace nix { + +static Path tempName(Path tmpRoot, const Path & prefix, bool includePid, + std::atomic<unsigned int> & counter) +{ + tmpRoot = canonPath(tmpRoot.empty() ? getEnv("TMPDIR").value_or("/tmp") : tmpRoot, true); + if (includePid) + return fmt("%1%/%2%-%3%-%4%", tmpRoot, prefix, getpid(), counter++); + else + return fmt("%1%/%2%-%3%", tmpRoot, prefix, counter++); +} + +Path createTempDir(const Path & tmpRoot, const Path & prefix, + bool includePid, bool useGlobalCounter, mode_t mode) +{ + static std::atomic<unsigned int> globalCounter = 0; + std::atomic<unsigned int> localCounter = 0; + auto & counter(useGlobalCounter ? globalCounter : localCounter); + + while (1) { + checkInterrupt(); + Path tmpDir = tempName(tmpRoot, prefix, includePid, counter); + if (mkdir(tmpDir.c_str(), mode) == 0) { +#if __FreeBSD__ + /* Explicitly set the group of the directory. This is to + work around around problems caused by BSD's group + ownership semantics (directories inherit the group of + the parent). For instance, the group of /tmp on + FreeBSD is "wheel", so all directories created in /tmp + will be owned by "wheel"; but if the user is not in + "wheel", then "tar" will fail to unpack archives that + have the setgid bit set on directories. */ + if (chown(tmpDir.c_str(), (uid_t) -1, getegid()) != 0) + throw SysError("setting group of directory '%1%'", tmpDir); +#endif + return tmpDir; + } + if (errno != EEXIST) + throw SysError("creating directory '%1%'", tmpDir); + } +} + + +std::pair<AutoCloseFD, Path> createTempFile(const Path & prefix) +{ + Path tmpl(getEnv("TMPDIR").value_or("/tmp") + "/" + prefix + ".XXXXXX"); + // Strictly speaking, this is UB, but who cares... + // FIXME: use O_TMPFILE. + AutoCloseFD fd(mkstemp((char *) tmpl.c_str())); + if (!fd) + throw SysError("creating temporary file '%s'", tmpl); + closeOnExec(fd.get()); + return {std::move(fd), tmpl}; +} + +void createSymlink(const Path & target, const Path & link, + std::optional<time_t> mtime) +{ + if (symlink(target.c_str(), link.c_str())) + throw SysError("creating symlink from '%1%' to '%2%'", link, target); + if (mtime) { + struct timeval times[2]; + times[0].tv_sec = *mtime; + times[0].tv_usec = 0; + times[1].tv_sec = *mtime; + times[1].tv_usec = 0; + if (lutimes(link.c_str(), times)) + throw SysError("setting time of symlink '%s'", link); + } +} + +void replaceSymlink(const Path & target, const Path & link, + std::optional<time_t> mtime) +{ + for (unsigned int n = 0; true; n++) { + Path tmp = canonPath(fmt("%s/.%d_%s", dirOf(link), n, baseNameOf(link))); + + try { + createSymlink(target, tmp, mtime); + } catch (SysError & e) { + if (e.errNo == EEXIST) continue; + throw; + } + + renameFile(tmp, link); + + break; + } +} + +void setWriteTime(const fs::path & p, const struct stat & st) +{ + struct timeval times[2]; + times[0] = { + .tv_sec = st.st_atime, + .tv_usec = 0, + }; + times[1] = { + .tv_sec = st.st_mtime, + .tv_usec = 0, + }; + if (lutimes(p.c_str(), times) != 0) + throw SysError("changing modification time of '%s'", p); +} + +void copy(const fs::directory_entry & from, const fs::path & to, bool andDelete) +{ + // TODO: Rewrite the `is_*` to use `symlink_status()` + auto statOfFrom = lstat(from.path().c_str()); + auto fromStatus = from.symlink_status(); + + // Mark the directory as writable so that we can delete its children + if (andDelete && fs::is_directory(fromStatus)) { + fs::permissions(from.path(), fs::perms::owner_write, fs::perm_options::add | fs::perm_options::nofollow); + } + + + if (fs::is_symlink(fromStatus) || fs::is_regular_file(fromStatus)) { + fs::copy(from.path(), to, fs::copy_options::copy_symlinks | fs::copy_options::overwrite_existing); + } else if (fs::is_directory(fromStatus)) { + fs::create_directory(to); + for (auto & entry : fs::directory_iterator(from.path())) { + copy(entry, to / entry.path().filename(), andDelete); + } + } else { + throw Error("file '%s' has an unsupported type", from.path()); + } + + setWriteTime(to, statOfFrom); + if (andDelete) { + if (!fs::is_symlink(fromStatus)) + fs::permissions(from.path(), fs::perms::owner_write, fs::perm_options::add | fs::perm_options::nofollow); + fs::remove(from.path()); + } +} + +void renameFile(const Path & oldName, const Path & newName) +{ + fs::rename(oldName, newName); +} + +void moveFile(const Path & oldName, const Path & newName) +{ + try { + renameFile(oldName, newName); + } catch (fs::filesystem_error & e) { + auto oldPath = fs::path(oldName); + auto newPath = fs::path(newName); + // For the move to be as atomic as possible, copy to a temporary + // directory + fs::path temp = createTempDir(newPath.parent_path(), "rename-tmp"); + Finally removeTemp = [&]() { fs::remove(temp); }; + auto tempCopyTarget = temp / "copy-target"; + if (e.code().value() == EXDEV) { + fs::remove(newPath); + warn("Can’t rename %s as %s, copying instead", oldName, newName); + copy(fs::directory_entry(oldPath), tempCopyTarget, true); + renameFile(tempCopyTarget, newPath); + } + } +} + +} diff --git a/src/libutil/fmt.hh b/src/libutil/fmt.hh index 7664e5c04..e11426b88 100644 --- a/src/libutil/fmt.hh +++ b/src/libutil/fmt.hh @@ -17,16 +17,6 @@ using boost::format; struct nop { template<typename... T> nop(T...) {} }; -struct FormatOrString -{ - std::string s; - FormatOrString(std::string s) : s(std::move(s)) { }; - template<class F> - FormatOrString(const F & f) : s(f.str()) { }; - FormatOrString(const char * s) : s(s) { }; -}; - - /* A helper for formatting strings. ‘fmt(format, a_0, ..., a_n)’ is equivalent to ‘boost::format(format) % a_0 % ... % ... a_n’. However, ‘fmt(s)’ is equivalent to ‘s’ (so no %-expansion @@ -53,11 +43,6 @@ inline std::string fmt(const char * s) return s; } -inline std::string fmt(const FormatOrString & fs) -{ - return fs.s; -} - template<typename... Args> inline std::string fmt(const std::string & fs, const Args & ... args) { @@ -148,7 +133,7 @@ inline hintformat hintfmt(const std::string & fs, const Args & ... args) return f; } -inline hintformat hintfmt(std::string plain_string) +inline hintformat hintfmt(const std::string & plain_string) { // we won't be receiving any args in this case, so just print the original string return hintfmt("%s", normaltxt(plain_string)); diff --git a/src/libutil/json-impls.hh b/src/libutil/json-impls.hh new file mode 100644 index 000000000..bd75748ad --- /dev/null +++ b/src/libutil/json-impls.hh @@ -0,0 +1,14 @@ +#pragma once + +#include "nlohmann/json_fwd.hpp" + +// Following https://github.com/nlohmann/json#how-can-i-use-get-for-non-default-constructiblenon-copyable-types +#define JSON_IMPL(TYPE) \ + namespace nlohmann { \ + using namespace nix; \ + template <> \ + struct adl_serializer<TYPE> { \ + static TYPE from_json(const json & json); \ + static void to_json(json & json, TYPE t); \ + }; \ + } diff --git a/src/libutil/json.cc b/src/libutil/json.cc deleted file mode 100644 index abe0e6e74..000000000 --- a/src/libutil/json.cc +++ /dev/null @@ -1,199 +0,0 @@ -#include "json.hh" - -#include <iomanip> -#include <cstdint> -#include <cstring> - -namespace nix { - -template<> -void toJSON<std::string_view>(std::ostream & str, const std::string_view & s) -{ - constexpr size_t BUF_SIZE = 4096; - char buf[BUF_SIZE + 7]; // BUF_SIZE + largest single sequence of puts - size_t bufPos = 0; - - const auto flush = [&] { - str.write(buf, bufPos); - bufPos = 0; - }; - const auto put = [&] (char c) { - buf[bufPos++] = c; - }; - - put('"'); - for (auto i = s.begin(); i != s.end(); i++) { - if (bufPos >= BUF_SIZE) flush(); - if (*i == '\"' || *i == '\\') { put('\\'); put(*i); } - else if (*i == '\n') { put('\\'); put('n'); } - else if (*i == '\r') { put('\\'); put('r'); } - else if (*i == '\t') { put('\\'); put('t'); } - else if (*i >= 0 && *i < 32) { - const char hex[17] = "0123456789abcdef"; - put('\\'); - put('u'); - put(hex[(uint16_t(*i) >> 12) & 0xf]); - put(hex[(uint16_t(*i) >> 8) & 0xf]); - put(hex[(uint16_t(*i) >> 4) & 0xf]); - put(hex[(uint16_t(*i) >> 0) & 0xf]); - } - else put(*i); - } - put('"'); - flush(); -} - -void toJSON(std::ostream & str, const char * s) -{ - if (!s) str << "null"; else toJSON(str, std::string_view(s)); -} - -template<> void toJSON<int>(std::ostream & str, const int & n) { str << n; } -template<> void toJSON<unsigned int>(std::ostream & str, const unsigned int & n) { str << n; } -template<> void toJSON<long>(std::ostream & str, const long & n) { str << n; } -template<> void toJSON<unsigned long>(std::ostream & str, const unsigned long & n) { str << n; } -template<> void toJSON<long long>(std::ostream & str, const long long & n) { str << n; } -template<> void toJSON<unsigned long long>(std::ostream & str, const unsigned long long & n) { str << n; } -template<> void toJSON<float>(std::ostream & str, const float & n) { str << n; } -template<> void toJSON<double>(std::ostream & str, const double & n) { str << n; } -template<> void toJSON<std::string>(std::ostream & str, const std::string & s) { toJSON(str, (std::string_view) s); } - -template<> void toJSON<bool>(std::ostream & str, const bool & b) -{ - str << (b ? "true" : "false"); -} - -template<> void toJSON<std::nullptr_t>(std::ostream & str, const std::nullptr_t & b) -{ - str << "null"; -} - -JSONWriter::JSONWriter(std::ostream & str, bool indent) - : state(new JSONState(str, indent)) -{ - state->stack++; -} - -JSONWriter::JSONWriter(JSONState * state) - : state(state) -{ - state->stack++; -} - -JSONWriter::~JSONWriter() -{ - if (state) { - assertActive(); - state->stack--; - if (state->stack == 0) delete state; - } -} - -void JSONWriter::comma() -{ - assertActive(); - if (first) { - first = false; - } else { - state->str << ','; - } - if (state->indent) indent(); -} - -void JSONWriter::indent() -{ - state->str << '\n' << std::string(state->depth * 2, ' '); -} - -void JSONList::open() -{ - state->depth++; - state->str << '['; -} - -JSONList::~JSONList() -{ - state->depth--; - if (state->indent && !first) indent(); - state->str << "]"; -} - -JSONList JSONList::list() -{ - comma(); - return JSONList(state); -} - -JSONObject JSONList::object() -{ - comma(); - return JSONObject(state); -} - -JSONPlaceholder JSONList::placeholder() -{ - comma(); - return JSONPlaceholder(state); -} - -void JSONObject::open() -{ - state->depth++; - state->str << '{'; -} - -JSONObject::~JSONObject() -{ - if (state) { - state->depth--; - if (state->indent && !first) indent(); - state->str << "}"; - } -} - -void JSONObject::attr(std::string_view s) -{ - comma(); - toJSON(state->str, s); - state->str << ':'; - if (state->indent) state->str << ' '; -} - -JSONList JSONObject::list(std::string_view name) -{ - attr(name); - return JSONList(state); -} - -JSONObject JSONObject::object(std::string_view name) -{ - attr(name); - return JSONObject(state); -} - -JSONPlaceholder JSONObject::placeholder(std::string_view name) -{ - attr(name); - return JSONPlaceholder(state); -} - -JSONList JSONPlaceholder::list() -{ - assertValid(); - first = false; - return JSONList(state); -} - -JSONObject JSONPlaceholder::object() -{ - assertValid(); - first = false; - return JSONObject(state); -} - -JSONPlaceholder::~JSONPlaceholder() -{ - assert(!first || std::uncaught_exceptions()); -} - -} diff --git a/src/libutil/json.hh b/src/libutil/json.hh deleted file mode 100644 index 3790b1a2e..000000000 --- a/src/libutil/json.hh +++ /dev/null @@ -1,185 +0,0 @@ -#pragma once - -#include <iostream> -#include <vector> -#include <cassert> - -namespace nix { - -void toJSON(std::ostream & str, const char * s); - -template<typename T> -void toJSON(std::ostream & str, const T & n); - -class JSONWriter -{ -protected: - - struct JSONState - { - std::ostream & str; - bool indent; - size_t depth = 0; - size_t stack = 0; - JSONState(std::ostream & str, bool indent) : str(str), indent(indent) { } - ~JSONState() - { - assert(stack == 0); - } - }; - - JSONState * state; - - bool first = true; - - JSONWriter(std::ostream & str, bool indent); - - JSONWriter(JSONState * state); - - ~JSONWriter(); - - void assertActive() - { - assert(state->stack != 0); - } - - void comma(); - - void indent(); -}; - -class JSONObject; -class JSONPlaceholder; - -class JSONList : JSONWriter -{ -private: - - friend class JSONObject; - friend class JSONPlaceholder; - - void open(); - - JSONList(JSONState * state) - : JSONWriter(state) - { - open(); - } - -public: - - JSONList(std::ostream & str, bool indent = false) - : JSONWriter(str, indent) - { - open(); - } - - ~JSONList(); - - template<typename T> - JSONList & elem(const T & v) - { - comma(); - toJSON(state->str, v); - return *this; - } - - JSONList list(); - - JSONObject object(); - - JSONPlaceholder placeholder(); -}; - -class JSONObject : JSONWriter -{ -private: - - friend class JSONList; - friend class JSONPlaceholder; - - void open(); - - JSONObject(JSONState * state) - : JSONWriter(state) - { - open(); - } - - void attr(std::string_view s); - -public: - - JSONObject(std::ostream & str, bool indent = false) - : JSONWriter(str, indent) - { - open(); - } - - JSONObject(const JSONObject & obj) = delete; - - JSONObject(JSONObject && obj) - : JSONWriter(obj.state) - { - obj.state = 0; - } - - ~JSONObject(); - - template<typename T> - JSONObject & attr(std::string_view name, const T & v) - { - attr(name); - toJSON(state->str, v); - return *this; - } - - JSONList list(std::string_view name); - - JSONObject object(std::string_view name); - - JSONPlaceholder placeholder(std::string_view name); -}; - -class JSONPlaceholder : JSONWriter -{ - -private: - - friend class JSONList; - friend class JSONObject; - - JSONPlaceholder(JSONState * state) - : JSONWriter(state) - { - } - - void assertValid() - { - assertActive(); - assert(first); - } - -public: - - JSONPlaceholder(std::ostream & str, bool indent = false) - : JSONWriter(str, indent) - { - } - - ~JSONPlaceholder(); - - template<typename T> - void write(const T & v) - { - assertValid(); - first = false; - toJSON(state->str, v); - } - - JSONList list(); - - JSONObject object(); -}; - -} diff --git a/src/libutil/logging.cc b/src/libutil/logging.cc index cb2b15b41..7cac75ce1 100644 --- a/src/libutil/logging.cc +++ b/src/libutil/logging.cc @@ -32,7 +32,8 @@ void Logger::warn(const std::string & msg) void Logger::writeToStdout(std::string_view s) { - std::cout << s << "\n"; + writeFull(STDOUT_FILENO, s); + writeFull(STDOUT_FILENO, "\n"); } class SimpleLogger : public Logger @@ -53,7 +54,7 @@ public: return printBuildLogs; } - void log(Verbosity lvl, const FormatOrString & fs) override + void log(Verbosity lvl, std::string_view s) override { if (lvl > verbosity) return; @@ -71,7 +72,7 @@ public: prefix = std::string("<") + c + ">"; } - writeToStderr(prefix + filterANSIEscapes(fs.s, !tty) + "\n"); + writeToStderr(prefix + filterANSIEscapes(s, !tty) + "\n"); } void logEI(const ErrorInfo & ei) override @@ -84,7 +85,7 @@ public: void startActivity(ActivityId act, Verbosity lvl, ActivityType type, const std::string & s, const Fields & fields, ActivityId parent) - override + override { if (lvl <= verbosity && !s.empty()) log(lvl, s + "..."); @@ -105,14 +106,6 @@ public: Verbosity verbosity = lvlInfo; -void warnOnce(bool & haveWarned, const FormatOrString & fs) -{ - if (!haveWarned) { - warn(fs.s); - haveWarned = true; - } -} - void writeToStderr(std::string_view s) { try { @@ -130,15 +123,30 @@ Logger * makeSimpleLogger(bool printBuildLogs) return new SimpleLogger(printBuildLogs); } -std::atomic<uint64_t> nextId{(uint64_t) getpid() << 32}; +std::atomic<uint64_t> nextId{0}; Activity::Activity(Logger & logger, Verbosity lvl, ActivityType type, const std::string & s, const Logger::Fields & fields, ActivityId parent) - : logger(logger), id(nextId++) + : logger(logger), id(nextId++ + (((uint64_t) getpid()) << 32)) { logger.startActivity(id, lvl, type, s, fields, parent); } +void to_json(nlohmann::json & json, std::shared_ptr<AbstractPos> pos) +{ + if (pos) { + json["line"] = pos->line; + json["column"] = pos->column; + std::ostringstream str; + pos->print(str); + json["file"] = str.str(); + } else { + json["line"] = nullptr; + json["column"] = nullptr; + json["file"] = nullptr; + } +} + struct JSONLogger : Logger { Logger & prevLogger; @@ -166,12 +174,12 @@ struct JSONLogger : Logger { prevLogger.log(lvlError, "@nix " + json.dump(-1, ' ', false, nlohmann::json::error_handler_t::replace)); } - void log(Verbosity lvl, const FormatOrString & fs) override + void log(Verbosity lvl, std::string_view s) override { nlohmann::json json; json["action"] = "msg"; json["level"] = lvl; - json["msg"] = fs.s; + json["msg"] = s; write(json); } @@ -185,27 +193,14 @@ struct JSONLogger : Logger { json["level"] = ei.level; json["msg"] = oss.str(); json["raw_msg"] = ei.msg.str(); - - if (ei.errPos.has_value() && (*ei.errPos)) { - json["line"] = ei.errPos->line; - json["column"] = ei.errPos->column; - json["file"] = ei.errPos->file; - } else { - json["line"] = nullptr; - json["column"] = nullptr; - json["file"] = nullptr; - } + to_json(json, ei.errPos); if (loggerSettings.showTrace.get() && !ei.traces.empty()) { nlohmann::json traces = nlohmann::json::array(); for (auto iter = ei.traces.rbegin(); iter != ei.traces.rend(); ++iter) { nlohmann::json stackFrame; stackFrame["raw_msg"] = iter->hint.str(); - if (iter->pos.has_value() && (*iter->pos)) { - stackFrame["line"] = iter->pos->line; - stackFrame["column"] = iter->pos->column; - stackFrame["file"] = iter->pos->file; - } + to_json(stackFrame, iter->pos); traces.push_back(stackFrame); } diff --git a/src/libutil/logging.hh b/src/libutil/logging.hh index 6f81b92de..59a707eef 100644 --- a/src/libutil/logging.hh +++ b/src/libutil/logging.hh @@ -75,14 +75,14 @@ public: // Whether the logger prints the whole build log virtual bool isVerbose() { return false; } - virtual void log(Verbosity lvl, const FormatOrString & fs) = 0; + virtual void log(Verbosity lvl, std::string_view s) = 0; - void log(const FormatOrString & fs) + void log(std::string_view s) { - log(lvlInfo, fs); + log(lvlInfo, s); } - virtual void logEI(const ErrorInfo &ei) = 0; + virtual void logEI(const ErrorInfo & ei) = 0; void logEI(Verbosity lvl, ErrorInfo ei) { @@ -102,15 +102,16 @@ public: virtual void writeToStdout(std::string_view s); template<typename... Args> - inline void cout(const std::string & fs, const Args & ... args) + inline void cout(const Args & ... args) { - boost::format f(fs); - formatHelper(f, args...); - writeToStdout(f.str()); + writeToStdout(fmt(args...)); } virtual std::optional<char> ask(std::string_view s) { return {}; } + + virtual void setPrintBuildLogs(bool printBuildLogs) + { } }; ActivityId getCurActivity(); @@ -222,7 +223,11 @@ inline void warn(const std::string & fs, const Args & ... args) logger->warn(f.str()); } -void warnOnce(bool & haveWarned, const FormatOrString & fs); +#define warnOnce(haveWarned, args...) \ + if (!haveWarned) { \ + haveWarned = true; \ + warn(args); \ + } void writeToStderr(std::string_view s); diff --git a/src/libutil/monitor-fd.hh b/src/libutil/monitor-fd.hh index 5ee0b88ef..9518cf8aa 100644 --- a/src/libutil/monitor-fd.hh +++ b/src/libutil/monitor-fd.hh @@ -22,27 +22,38 @@ public: { thread = std::thread([fd]() { while (true) { - /* Wait indefinitely until a POLLHUP occurs. */ - struct pollfd fds[1]; - fds[0].fd = fd; - /* This shouldn't be necessary, but macOS doesn't seem to - like a zeroed out events field. - See rdar://37537852. - */ - fds[0].events = POLLHUP; - auto count = poll(fds, 1, -1); - if (count == -1) abort(); // can't happen - /* This shouldn't happen, but can on macOS due to a bug. - See rdar://37550628. - - This may eventually need a delay or further - coordination with the main thread if spinning proves - too harmful. - */ - if (count == 0) continue; - assert(fds[0].revents & POLLHUP); - triggerInterrupt(); - break; + /* Wait indefinitely until a POLLHUP occurs. */ + struct pollfd fds[1]; + fds[0].fd = fd; + /* Polling for no specific events (i.e. just waiting + for an error/hangup) doesn't work on macOS + anymore. So wait for read events and ignore + them. */ + fds[0].events = + #ifdef __APPLE__ + POLLRDNORM + #else + 0 + #endif + ; + auto count = poll(fds, 1, -1); + if (count == -1) abort(); // can't happen + /* This shouldn't happen, but can on macOS due to a bug. + See rdar://37550628. + + This may eventually need a delay or further + coordination with the main thread if spinning proves + too harmful. + */ + if (count == 0) continue; + if (fds[0].revents & POLLHUP) { + triggerInterrupt(); + break; + } + /* This will only happen on macOS. We sleep a bit to + avoid waking up too often if the client is sending + input. */ + sleep(1); } }); }; diff --git a/src/libutil/namespaces.cc b/src/libutil/namespaces.cc new file mode 100644 index 000000000..f66accb10 --- /dev/null +++ b/src/libutil/namespaces.cc @@ -0,0 +1,97 @@ +#if __linux__ + +#include "namespaces.hh" +#include "util.hh" +#include "finally.hh" + +#include <sys/mount.h> + +namespace nix { + +bool userNamespacesSupported() +{ + static auto res = [&]() -> bool + { + if (!pathExists("/proc/self/ns/user")) { + debug("'/proc/self/ns/user' does not exist; your kernel was likely built without CONFIG_USER_NS=y"); + return false; + } + + Path maxUserNamespaces = "/proc/sys/user/max_user_namespaces"; + if (!pathExists(maxUserNamespaces) || + trim(readFile(maxUserNamespaces)) == "0") + { + debug("user namespaces appear to be disabled; check '/proc/sys/user/max_user_namespaces'"); + return false; + } + + Path procSysKernelUnprivilegedUsernsClone = "/proc/sys/kernel/unprivileged_userns_clone"; + if (pathExists(procSysKernelUnprivilegedUsernsClone) + && trim(readFile(procSysKernelUnprivilegedUsernsClone)) == "0") + { + debug("user namespaces appear to be disabled; check '/proc/sys/kernel/unprivileged_userns_clone'"); + return false; + } + + try { + Pid pid = startProcess([&]() + { + _exit(0); + }, { + .cloneFlags = CLONE_NEWUSER + }); + + auto r = pid.wait(); + assert(!r); + } catch (SysError & e) { + debug("user namespaces do not work on this system: %s", e.msg()); + return false; + } + + return true; + }(); + return res; +} + +bool mountAndPidNamespacesSupported() +{ + static auto res = [&]() -> bool + { + try { + + Pid pid = startProcess([&]() + { + /* Make sure we don't remount the parent's /proc. */ + if (mount(0, "/", 0, MS_PRIVATE | MS_REC, 0) == -1) + _exit(1); + + /* Test whether we can remount /proc. The kernel disallows + this if /proc is not fully visible, i.e. if there are + filesystems mounted on top of files inside /proc. See + https://lore.kernel.org/lkml/87tvsrjai0.fsf@xmission.com/T/. */ + if (mount("none", "/proc", "proc", 0, 0) == -1) + _exit(2); + + _exit(0); + }, { + .cloneFlags = CLONE_NEWNS | CLONE_NEWPID | (userNamespacesSupported() ? CLONE_NEWUSER : 0) + }); + + if (pid.wait()) { + debug("PID namespaces do not work on this system: cannot remount /proc"); + return false; + } + + } catch (SysError & e) { + debug("mount namespaces do not work on this system: %s", e.msg()); + return false; + } + + return true; + }(); + return res; +} + +} + +#endif diff --git a/src/libutil/namespaces.hh b/src/libutil/namespaces.hh new file mode 100644 index 000000000..e82379b9c --- /dev/null +++ b/src/libutil/namespaces.hh @@ -0,0 +1,13 @@ +#pragma once + +namespace nix { + +#if __linux__ + +bool userNamespacesSupported(); + +bool mountAndPidNamespacesSupported(); + +#endif + +} diff --git a/src/libutil/ref.hh b/src/libutil/ref.hh index bf26321db..7d38b059c 100644 --- a/src/libutil/ref.hh +++ b/src/libutil/ref.hh @@ -83,6 +83,11 @@ public: return p != other.p; } + bool operator < (const ref<T> & other) const + { + return p < other.p; + } + private: template<typename T2, typename... Args> diff --git a/src/libutil/regex-combinators.hh b/src/libutil/regex-combinators.hh new file mode 100644 index 000000000..0b997b25a --- /dev/null +++ b/src/libutil/regex-combinators.hh @@ -0,0 +1,30 @@ +#pragma once + +#include <string_view> + +namespace nix::regex { + +// TODO use constexpr string building like +// https://github.com/akrzemi1/static_string/blob/master/include/ak_toolkit/static_string.hpp + +static inline std::string either(std::string_view a, std::string_view b) +{ + return std::string { a } + "|" + b; +} + +static inline std::string group(std::string_view a) +{ + return std::string { "(" } + a + ")"; +} + +static inline std::string many(std::string_view a) +{ + return std::string { "(?:" } + a + ")*"; +} + +static inline std::string list(std::string_view a) +{ + return std::string { a } + many(group("," + a)); +} + +} diff --git a/src/libutil/serialise.cc b/src/libutil/serialise.cc index 8ff904583..7476e6f6c 100644 --- a/src/libutil/serialise.cc +++ b/src/libutil/serialise.cc @@ -48,24 +48,9 @@ FdSink::~FdSink() } -size_t threshold = 256 * 1024 * 1024; - -static void warnLargeDump() -{ - warn("dumping very large path (> 256 MiB); this may run out of memory"); -} - - void FdSink::write(std::string_view data) { written += data.size(); - static bool warned = false; - if (warn && !warned) { - if (written > threshold) { - warnLargeDump(); - warned = true; - } - } try { writeFull(fd, data); } catch (SysError & e) { @@ -353,7 +338,7 @@ Sink & operator << (Sink & sink, const StringSet & s) Sink & operator << (Sink & sink, const Error & ex) { - auto info = ex.info(); + auto & info = ex.info(); sink << "Error" << info.level @@ -430,7 +415,7 @@ Error readError(Source & source) auto msg = readString(source); ErrorInfo info { .level = level, - .msg = hintformat(std::move(format("%s") % msg)), + .msg = hintformat(fmt("%s", msg)), }; auto havePos = readNum<size_t>(source); assert(havePos == 0); @@ -439,7 +424,7 @@ Error readError(Source & source) havePos = readNum<size_t>(source); assert(havePos == 0); info.traces.push_back(Trace { - .hint = hintformat(std::move(format("%s") % readString(source))) + .hint = hintformat(fmt("%s", readString(source))) }); } return Error(std::move(info)); @@ -448,11 +433,6 @@ Error readError(Source & source) void StringSink::operator () (std::string_view data) { - static bool warned = false; - if (!warned && s.size() > threshold) { - warnLargeDump(); - warned = true; - } s.append(data); } diff --git a/src/libutil/serialise.hh b/src/libutil/serialise.hh index 13da26c6a..7da5b07fd 100644 --- a/src/libutil/serialise.hh +++ b/src/libutil/serialise.hh @@ -97,19 +97,17 @@ protected: struct FdSink : BufferedSink { int fd; - bool warn = false; size_t written = 0; FdSink() : fd(-1) { } FdSink(int fd) : fd(fd) { } FdSink(FdSink&&) = default; - FdSink& operator=(FdSink && s) + FdSink & operator=(FdSink && s) { flush(); fd = s.fd; s.fd = -1; - warn = s.warn; written = s.written; return *this; } @@ -333,17 +331,9 @@ T readNum(Source & source) unsigned char buf[8]; source((char *) buf, sizeof(buf)); - uint64_t n = - ((uint64_t) buf[0]) | - ((uint64_t) buf[1] << 8) | - ((uint64_t) buf[2] << 16) | - ((uint64_t) buf[3] << 24) | - ((uint64_t) buf[4] << 32) | - ((uint64_t) buf[5] << 40) | - ((uint64_t) buf[6] << 48) | - ((uint64_t) buf[7] << 56); - - if (n > (uint64_t)std::numeric_limits<T>::max()) + auto n = readLittleEndian<uint64_t>(buf); + + if (n > (uint64_t) std::numeric_limits<T>::max()) throw SerialisationError("serialised integer %d is too large for type '%s'", n, typeid(T).name()); return (T) n; diff --git a/src/libutil/tarfile.cc b/src/libutil/tarfile.cc index a7db58559..238d0a7a6 100644 --- a/src/libutil/tarfile.cc +++ b/src/libutil/tarfile.cc @@ -77,9 +77,7 @@ TarArchive::~TarArchive() static void extract_archive(TarArchive & archive, const Path & destDir) { - int flags = ARCHIVE_EXTRACT_FFLAGS - | ARCHIVE_EXTRACT_PERM - | ARCHIVE_EXTRACT_TIME + int flags = ARCHIVE_EXTRACT_TIME | ARCHIVE_EXTRACT_SECURE_SYMLINKS | ARCHIVE_EXTRACT_SECURE_NODOTDOT; @@ -98,6 +96,10 @@ static void extract_archive(TarArchive & archive, const Path & destDir) archive_entry_copy_pathname(entry, (destDir + "/" + name).c_str()); + // sources can and do contain dirs with no rx bits + if (archive_entry_filetype(entry) == AE_IFDIR && (archive_entry_mode(entry) & 0500) != 0500) + archive_entry_set_mode(entry, archive_entry_mode(entry) | 0500); + // Patch hardlink path const char *original_hardlink = archive_entry_hardlink(entry); if (original_hardlink) { diff --git a/src/libutil/tests/canon-path.cc b/src/libutil/tests/canon-path.cc new file mode 100644 index 000000000..c1c5adadf --- /dev/null +++ b/src/libutil/tests/canon-path.cc @@ -0,0 +1,155 @@ +#include "canon-path.hh" + +#include <gtest/gtest.h> + +namespace nix { + + TEST(CanonPath, basic) { + { + CanonPath p("/"); + ASSERT_EQ(p.abs(), "/"); + ASSERT_EQ(p.rel(), ""); + ASSERT_EQ(p.baseName(), std::nullopt); + ASSERT_EQ(p.dirOf(), std::nullopt); + ASSERT_FALSE(p.parent()); + } + + { + CanonPath p("/foo//"); + ASSERT_EQ(p.abs(), "/foo"); + ASSERT_EQ(p.rel(), "foo"); + ASSERT_EQ(*p.baseName(), "foo"); + ASSERT_EQ(*p.dirOf(), ""); // FIXME: do we want this? + ASSERT_EQ(p.parent()->abs(), "/"); + } + + { + CanonPath p("foo/bar"); + ASSERT_EQ(p.abs(), "/foo/bar"); + ASSERT_EQ(p.rel(), "foo/bar"); + ASSERT_EQ(*p.baseName(), "bar"); + ASSERT_EQ(*p.dirOf(), "/foo"); + ASSERT_EQ(p.parent()->abs(), "/foo"); + } + + { + CanonPath p("foo//bar/"); + ASSERT_EQ(p.abs(), "/foo/bar"); + ASSERT_EQ(p.rel(), "foo/bar"); + ASSERT_EQ(*p.baseName(), "bar"); + ASSERT_EQ(*p.dirOf(), "/foo"); + } + } + + TEST(CanonPath, pop) { + CanonPath p("foo/bar/x"); + ASSERT_EQ(p.abs(), "/foo/bar/x"); + p.pop(); + ASSERT_EQ(p.abs(), "/foo/bar"); + p.pop(); + ASSERT_EQ(p.abs(), "/foo"); + p.pop(); + ASSERT_EQ(p.abs(), "/"); + } + + TEST(CanonPath, removePrefix) { + CanonPath p1("foo/bar"); + CanonPath p2("foo/bar/a/b/c"); + ASSERT_EQ(p2.removePrefix(p1).abs(), "/a/b/c"); + ASSERT_EQ(p1.removePrefix(p1).abs(), "/"); + ASSERT_EQ(p1.removePrefix(CanonPath("/")).abs(), "/foo/bar"); + } + + TEST(CanonPath, iter) { + { + CanonPath p("a//foo/bar//"); + std::vector<std::string_view> ss; + for (auto & c : p) ss.push_back(c); + ASSERT_EQ(ss, std::vector<std::string_view>({"a", "foo", "bar"})); + } + + { + CanonPath p("/"); + std::vector<std::string_view> ss; + for (auto & c : p) ss.push_back(c); + ASSERT_EQ(ss, std::vector<std::string_view>()); + } + } + + TEST(CanonPath, concat) { + { + CanonPath p1("a//foo/bar//"); + CanonPath p2("xyzzy/bla"); + ASSERT_EQ((p1 + p2).abs(), "/a/foo/bar/xyzzy/bla"); + } + + { + CanonPath p1("/"); + CanonPath p2("/a/b"); + ASSERT_EQ((p1 + p2).abs(), "/a/b"); + } + + { + CanonPath p1("/a/b"); + CanonPath p2("/"); + ASSERT_EQ((p1 + p2).abs(), "/a/b"); + } + + { + CanonPath p("/foo/bar"); + ASSERT_EQ((p + "x").abs(), "/foo/bar/x"); + } + + { + CanonPath p("/"); + ASSERT_EQ((p + "foo" + "bar").abs(), "/foo/bar"); + } + } + + TEST(CanonPath, within) { + { + ASSERT_TRUE(CanonPath("foo").isWithin(CanonPath("foo"))); + ASSERT_FALSE(CanonPath("foo").isWithin(CanonPath("bar"))); + ASSERT_FALSE(CanonPath("foo").isWithin(CanonPath("fo"))); + ASSERT_TRUE(CanonPath("foo/bar").isWithin(CanonPath("foo"))); + ASSERT_FALSE(CanonPath("foo").isWithin(CanonPath("foo/bar"))); + ASSERT_TRUE(CanonPath("/foo/bar/default.nix").isWithin(CanonPath("/"))); + ASSERT_TRUE(CanonPath("/").isWithin(CanonPath("/"))); + } + } + + TEST(CanonPath, sort) { + ASSERT_FALSE(CanonPath("foo") < CanonPath("foo")); + ASSERT_TRUE (CanonPath("foo") < CanonPath("foo/bar")); + ASSERT_TRUE (CanonPath("foo/bar") < CanonPath("foo!")); + ASSERT_FALSE(CanonPath("foo!") < CanonPath("foo")); + ASSERT_TRUE (CanonPath("foo") < CanonPath("foo!")); + } + + TEST(CanonPath, allowed) { + { + std::set<CanonPath> allowed { + CanonPath("foo/bar"), + CanonPath("foo!"), + CanonPath("xyzzy"), + CanonPath("a/b/c"), + }; + + ASSERT_TRUE (CanonPath("foo/bar").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("foo/bar/bla").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("foo").isAllowed(allowed)); + ASSERT_FALSE(CanonPath("bar").isAllowed(allowed)); + ASSERT_FALSE(CanonPath("bar/a").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("a").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("a/b").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("a/b/c").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("a/b/c/d").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("a/b/c/d/e").isAllowed(allowed)); + ASSERT_FALSE(CanonPath("a/b/a").isAllowed(allowed)); + ASSERT_FALSE(CanonPath("a/b/d").isAllowed(allowed)); + ASSERT_FALSE(CanonPath("aaa").isAllowed(allowed)); + ASSERT_FALSE(CanonPath("zzz").isAllowed(allowed)); + ASSERT_TRUE (CanonPath("/").isAllowed(allowed)); + } + } +} diff --git a/src/libutil/tests/hash.cc b/src/libutil/tests/hash.cc index 412c03030..e4e928b3b 100644 --- a/src/libutil/tests/hash.cc +++ b/src/libutil/tests/hash.cc @@ -1,5 +1,12 @@ -#include "hash.hh" +#include <regex> + +#include <nlohmann/json.hpp> #include <gtest/gtest.h> +#include <rapidcheck/gtest.h> + +#include <hash.hh> + +#include "tests/hash.hh" namespace nix { @@ -73,3 +80,16 @@ namespace nix { "c7d329eeb6dd26545e96e55b874be909"); } } + +namespace rc { +using namespace nix; + +Gen<Hash> Arbitrary<Hash>::arbitrary() +{ + Hash hash(htSHA1); + for (size_t i = 0; i < hash.hashSize; ++i) + hash.hash[i] = *gen::arbitrary<uint8_t>(); + return gen::just(hash); +} + +} diff --git a/src/libutil/tests/hash.hh b/src/libutil/tests/hash.hh new file mode 100644 index 000000000..9e9650e6e --- /dev/null +++ b/src/libutil/tests/hash.hh @@ -0,0 +1,15 @@ +#pragma once + +#include <rapidcheck/gen/Arbitrary.h> + +#include <hash.hh> + +namespace rc { +using namespace nix; + +template<> +struct Arbitrary<Hash> { + static Gen<Hash> arbitrary(); +}; + +} diff --git a/src/libutil/tests/json.cc b/src/libutil/tests/json.cc deleted file mode 100644 index 156286999..000000000 --- a/src/libutil/tests/json.cc +++ /dev/null @@ -1,193 +0,0 @@ -#include "json.hh" -#include <gtest/gtest.h> -#include <sstream> - -namespace nix { - - /* ---------------------------------------------------------------------------- - * toJSON - * --------------------------------------------------------------------------*/ - - TEST(toJSON, quotesCharPtr) { - const char* input = "test"; - std::stringstream out; - toJSON(out, input); - - ASSERT_EQ(out.str(), "\"test\""); - } - - TEST(toJSON, quotesStdString) { - std::string input = "test"; - std::stringstream out; - toJSON(out, input); - - ASSERT_EQ(out.str(), "\"test\""); - } - - TEST(toJSON, convertsNullptrtoNull) { - auto input = nullptr; - std::stringstream out; - toJSON(out, input); - - ASSERT_EQ(out.str(), "null"); - } - - TEST(toJSON, convertsNullToNull) { - const char* input = 0; - std::stringstream out; - toJSON(out, input); - - ASSERT_EQ(out.str(), "null"); - } - - - TEST(toJSON, convertsFloat) { - auto input = 1.024f; - std::stringstream out; - toJSON(out, input); - - ASSERT_EQ(out.str(), "1.024"); - } - - TEST(toJSON, convertsDouble) { - const double input = 1.024; - std::stringstream out; - toJSON(out, input); - - ASSERT_EQ(out.str(), "1.024"); - } - - TEST(toJSON, convertsBool) { - auto input = false; - std::stringstream out; - toJSON(out, input); - - ASSERT_EQ(out.str(), "false"); - } - - TEST(toJSON, quotesTab) { - std::stringstream out; - toJSON(out, "\t"); - - ASSERT_EQ(out.str(), "\"\\t\""); - } - - TEST(toJSON, quotesNewline) { - std::stringstream out; - toJSON(out, "\n"); - - ASSERT_EQ(out.str(), "\"\\n\""); - } - - TEST(toJSON, quotesCreturn) { - std::stringstream out; - toJSON(out, "\r"); - - ASSERT_EQ(out.str(), "\"\\r\""); - } - - TEST(toJSON, quotesCreturnNewLine) { - std::stringstream out; - toJSON(out, "\r\n"); - - ASSERT_EQ(out.str(), "\"\\r\\n\""); - } - - TEST(toJSON, quotesDoublequotes) { - std::stringstream out; - toJSON(out, "\""); - - ASSERT_EQ(out.str(), "\"\\\"\""); - } - - TEST(toJSON, substringEscape) { - std::stringstream out; - std::string_view s = "foo\t"; - toJSON(out, s.substr(3)); - - ASSERT_EQ(out.str(), "\"\\t\""); - } - - /* ---------------------------------------------------------------------------- - * JSONObject - * --------------------------------------------------------------------------*/ - - TEST(JSONObject, emptyObject) { - std::stringstream out; - { - JSONObject t(out); - } - ASSERT_EQ(out.str(), "{}"); - } - - TEST(JSONObject, objectWithList) { - std::stringstream out; - { - JSONObject t(out); - auto l = t.list("list"); - l.elem("element"); - } - ASSERT_EQ(out.str(), R"#({"list":["element"]})#"); - } - - TEST(JSONObject, objectWithListIndent) { - std::stringstream out; - { - JSONObject t(out, true); - auto l = t.list("list"); - l.elem("element"); - } - ASSERT_EQ(out.str(), -R"#({ - "list": [ - "element" - ] -})#"); - } - - TEST(JSONObject, objectWithPlaceholderAndList) { - std::stringstream out; - { - JSONObject t(out); - auto l = t.placeholder("list"); - l.list().elem("element"); - } - - ASSERT_EQ(out.str(), R"#({"list":["element"]})#"); - } - - TEST(JSONObject, objectWithPlaceholderAndObject) { - std::stringstream out; - { - JSONObject t(out); - auto l = t.placeholder("object"); - l.object().attr("key", "value"); - } - - ASSERT_EQ(out.str(), R"#({"object":{"key":"value"}})#"); - } - - /* ---------------------------------------------------------------------------- - * JSONList - * --------------------------------------------------------------------------*/ - - TEST(JSONList, empty) { - std::stringstream out; - { - JSONList l(out); - } - ASSERT_EQ(out.str(), R"#([])#"); - } - - TEST(JSONList, withElements) { - std::stringstream out; - { - JSONList l(out); - l.elem("one"); - l.object(); - l.placeholder().write("three"); - } - ASSERT_EQ(out.str(), R"#(["one",{},"three"])#"); - } -} - diff --git a/src/libutil/tests/local.mk b/src/libutil/tests/local.mk index 815e18560..167915439 100644 --- a/src/libutil/tests/local.mk +++ b/src/libutil/tests/local.mk @@ -2,14 +2,28 @@ check: libutil-tests_RUN programs += libutil-tests +libutil-tests-exe_NAME = libnixutil-tests + +libutil-tests-exe_DIR := $(d) + +libutil-tests-exe_INSTALL_DIR := + +libutil-tests-exe_LIBS = libutil-tests + +libutil-tests-exe_LDFLAGS := $(GTEST_LIBS) + +libraries += libutil-tests + +libutil-tests_NAME = libnixutil-tests + libutil-tests_DIR := $(d) libutil-tests_INSTALL_DIR := libutil-tests_SOURCES := $(wildcard $(d)/*.cc) -libutil-tests_CXXFLAGS += -I src/libutil -I src/libexpr +libutil-tests_CXXFLAGS += -I src/libutil libutil-tests_LIBS = libutil -libutil-tests_LDFLAGS := $(GTEST_LIBS) +libutil-tests_LDFLAGS := -lrapidcheck $(GTEST_LIBS) diff --git a/src/libutil/tests/tests.cc b/src/libutil/tests/tests.cc index 6e325db98..250e83a38 100644 --- a/src/libutil/tests/tests.cc +++ b/src/libutil/tests/tests.cc @@ -312,6 +312,42 @@ namespace nix { } /* ---------------------------------------------------------------------------- + * getLine + * --------------------------------------------------------------------------*/ + + TEST(getLine, all) { + { + auto [line, rest] = getLine("foo\nbar\nxyzzy"); + ASSERT_EQ(line, "foo"); + ASSERT_EQ(rest, "bar\nxyzzy"); + } + + { + auto [line, rest] = getLine("foo\r\nbar\r\nxyzzy"); + ASSERT_EQ(line, "foo"); + ASSERT_EQ(rest, "bar\r\nxyzzy"); + } + + { + auto [line, rest] = getLine("foo\n"); + ASSERT_EQ(line, "foo"); + ASSERT_EQ(rest, ""); + } + + { + auto [line, rest] = getLine("foo"); + ASSERT_EQ(line, "foo"); + ASSERT_EQ(rest, ""); + } + + { + auto [line, rest] = getLine(""); + ASSERT_EQ(line, ""); + ASSERT_EQ(rest, ""); + } + } + + /* ---------------------------------------------------------------------------- * toLower * --------------------------------------------------------------------------*/ diff --git a/src/libutil/tests/url.cc b/src/libutil/tests/url.cc index c3b233797..a908631e6 100644 --- a/src/libutil/tests/url.cc +++ b/src/libutil/tests/url.cc @@ -99,6 +99,27 @@ namespace nix { ASSERT_EQ(parsed, expected); } + TEST(parseURL, parsesFilePlusHttpsUrl) { + auto s = "file+https://www.example.org/video.mp4"; + auto parsed = parseURL(s); + + ParsedURL expected { + .url = "file+https://www.example.org/video.mp4", + .base = "https://www.example.org/video.mp4", + .scheme = "file+https", + .authority = "www.example.org", + .path = "/video.mp4", + .query = (StringMap) { }, + .fragment = "", + }; + + ASSERT_EQ(parsed, expected); + } + + TEST(parseURL, rejectsAuthorityInUrlsWithFileTransportation) { + auto s = "file://www.example.org/video.mp4"; + ASSERT_THROW(parseURL(s), Error); + } TEST(parseURL, parseIPv4Address) { auto s = "http://127.0.0.1:8080/file.tar.gz?download=fast&when=now#hello"; @@ -281,4 +302,37 @@ namespace nix { ASSERT_EQ(d, s); } + + /* ---------------------------------------------------------------------------- + * percentEncode + * --------------------------------------------------------------------------*/ + + TEST(percentEncode, encodesUrlEncodedString) { + std::string s = percentEncode("==@=="); + std::string d = "%3D%3D%40%3D%3D"; + ASSERT_EQ(d, s); + } + + TEST(percentEncode, keepArgument) { + std::string a = percentEncode("abd / def"); + std::string b = percentEncode("abd / def", "/"); + ASSERT_EQ(a, "abd%20%2F%20def"); + ASSERT_EQ(b, "abd%20/%20def"); + } + + TEST(percentEncode, inverseOfDecode) { + std::string original = "%3D%3D%40%3D%3D"; + std::string once = percentEncode(original); + std::string back = percentDecode(once); + + ASSERT_EQ(back, original); + } + + TEST(percentEncode, trailingPercent) { + std::string s = percentEncode("==@==%"); + std::string d = "%3D%3D%40%3D%3D%25"; + + ASSERT_EQ(d, s); + } + } diff --git a/src/libutil/url.cc b/src/libutil/url.cc index 5b7abeb49..9e44241ac 100644 --- a/src/libutil/url.cc +++ b/src/libutil/url.cc @@ -30,13 +30,13 @@ ParsedURL parseURL(const std::string & url) auto & query = match[6]; auto & fragment = match[7]; - auto isFile = scheme.find("file") != std::string::npos; + auto transportIsFile = parseUrlScheme(scheme).transport == "file"; - if (authority && *authority != "" && isFile) + if (authority && *authority != "" && transportIsFile) throw BadURL("file:// URL '%s' has unexpected authority '%s'", url, *authority); - if (isFile && path.empty()) + if (transportIsFile && path.empty()) path = "/"; return ParsedURL{ @@ -88,17 +88,22 @@ std::map<std::string, std::string> decodeQuery(const std::string & query) return result; } -std::string percentEncode(std::string_view s) +const static std::string allowedInQuery = ":@/?"; +const static std::string allowedInPath = ":@/"; + +std::string percentEncode(std::string_view s, std::string_view keep) { std::string res; for (auto & c : s) + // unreserved + keep if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || (c >= '0' && c <= '9') - || strchr("-._~!$&'()*+,;=:@", c)) + || strchr("-._~", c) + || keep.find(c) != std::string::npos) res += c; else - res += fmt("%%%02x", (unsigned int) c); + res += fmt("%%%02X", (unsigned int) c); return res; } @@ -109,9 +114,9 @@ std::string encodeQuery(const std::map<std::string, std::string> & ss) for (auto & [name, value] : ss) { if (!first) res += '&'; first = false; - res += percentEncode(name); + res += percentEncode(name, allowedInQuery); res += '='; - res += percentEncode(value); + res += percentEncode(value, allowedInQuery); } return res; } @@ -122,7 +127,7 @@ std::string ParsedURL::to_string() const scheme + ":" + (authority ? "//" + *authority : "") - + path + + percentEncode(path, allowedInPath) + (query.empty() ? "" : "?" + encodeQuery(query)) + (fragment.empty() ? "" : "#" + percentEncode(fragment)); } diff --git a/src/libutil/url.hh b/src/libutil/url.hh index 2a9fb34c1..ddd673d65 100644 --- a/src/libutil/url.hh +++ b/src/libutil/url.hh @@ -22,6 +22,7 @@ struct ParsedURL MakeError(BadURL, Error); std::string percentDecode(std::string_view in); +std::string percentEncode(std::string_view s, std::string_view keep=""); std::map<std::string, std::string> decodeQuery(const std::string & query); diff --git a/src/libutil/util.cc b/src/libutil/util.cc index be6fe091f..843a10eab 100644 --- a/src/libutil/util.cc +++ b/src/libutil/util.cc @@ -2,6 +2,7 @@ #include "sync.hh" #include "finally.hh" #include "serialise.hh" +#include "cgroup.hh" #include <array> #include <cctype> @@ -35,8 +36,8 @@ #ifdef __linux__ #include <sys/prctl.h> #include <sys/resource.h> +#include <sys/mman.h> -#include <mntent.h> #include <cmath> #endif @@ -53,6 +54,11 @@ std::optional<std::string> getEnv(const std::string & key) return std::string(value); } +std::optional<std::string> getEnvNonEmpty(const std::string & key) { + auto value = getEnv(key); + if (value == "") return {}; + return value; +} std::map<std::string, std::string> getEnv() { @@ -353,7 +359,7 @@ void readFile(const Path & path, Sink & sink) } -void writeFile(const Path & path, std::string_view s, mode_t mode) +void writeFile(const Path & path, std::string_view s, mode_t mode, bool sync) { AutoCloseFD fd = open(path.c_str(), O_WRONLY | O_TRUNC | O_CREAT | O_CLOEXEC, mode); if (!fd) @@ -364,10 +370,16 @@ void writeFile(const Path & path, std::string_view s, mode_t mode) e.addTrace({}, "writing file '%1%'", path); throw; } + if (sync) + fd.fsync(); + // Explicitly close to make sure exceptions are propagated. + fd.close(); + if (sync) + syncParent(path); } -void writeFile(const Path & path, Source & source, mode_t mode) +void writeFile(const Path & path, Source & source, mode_t mode, bool sync) { AutoCloseFD fd = open(path.c_str(), O_WRONLY | O_TRUNC | O_CREAT | O_CLOEXEC, mode); if (!fd) @@ -386,6 +398,20 @@ void writeFile(const Path & path, Source & source, mode_t mode) e.addTrace({}, "writing file '%1%'", path); throw; } + if (sync) + fd.fsync(); + // Explicitly close to make sure exceptions are propagated. + fd.close(); + if (sync) + syncParent(path); +} + +void syncParent(const Path & path) +{ + AutoCloseFD fd = open(dirOf(path).c_str(), O_RDONLY, 0); + if (!fd) + throw SysError("opening file '%1%'", path); + fd.fsync(); } std::string readLine(int fd) @@ -502,67 +528,12 @@ void deletePath(const Path & path) void deletePath(const Path & path, uint64_t & bytesFreed) { - //Activity act(*logger, lvlDebug, format("recursively deleting path '%1%'") % path); + //Activity act(*logger, lvlDebug, "recursively deleting path '%1%'", path); bytesFreed = 0; _deletePath(path, bytesFreed); } -static Path tempName(Path tmpRoot, const Path & prefix, bool includePid, - int & counter) -{ - tmpRoot = canonPath(tmpRoot.empty() ? getEnv("TMPDIR").value_or("/tmp") : tmpRoot, true); - if (includePid) - return (format("%1%/%2%-%3%-%4%") % tmpRoot % prefix % getpid() % counter++).str(); - else - return (format("%1%/%2%-%3%") % tmpRoot % prefix % counter++).str(); -} - - -Path createTempDir(const Path & tmpRoot, const Path & prefix, - bool includePid, bool useGlobalCounter, mode_t mode) -{ - static int globalCounter = 0; - int localCounter = 0; - int & counter(useGlobalCounter ? globalCounter : localCounter); - - while (1) { - checkInterrupt(); - Path tmpDir = tempName(tmpRoot, prefix, includePid, counter); - if (mkdir(tmpDir.c_str(), mode) == 0) { -#if __FreeBSD__ - /* Explicitly set the group of the directory. This is to - work around around problems caused by BSD's group - ownership semantics (directories inherit the group of - the parent). For instance, the group of /tmp on - FreeBSD is "wheel", so all directories created in /tmp - will be owned by "wheel"; but if the user is not in - "wheel", then "tar" will fail to unpack archives that - have the setgid bit set on directories. */ - if (chown(tmpDir.c_str(), (uid_t) -1, getegid()) != 0) - throw SysError("setting group of directory '%1%'", tmpDir); -#endif - return tmpDir; - } - if (errno != EEXIST) - throw SysError("creating directory '%1%'", tmpDir); - } -} - - -std::pair<AutoCloseFD, Path> createTempFile(const Path & prefix) -{ - Path tmpl(getEnv("TMPDIR").value_or("/tmp") + "/" + prefix + ".XXXXXX"); - // Strictly speaking, this is UB, but who cares... - // FIXME: use O_TMPFILE. - AutoCloseFD fd(mkstemp((char *) tmpl.c_str())); - if (!fd) - throw SysError("creating temporary file '%s'", tmpl); - closeOnExec(fd.get()); - return {std::move(fd), tmpl}; -} - - std::string getUserName() { auto pw = getpwuid(geteuid()); @@ -572,11 +543,22 @@ std::string getUserName() return name; } +Path getHomeOf(uid_t userId) +{ + std::vector<char> buf(16384); + struct passwd pwbuf; + struct passwd * pw; + if (getpwuid_r(userId, &pwbuf, buf.data(), buf.size(), &pw) != 0 + || !pw || !pw->pw_dir || !pw->pw_dir[0]) + throw Error("cannot determine user's home directory"); + return pw->pw_dir; +} Path getHome() { static Path homeDir = []() { + std::optional<std::string> unownedUserHomeDir = {}; auto homeDir = getEnv("HOME"); if (homeDir) { // Only use $HOME if doesn't exist or is owned by the current user. @@ -588,18 +570,14 @@ Path getHome() homeDir.reset(); } } else if (st.st_uid != geteuid()) { - warn("$HOME ('%s') is not owned by you, falling back to the one defined in the 'passwd' file", *homeDir); - homeDir.reset(); + unownedUserHomeDir.swap(homeDir); } } if (!homeDir) { - std::vector<char> buf(16384); - struct passwd pwbuf; - struct passwd * pw; - if (getpwuid_r(geteuid(), &pwbuf, buf.data(), buf.size(), &pw) != 0 - || !pw || !pw->pw_dir || !pw->pw_dir[0]) - throw Error("cannot determine user's home directory"); - homeDir = pw->pw_dir; + homeDir = getHomeOf(geteuid()); + if (unownedUserHomeDir.has_value() && unownedUserHomeDir != homeDir) { + warn("$HOME ('%s') is not owned by you, falling back to the one defined in the 'passwd' file ('%s')", *unownedUserHomeDir, *homeDir); + } } return *homeDir; }(); @@ -636,6 +614,19 @@ Path getDataDir() return dataDir ? *dataDir : getHome() + "/.local/share"; } +Path getStateDir() +{ + auto stateDir = getEnv("XDG_STATE_HOME"); + return stateDir ? *stateDir : getHome() + "/.local/state"; +} + +Path createNixStateDir() +{ + Path dir = getStateDir() + "/nix"; + createDirs(dir); + return dir; +} + std::optional<Path> getSelfExe() { @@ -681,44 +672,6 @@ Paths createDirs(const Path & path) } -void createSymlink(const Path & target, const Path & link, - std::optional<time_t> mtime) -{ - if (symlink(target.c_str(), link.c_str())) - throw SysError("creating symlink from '%1%' to '%2%'", link, target); - if (mtime) { - struct timeval times[2]; - times[0].tv_sec = *mtime; - times[0].tv_usec = 0; - times[1].tv_sec = *mtime; - times[1].tv_usec = 0; - if (lutimes(link.c_str(), times)) - throw SysError("setting time of symlink '%s'", link); - } -} - - -void replaceSymlink(const Path & target, const Path & link, - std::optional<time_t> mtime) -{ - for (unsigned int n = 0; true; n++) { - Path tmp = canonPath(fmt("%s/.%d_%s", dirOf(link), n, baseNameOf(link))); - - try { - createSymlink(target, tmp, mtime); - } catch (SysError & e) { - if (e.errNo == EEXIST) continue; - throw; - } - - if (rename(tmp.c_str(), link.c_str()) != 0) - throw SysError("renaming '%1%' to '%2%'", tmp, link); - - break; - } -} - - void readFull(int fd, char * buf, size_t count) { while (count) { @@ -797,45 +750,22 @@ unsigned int getMaxCPU() { #if __linux__ try { - FILE *fp = fopen("/proc/mounts", "r"); - if (!fp) - return 0; - - Strings cgPathParts; - - struct mntent *ent; - while ((ent = getmntent(fp))) { - std::string mountType, mountPath; - - mountType = ent->mnt_type; - mountPath = ent->mnt_dir; - - if (mountType == "cgroup2") { - cgPathParts.push_back(mountPath); - break; - } - } - - fclose(fp); - - if (cgPathParts.size() > 0 && pathExists("/proc/self/cgroup")) { - std::string currentCgroup = readFile("/proc/self/cgroup"); - Strings cgValues = tokenizeString<Strings>(currentCgroup, ":"); - cgPathParts.push_back(trim(cgValues.back(), "\n")); - cgPathParts.push_back("cpu.max"); - std::string fullCgPath = canonPath(concatStringsSep("/", cgPathParts)); - - if (pathExists(fullCgPath)) { - std::string cpuMax = readFile(fullCgPath); - std::vector<std::string> cpuMaxParts = tokenizeString<std::vector<std::string>>(cpuMax, " "); - std::string quota = cpuMaxParts[0]; - std::string period = trim(cpuMaxParts[1], "\n"); - - if (quota != "max") - return std::ceil(std::stoi(quota) / std::stof(period)); - } - } - } catch (Error &) { ignoreException(); } + auto cgroupFS = getCgroupFS(); + if (!cgroupFS) return 0; + + auto cgroups = getCgroups("/proc/self/cgroup"); + auto cgroup = cgroups[""]; + if (cgroup == "") return 0; + + auto cpuFile = *cgroupFS + "/" + cgroup + "/cpu.max"; + + auto cpuMax = readFile(cpuFile); + auto cpuMaxParts = tokenizeString<std::vector<std::string>>(cpuMax, " \n"); + auto quota = cpuMaxParts[0]; + auto period = cpuMaxParts[1]; + if (quota != "max") + return std::ceil(std::stoi(quota) / std::stof(period)); + } catch (Error &) { ignoreException(lvlDebug); } #endif return 0; @@ -931,6 +861,20 @@ void AutoCloseFD::close() } } +void AutoCloseFD::fsync() +{ + if (fd != -1) { + int result; +#if __APPLE__ + result = ::fcntl(fd, F_FULLFSYNC); +#else + result = ::fsync(fd); +#endif + if (result == -1) + throw SysError("fsync file descriptor %1%", fd); + } +} + AutoCloseFD::operator bool() const { @@ -1126,9 +1070,19 @@ static pid_t doFork(bool allowVfork, std::function<void()> fun) } +#if __linux__ +static int childEntry(void * arg) +{ + auto main = (std::function<void()> *) arg; + (*main)(); + return 1; +} +#endif + + pid_t startProcess(std::function<void()> fun, const ProcessOptions & options) { - auto wrapper = [&]() { + std::function<void()> wrapper = [&]() { if (!options.allowVfork) logger = makeSimpleLogger(); try { @@ -1148,7 +1102,27 @@ pid_t startProcess(std::function<void()> fun, const ProcessOptions & options) _exit(1); }; - pid_t pid = doFork(options.allowVfork, wrapper); + pid_t pid = -1; + + if (options.cloneFlags) { + #ifdef __linux__ + // Not supported, since then we don't know when to free the stack. + assert(!(options.cloneFlags & CLONE_VM)); + + size_t stackSize = 1 * 1024 * 1024; + auto stack = (char *) mmap(0, stackSize, + PROT_WRITE | PROT_READ, MAP_PRIVATE | MAP_ANONYMOUS | MAP_STACK, -1, 0); + if (stack == MAP_FAILED) throw SysError("allocating stack"); + + Finally freeStack([&]() { munmap(stack, stackSize); }); + + pid = clone(childEntry, stack + stackSize, options.cloneFlags | SIGCHLD, &wrapper); + #else + throw Error("clone flags are only supported on Linux"); + #endif + } else + pid = doFork(options.allowVfork, wrapper); + if (pid == -1) throw SysError("unable to fork"); return pid; @@ -1427,14 +1401,14 @@ std::string statusToString(int status) { if (!WIFEXITED(status) || WEXITSTATUS(status) != 0) { if (WIFEXITED(status)) - return (format("failed with exit code %1%") % WEXITSTATUS(status)).str(); + return fmt("failed with exit code %1%", WEXITSTATUS(status)); else if (WIFSIGNALED(status)) { int sig = WTERMSIG(status); #if HAVE_STRSIGNAL const char * description = strsignal(sig); - return (format("failed due to signal %1% (%2%)") % sig % description).str(); + return fmt("failed due to signal %1% (%2%)", sig, description); #else - return (format("failed due to signal %1%") % sig).str(); + return fmt("failed due to signal %1%", sig); #endif } else @@ -1483,7 +1457,7 @@ std::string shellEscape(const std::string_view s) } -void ignoreException() +void ignoreException(Verbosity lvl) { /* Make sure no exceptions leave this function. printError() also throws when remote is closed. */ @@ -1491,7 +1465,7 @@ void ignoreException() try { throw; } catch (std::exception & e) { - printError("error (ignored): %1%", e.what()); + printMsg(lvl, "error (ignored): %1%", e.what()); } } catch (...) { } } @@ -1503,7 +1477,7 @@ bool shouldANSI() && !getEnv("NO_COLOR").has_value(); } -std::string filterANSIEscapes(const std::string & s, bool filterAll, unsigned int width) +std::string filterANSIEscapes(std::string_view s, bool filterAll, unsigned int width) { std::string t, e; size_t w = 0; @@ -1673,6 +1647,21 @@ std::string stripIndentation(std::string_view s) } +std::pair<std::string_view, std::string_view> getLine(std::string_view s) +{ + auto newline = s.find('\n'); + + if (newline == s.npos) { + return {s, ""}; + } else { + auto line = s.substr(0, newline); + if (!line.empty() && line[line.size() - 1] == '\r') + line = line.substr(0, line.size() - 1); + return {line, s.substr(newline + 1)}; + } +} + + ////////////////////////////////////////////////////////////////////// static Sync<std::pair<unsigned short, unsigned short>> windowSize{{0, 0}}; @@ -1979,7 +1968,7 @@ std::string showBytes(uint64_t bytes) // FIXME: move to libstore/build -void commonChildInit(Pipe & logPipe) +void commonChildInit() { logger = makeSimpleLogger(); @@ -1993,10 +1982,6 @@ void commonChildInit(Pipe & logPipe) if (setsid() == -1) throw SysError("creating a new session"); - /* Dup the write side of the logger pipe into stderr. */ - if (dup2(logPipe.writeSide.get(), STDERR_FILENO) == -1) - throw SysError("cannot pipe standard error into log file"); - /* Dup stderr to stdout. */ if (dup2(STDERR_FILENO, STDOUT_FILENO) == -1) throw SysError("cannot dup stderr into stdout"); diff --git a/src/libutil/util.hh b/src/libutil/util.hh index 29227ecc6..a8e6f9fbb 100644 --- a/src/libutil/util.hh +++ b/src/libutil/util.hh @@ -39,6 +39,10 @@ extern const std::string nativeSystem; /* Return an environment variable. */ std::optional<std::string> getEnv(const std::string & key); +/* Return a non empty environment variable. Returns nullopt if the env +variable is set to "" */ +std::optional<std::string> getEnvNonEmpty(const std::string & key); + /* Get the entire environment. */ std::map<std::string, std::string> getEnv(); @@ -115,9 +119,12 @@ std::string readFile(const Path & path); void readFile(const Path & path, Sink & sink); /* Write a string to a file. */ -void writeFile(const Path & path, std::string_view s, mode_t mode = 0666); +void writeFile(const Path & path, std::string_view s, mode_t mode = 0666, bool sync = false); + +void writeFile(const Path & path, Source & source, mode_t mode = 0666, bool sync = false); -void writeFile(const Path & path, Source & source, mode_t mode = 0666); +/* Flush a file's parent directory to disk */ +void syncParent(const Path & path); /* Read a line from a file descriptor. */ std::string readLine(int fd); @@ -134,6 +141,9 @@ void deletePath(const Path & path, uint64_t & bytesFreed); std::string getUserName(); +/* Return the given user's home directory from /etc/passwd. */ +Path getHomeOf(uid_t userId); + /* Return $HOME or the user's home directory from /etc/passwd. */ Path getHome(); @@ -152,6 +162,12 @@ Path getDataDir(); /* Return the path of the current executable. */ std::optional<Path> getSelfExe(); +/* Return $XDG_STATE_HOME or $HOME/.local/state. */ +Path getStateDir(); + +/* Create the Nix state directory and return the path to it. */ +Path createNixStateDir(); + /* Create a directory and all its parents, if necessary. Returns the list of created directories, in order of creation. */ Paths createDirs(const Path & path); @@ -168,6 +184,17 @@ void createSymlink(const Path & target, const Path & link, void replaceSymlink(const Path & target, const Path & link, std::optional<time_t> mtime = {}); +void renameFile(const Path & src, const Path & dst); + +/** + * Similar to 'renameFile', but fallback to a copy+remove if `src` and `dst` + * are on a different filesystem. + * + * Beware that this might not be atomic because of the copy that happens behind + * the scenes + */ +void moveFile(const Path & src, const Path & dst); + /* Wrappers arount read()/write() that read/write exactly the requested number of bytes. */ @@ -220,6 +247,7 @@ public: explicit operator bool() const; int release(); void close(); + void fsync(); }; @@ -283,6 +311,7 @@ struct ProcessOptions bool dieWithParent = true; bool runExitHandlers = false; bool allowVfork = false; + int cloneFlags = 0; // use clone() with the specified flags (Linux only) }; pid_t startProcess(std::function<void()> fun, const ProcessOptions & options = ProcessOptions()); @@ -495,6 +524,18 @@ std::optional<N> string2Float(const std::string_view s) } +/* Convert a little-endian integer to host order. */ +template<typename T> +T readLittleEndian(unsigned char * p) +{ + T x = 0; + for (size_t i = 0; i < sizeof(x); ++i, ++p) { + x |= ((T) *p) << (i * 8); + } + return x; +} + + /* Return true iff `s' starts with `prefix'. */ bool hasPrefix(std::string_view s, std::string_view prefix); @@ -513,7 +554,7 @@ std::string shellEscape(const std::string_view s); /* Exception handling in destructors: print an error message, then ignore the exception. */ -void ignoreException(); +void ignoreException(Verbosity lvl = lvlError); @@ -532,7 +573,7 @@ bool shouldANSI(); some escape sequences (such as colour setting) are copied but not included in the character count. Also, tabs are expanded to spaces. */ -std::string filterANSIEscapes(const std::string & s, +std::string filterANSIEscapes(std::string_view s, bool filterAll = false, unsigned int width = std::numeric_limits<unsigned int>::max()); @@ -548,6 +589,12 @@ std::string base64Decode(std::string_view s); std::string stripIndentation(std::string_view s); +/* Get the prefix of 's' up to and excluding the next line break (LF + optionally preceded by CR), and the remainder following the line + break. */ +std::pair<std::string_view, std::string_view> getLine(std::string_view s); + + /* Get a value for the specified key from an associate container. */ template <class T> const typename T::mapped_type * get(const T & map, const typename T::key_type & key) @@ -657,7 +704,7 @@ typedef std::function<bool(const Path & path)> PathFilter; extern PathFilter defaultPathFilter; /* Common initialisation performed in child processes. */ -void commonChildInit(Pipe & logPipe); +void commonChildInit(); /* Create a Unix domain socket. */ AutoCloseFD createUnixDomainSocket(); @@ -722,4 +769,13 @@ inline std::string operator + (std::string && s, std::string_view s2) return std::move(s); } +inline std::string operator + (std::string_view s1, const char * s2) +{ + std::string s; + s.reserve(s1.size() + strlen(s2)); + s.append(s1); + s.append(s2); + return s; +} + } diff --git a/src/nix-build/nix-build.cc b/src/nix-build/nix-build.cc index 7eb8c8f6a..a4b3b1f96 100644 --- a/src/nix-build/nix-build.cc +++ b/src/nix-build/nix-build.cc @@ -85,7 +85,6 @@ static void main_nix_build(int argc, char * * argv) Strings attrPaths; Strings left; RepairFlag repair = NoRepair; - Path gcRoot; BuildMode buildMode = bmNormal; bool readStdin = false; @@ -167,9 +166,6 @@ static void main_nix_build(int argc, char * * argv) else if (*arg == "--out-link" || *arg == "-o") outLink = getArg(*arg, arg, end); - else if (*arg == "--add-root") - gcRoot = getArg(*arg, arg, end); - else if (*arg == "--dry-run") dryRun = true; @@ -223,9 +219,9 @@ static void main_nix_build(int argc, char * * argv) // read the shebang to understand which packages to read from. Since // this is handled via nix-shell -p, we wrap our ruby script execution // in ruby -e 'load' which ignores the shebangs. - envCommand = (format("exec %1% %2% -e 'load(ARGV.shift)' -- %3% %4%") % execArgs % interpreter % shellEscape(script) % joined.str()).str(); + envCommand = fmt("exec %1% %2% -e 'load(ARGV.shift)' -- %3% %4%", execArgs, interpreter, shellEscape(script), joined.str()); } else { - envCommand = (format("exec %1% %2% %3% %4%") % execArgs % interpreter % shellEscape(script) % joined.str()).str(); + envCommand = fmt("exec %1% %2% %3% %4%", execArgs, interpreter, shellEscape(script), joined.str()); } } @@ -401,7 +397,7 @@ static void main_nix_build(int argc, char * * argv) auto bashDrv = drv->requireDrvPath(); pathsToBuild.push_back(DerivedPath::Built { .drvPath = bashDrv, - .outputs = {}, + .outputs = OutputsSpec::Names {"out"}, }); pathsToCopy.insert(bashDrv); shellDrv = bashDrv; @@ -425,7 +421,7 @@ static void main_nix_build(int argc, char * * argv) { pathsToBuild.push_back(DerivedPath::Built { .drvPath = inputDrv, - .outputs = inputOutputs + .outputs = OutputsSpec::Names { inputOutputs }, }); pathsToCopy.insert(inputDrv); } @@ -540,7 +536,9 @@ static void main_nix_build(int argc, char * * argv) "SHELL=%5%; " "BASH=%5%; " "set +e; " - R"s([ -n "$PS1" -a -z "$NIX_SHELL_PRESERVE_PROMPT" ] && PS1='\n\[\033[1;32m\][nix-shell:\w]\$\[\033[0m\] '; )s" + R"s([ -n "$PS1" -a -z "$NIX_SHELL_PRESERVE_PROMPT" ] && )s" + + (getuid() == 0 ? R"s(PS1='\n\[\033[1;31m\][nix-shell:\w]\$\[\033[0m\] '; )s" + : R"s(PS1='\n\[\033[1;32m\][nix-shell:\w]\$\[\033[0m\] '; )s") + "if [ \"$(type -t runHook)\" = function ]; then runHook shellHook; fi; " "unset NIX_ENFORCE_PURITY; " "shopt -u nullglob; " @@ -595,7 +593,7 @@ static void main_nix_build(int argc, char * * argv) if (outputName == "") throw Error("derivation '%s' lacks an 'outputName' attribute", store->printStorePath(drvPath)); - pathsToBuild.push_back(DerivedPath::Built{drvPath, {outputName}}); + pathsToBuild.push_back(DerivedPath::Built{drvPath, OutputsSpec::Names{outputName}}); pathsToBuildOrdered.push_back({drvPath, {outputName}}); drvsToCopy.insert(drvPath); diff --git a/src/nix-channel/nix-channel.cc b/src/nix-channel/nix-channel.cc index cf52b03b4..338a7d18e 100755 --- a/src/nix-channel/nix-channel.cc +++ b/src/nix-channel/nix-channel.cc @@ -1,9 +1,11 @@ +#include "profiles.hh" #include "shared.hh" #include "globals.hh" #include "filetransfer.hh" #include "store-api.hh" #include "legacy.hh" #include "fetchers.hh" +#include "util.hh" #include <fcntl.h> #include <regex> @@ -162,11 +164,11 @@ static int main_nix_channel(int argc, char ** argv) { // Figure out the name of the `.nix-channels' file to use auto home = getHome(); - channelsList = home + "/.nix-channels"; - nixDefExpr = home + "/.nix-defexpr"; + channelsList = settings.useXDGBaseDirectories ? createNixStateDir() + "/channels" : home + "/.nix-channels"; + nixDefExpr = settings.useXDGBaseDirectories ? createNixStateDir() + "/defexpr" : home + "/.nix-defexpr"; // Figure out the name of the channels profile. - profile = fmt("%s/profiles/per-user/%s/channels", settings.nixStateDir, getUserName()); + profile = profilesDir() + "/channels"; enum { cNone, diff --git a/src/nix-collect-garbage/nix-collect-garbage.cc b/src/nix-collect-garbage/nix-collect-garbage.cc index e413faffe..3cc57af4e 100644 --- a/src/nix-collect-garbage/nix-collect-garbage.cc +++ b/src/nix-collect-garbage/nix-collect-garbage.cc @@ -40,7 +40,7 @@ void removeOldGenerations(std::string dir) throw; } if (link.find("link") != std::string::npos) { - printInfo(format("removing old generations of profile %1%") % path); + printInfo("removing old generations of profile %s", path); if (deleteOlderThan != "") deleteGenerationsOlderThan(path, deleteOlderThan, dryRun); else @@ -77,8 +77,7 @@ static int main_nix_collect_garbage(int argc, char * * argv) return true; }); - auto profilesDir = settings.nixStateDir + "/profiles"; - if (removeOld) removeOldGenerations(profilesDir); + if (removeOld) removeOldGenerations(profilesDir()); // Run the actual garbage collector. if (!dryRun) { diff --git a/src/nix-copy-closure/nix-copy-closure.cc b/src/nix-copy-closure/nix-copy-closure.cc index 841d50fd3..7f2bb93b6 100755 --- a/src/nix-copy-closure/nix-copy-closure.cc +++ b/src/nix-copy-closure/nix-copy-closure.cc @@ -22,7 +22,7 @@ static int main_nix_copy_closure(int argc, char ** argv) printVersion("nix-copy-closure"); else if (*arg == "--gzip" || *arg == "--bzip2" || *arg == "--xz") { if (*arg != "--gzip") - printMsg(lvlError, format("Warning: '%1%' is not implemented, falling back to gzip") % *arg); + warn("'%1%' is not implemented, falling back to gzip", *arg); gzip = true; } else if (*arg == "--from") toMode = false; diff --git a/src/nix-env/nix-env.cc b/src/nix-env/nix-env.cc index a69d3700d..3a012638b 100644 --- a/src/nix-env/nix-env.cc +++ b/src/nix-env/nix-env.cc @@ -12,7 +12,6 @@ #include "local-fs-store.hh" #include "user-env.hh" #include "util.hh" -#include "json.hh" #include "value-to-json.hh" #include "xml-writer.hh" #include "legacy.hh" @@ -26,6 +25,7 @@ #include <sys/types.h> #include <sys/stat.h> #include <unistd.h> +#include <nlohmann/json.hpp> using namespace nix; using std::cout; @@ -478,9 +478,14 @@ static void printMissing(EvalState & state, DrvInfos & elems) std::vector<DerivedPath> targets; for (auto & i : elems) if (auto drvPath = i.queryDrvPath()) - targets.push_back(DerivedPath::Built{*drvPath}); + targets.push_back(DerivedPath::Built{ + .drvPath = *drvPath, + .outputs = OutputsSpec::All { }, + }); else - targets.push_back(DerivedPath::Opaque{i.queryOutPath()}); + targets.push_back(DerivedPath::Opaque{ + .path = i.queryOutPath(), + }); printMissing(state.store, targets); } @@ -495,7 +500,7 @@ static bool keep(DrvInfo & drv) static void installDerivations(Globals & globals, const Strings & args, const Path & profile) { - debug(format("installing derivations")); + debug("installing derivations"); /* Get the set of user environment elements to be installed. */ DrvInfos newElems, newElemsTmp; @@ -574,7 +579,7 @@ typedef enum { utLt, utLeq, utEq, utAlways } UpgradeType; static void upgradeDerivations(Globals & globals, const Strings & args, UpgradeType upgradeType) { - debug(format("upgrading derivations")); + debug("upgrading derivations"); /* Upgrade works as follows: we take all currently installed derivations, and for any derivation matching any selector, look @@ -647,7 +652,7 @@ static void upgradeDerivations(Globals & globals, } else newElems.push_back(i); } catch (Error & e) { - e.addTrace(std::nullopt, "while trying to find an upgrade for '%s'", i.queryName()); + e.addTrace(nullptr, "while trying to find an upgrade for '%s'", i.queryName()); throw; } } @@ -751,14 +756,19 @@ static void opSet(Globals & globals, Strings opFlags, Strings opArgs) auto drvPath = drv.queryDrvPath(); std::vector<DerivedPath> paths { drvPath - ? (DerivedPath) (DerivedPath::Built { *drvPath }) - : (DerivedPath) (DerivedPath::Opaque { drv.queryOutPath() }), + ? (DerivedPath) (DerivedPath::Built { + .drvPath = *drvPath, + .outputs = OutputsSpec::All { }, + }) + : (DerivedPath) (DerivedPath::Opaque { + .path = drv.queryOutPath(), + }), }; printMissing(globals.state->store, paths); if (globals.dryRun) return; globals.state->store->buildPaths(paths, globals.state->repair ? bmRepair : bmNormal); - debug(format("switching to new user environment")); + debug("switching to new user environment"); Path generation = createGeneration( ref<LocalFSStore>(store2), globals.profile, @@ -911,53 +921,58 @@ static VersionDiff compareVersionAgainstSet( static void queryJSON(Globals & globals, std::vector<DrvInfo> & elems, bool printOutPath, bool printMeta) { - JSONObject topObj(cout, true); + using nlohmann::json; + json topObj = json::object(); for (auto & i : elems) { try { if (i.hasFailed()) continue; - JSONObject pkgObj = topObj.object(i.attrPath); auto drvName = DrvName(i.queryName()); - pkgObj.attr("name", drvName.fullName); - pkgObj.attr("pname", drvName.name); - pkgObj.attr("version", drvName.version); - pkgObj.attr("system", i.querySystem()); - pkgObj.attr("outputName", i.queryOutputName()); + json &pkgObj = topObj[i.attrPath]; + pkgObj = { + {"name", drvName.fullName}, + {"pname", drvName.name}, + {"version", drvName.version}, + {"system", i.querySystem()}, + {"outputName", i.queryOutputName()}, + }; { DrvInfo::Outputs outputs = i.queryOutputs(printOutPath); - JSONObject outputObj = pkgObj.object("outputs"); + json &outputObj = pkgObj["outputs"]; + outputObj = json::object(); for (auto & j : outputs) { if (j.second) - outputObj.attr(j.first, globals.state->store->printStorePath(*j.second)); + outputObj[j.first] = globals.state->store->printStorePath(*j.second); else - outputObj.attr(j.first, nullptr); + outputObj[j.first] = nullptr; } } if (printMeta) { - JSONObject metaObj = pkgObj.object("meta"); + json &metaObj = pkgObj["meta"]; + metaObj = json::object(); StringSet metaNames = i.queryMetaNames(); for (auto & j : metaNames) { - auto placeholder = metaObj.placeholder(j); Value * v = i.queryMeta(j); if (!v) { printError("derivation '%s' has invalid meta attribute '%s'", i.queryName(), j); - placeholder.write(nullptr); + metaObj[j] = nullptr; } else { PathSet context; - printValueAsJSON(*globals.state, true, *v, noPos, placeholder, context); + metaObj[j] = printValueAsJSON(*globals.state, true, *v, noPos, context); } } } } catch (AssertionError & e) { printMsg(lvlTalkative, "skipping derivation named '%1%' which gives an assertion failure", i.queryName()); } catch (Error & e) { - e.addTrace(std::nullopt, "while querying the derivation named '%1%'", i.queryName()); + e.addTrace(nullptr, "while querying the derivation named '%1%'", i.queryName()); throw; } } + std::cout << topObj.dump(2); } @@ -1078,7 +1093,7 @@ static void opQuery(Globals & globals, Strings opFlags, Strings opArgs) try { if (i.hasFailed()) continue; - //Activity act(*logger, lvlDebug, format("outputting query result '%1%'") % i.attrPath); + //Activity act(*logger, lvlDebug, "outputting query result '%1%'", i.attrPath); if (globals.prebuiltOnly && !validPaths.count(i.queryOutPath()) && @@ -1214,11 +1229,11 @@ static void opQuery(Globals & globals, Strings opFlags, Strings opArgs) xml.writeEmptyElement("meta", attrs2); } else if (v->type() == nInt) { attrs2["type"] = "int"; - attrs2["value"] = (format("%1%") % v->integer).str(); + attrs2["value"] = fmt("%1%", v->integer); xml.writeEmptyElement("meta", attrs2); } else if (v->type() == nFloat) { attrs2["type"] = "float"; - attrs2["value"] = (format("%1%") % v->fpoint).str(); + attrs2["value"] = fmt("%1%", v->fpoint); xml.writeEmptyElement("meta", attrs2); } else if (v->type() == nBool) { attrs2["type"] = "bool"; @@ -1257,7 +1272,7 @@ static void opQuery(Globals & globals, Strings opFlags, Strings opArgs) } catch (AssertionError & e) { printMsg(lvlTalkative, "skipping derivation named '%1%' which gives an assertion failure", i.queryName()); } catch (Error & e) { - e.addTrace(std::nullopt, "while querying the derivation named '%1%'", i.queryName()); + e.addTrace(nullptr, "while querying the derivation named '%1%'", i.queryName()); throw; } } @@ -1274,7 +1289,7 @@ static void opSwitchProfile(Globals & globals, Strings opFlags, Strings opArgs) throw UsageError("exactly one argument expected"); Path profile = absPath(opArgs.front()); - Path profileLink = getHome() + "/.nix-profile"; + Path profileLink = settings.useXDGBaseDirectories ? createNixStateDir() + "/profile" : getHome() + "/.nix-profile"; switchLink(profileLink, profile); } @@ -1322,11 +1337,11 @@ static void opListGenerations(Globals & globals, Strings opFlags, Strings opArgs for (auto & i : gens) { tm t; if (!localtime_r(&i.creationTime, &t)) throw Error("cannot convert time"); - cout << format("%|4| %|4|-%|02|-%|02| %|02|:%|02|:%|02| %||\n") - % i.number - % (t.tm_year + 1900) % (t.tm_mon + 1) % t.tm_mday - % t.tm_hour % t.tm_min % t.tm_sec - % (i.number == curGen ? "(current)" : ""); + logger->cout("%|4| %|4|-%|02|-%|02| %|02|:%|02|:%|02| %||", + i.number, + t.tm_year + 1900, t.tm_mon + 1, t.tm_mday, + t.tm_hour, t.tm_min, t.tm_sec, + i.number == curGen ? "(current)" : ""); } } @@ -1378,7 +1393,10 @@ static int main_nix_env(int argc, char * * argv) Globals globals; globals.instSource.type = srcUnknown; - globals.instSource.nixExprPath = getHome() + "/.nix-defexpr"; + { + Path nixExprPath = settings.useXDGBaseDirectories ? createNixStateDir() + "/defexpr" : getHome() + "/.nix-defexpr"; + globals.instSource.nixExprPath = nixExprPath; + } globals.instSource.systemFilter = "*"; if (!pathExists(globals.instSource.nixExprPath)) { diff --git a/src/nix-env/user-env.cc b/src/nix-env/user-env.cc index 4b1202be3..745e9e174 100644 --- a/src/nix-env/user-env.cc +++ b/src/nix-env/user-env.cc @@ -41,7 +41,7 @@ bool createUserEnv(EvalState & state, DrvInfos & elems, if (auto drvPath = i.queryDrvPath()) drvsToBuild.push_back({*drvPath}); - debug(format("building user environment dependencies")); + debug("building user environment dependencies"); state.store->buildPaths( toDerivedPaths(drvsToBuild), state.repair ? bmRepair : bmNormal); @@ -134,9 +134,9 @@ bool createUserEnv(EvalState & state, DrvInfos & elems, state.forceValue(topLevel, [&]() { return topLevel.determinePos(noPos); }); PathSet context; Attr & aDrvPath(*topLevel.attrs->find(state.sDrvPath)); - auto topLevelDrv = state.coerceToStorePath(aDrvPath.pos, *aDrvPath.value, context); + auto topLevelDrv = state.coerceToStorePath(aDrvPath.pos, *aDrvPath.value, context, ""); Attr & aOutPath(*topLevel.attrs->find(state.sOutPath)); - auto topLevelOut = state.coerceToStorePath(aOutPath.pos, *aOutPath.value, context); + auto topLevelOut = state.coerceToStorePath(aOutPath.pos, *aOutPath.value, context, ""); /* Realise the resulting store expression. */ debug("building user environment"); @@ -159,7 +159,7 @@ bool createUserEnv(EvalState & state, DrvInfos & elems, return false; } - debug(format("switching to new user environment")); + debug("switching to new user environment"); Path generation = createGeneration(ref<LocalFSStore>(store2), profile, topLevelOut); switchLink(profile, generation); } diff --git a/src/nix-instantiate/nix-instantiate.cc b/src/nix-instantiate/nix-instantiate.cc index d3144e131..6b5ba595d 100644 --- a/src/nix-instantiate/nix-instantiate.cc +++ b/src/nix-instantiate/nix-instantiate.cc @@ -52,9 +52,10 @@ void processExpr(EvalState & state, const Strings & attrPaths, state.autoCallFunction(autoArgs, v, vRes); if (output == okXML) printValueAsXML(state, strict, location, vRes, std::cout, context, noPos); - else if (output == okJSON) + else if (output == okJSON) { printValueAsJSON(state, strict, vRes, v.determinePos(noPos), std::cout, context); - else { + std::cout << std::endl; + } else { if (strict) state.forceValueDeep(vRes); vRes.print(state.symbols, std::cout); std::cout << std::endl; diff --git a/src/nix-store/nix-store.cc b/src/nix-store/nix-store.cc index b453ea1ca..54479489f 100644 --- a/src/nix-store/nix-store.cc +++ b/src/nix-store/nix-store.cc @@ -72,11 +72,13 @@ static PathSet realisePath(StorePathWithOutputs path, bool build = true) Derivation drv = store->derivationFromPath(path.path); rootNr++; + /* FIXME: Encode this empty special case explicitly in the type. */ if (path.outputs.empty()) for (auto & i : drv.outputs) path.outputs.insert(i.first); PathSet outputs; for (auto & j : path.outputs) { + /* Match outputs of a store path with outputs of the derivation that produces it. */ DerivationOutputs::iterator i = drv.outputs.find(j); if (i == drv.outputs.end()) throw Error("derivation '%s' does not have an output named '%s'", @@ -141,6 +143,7 @@ static void opRealise(Strings opFlags, Strings opArgs) toDerivedPaths(paths), willBuild, willSubstitute, unknown, downloadSize, narSize); + /* Filter out unknown paths from `paths`. */ if (ignoreUnknown) { std::vector<StorePathWithOutputs> paths2; for (auto & i : paths) @@ -457,7 +460,7 @@ static void opPrintEnv(Strings opFlags, Strings opArgs) /* Print each environment variable in the derivation in a format * that can be sourced by the shell. */ for (auto & i : drv.env) - cout << format("export %1%; %1%=%2%\n") % i.first % shellEscape(i.second); + logger->cout("export %1%; %1%=%2%\n", i.first, shellEscape(i.second)); /* Also output the arguments. This doesn't preserve whitespace in arguments. */ @@ -516,7 +519,7 @@ static void registerValidity(bool reregister, bool hashGiven, bool canonicalise) if (!store->isValidPath(info->path) || reregister) { /* !!! races */ if (canonicalise) - canonicalisePathMetaData(store->printStorePath(info->path), -1); + canonicalisePathMetaData(store->printStorePath(info->path), {}); if (!hashGiven) { HashResult hash = hashPath(htSHA256, store->printStorePath(info->path)); info->narHash = hash.first; @@ -808,14 +811,23 @@ static void opServe(Strings opFlags, Strings opArgs) if (GET_PROTOCOL_MINOR(clientVersion) >= 2) settings.maxLogSize = readNum<unsigned long>(in); if (GET_PROTOCOL_MINOR(clientVersion) >= 3) { - settings.buildRepeat = readInt(in); - settings.enforceDeterminism = readInt(in); + auto nrRepeats = readInt(in); + if (nrRepeats != 0) { + throw Error("client requested repeating builds, but this is not currently implemented"); + } + // Ignore 'enforceDeterminism'. It used to be true by + // default, but also only never had any effect when + // `nrRepeats == 0`. We have already asserted that + // `nrRepeats` in fact is 0, so we can safely ignore this + // without doing something other than what the client + // asked for. + readInt(in); + settings.runDiffHook = true; } if (GET_PROTOCOL_MINOR(clientVersion) >= 7) { settings.keepFailed = (bool) readInt(in); } - settings.printRepeatedBuilds = false; }; while (true) { @@ -922,11 +934,10 @@ static void opServe(Strings opFlags, Strings opArgs) if (GET_PROTOCOL_MINOR(clientVersion) >= 3) out << status.timesBuilt << status.isNonDeterministic << status.startTime << status.stopTime; - if (GET_PROTOCOL_MINOR(clientVersion >= 6)) { + if (GET_PROTOCOL_MINOR(clientVersion) >= 6) { worker_proto::write(*store, out, status.builtOutputs); } - break; } @@ -1012,6 +1023,7 @@ static int main_nix_store(int argc, char * * argv) { Strings opFlags, opArgs; Operation op = 0; + bool readFromStdIn = false; parseCmdLine(argc, argv, [&](Strings::iterator & arg, const Strings::iterator & end) { Operation oldOp = op; @@ -1070,6 +1082,8 @@ static int main_nix_store(int argc, char * * argv) op = opGenerateBinaryCacheKey; else if (*arg == "--add-root") gcRoot = absPath(getArg(*arg, arg, end)); + else if (*arg == "--stdin" && !isatty(STDIN_FILENO)) + readFromStdIn = true; else if (*arg == "--indirect") ; else if (*arg == "--no-output") @@ -1082,6 +1096,13 @@ static int main_nix_store(int argc, char * * argv) else opArgs.push_back(*arg); + if (readFromStdIn && op != opImport && op != opRestore && op != opServe) { + std::string word; + while (std::cin >> word) { + opArgs.emplace_back(std::move(word)); + }; + } + if (oldOp && oldOp != op) throw UsageError("only one operation may be specified"); diff --git a/src/nix/app.cc b/src/nix/app.cc index 821964f86..bfd75e278 100644 --- a/src/nix/app.cc +++ b/src/nix/app.cc @@ -1,4 +1,5 @@ #include "installables.hh" +#include "installable-derived-path.hh" #include "store-api.hh" #include "eval-inline.hh" #include "eval-cache.hh" @@ -8,40 +9,17 @@ namespace nix { -struct InstallableDerivedPath : Installable -{ - ref<Store> store; - const DerivedPath derivedPath; - - InstallableDerivedPath(ref<Store> store, const DerivedPath & derivedPath) - : store(store) - , derivedPath(derivedPath) - { - } - - - std::string what() const override { return derivedPath.to_string(*store); } - - DerivedPaths toDerivedPaths() override - { - return {derivedPath}; - } - - std::optional<StorePath> getStorePath() override - { - return std::nullopt; - } -}; - /** * Return the rewrites that are needed to resolve a string whose context is * included in `dependencies`. */ -StringPairs resolveRewrites(Store & store, const BuiltPaths dependencies) +StringPairs resolveRewrites( + Store & store, + const std::vector<BuiltPathWithResult> & dependencies) { StringPairs res; for (auto & dep : dependencies) - if (auto drvDep = std::get_if<BuiltPathBuilt>(&dep)) + if (auto drvDep = std::get_if<BuiltPathBuilt>(&dep.path)) for (auto & [ outputName, outputPath ] : drvDep->outputs) res.emplace( downstreamPlaceholder(store, drvDep->drvPath, outputName), @@ -53,7 +31,10 @@ StringPairs resolveRewrites(Store & store, const BuiltPaths dependencies) /** * Resolve the given string assuming the given context. */ -std::string resolveString(Store & store, const std::string & toResolve, const BuiltPaths dependencies) +std::string resolveString( + Store & store, + const std::string & toResolve, + const std::vector<BuiltPathWithResult> & dependencies) { auto rewrites = resolveRewrites(store, dependencies); return rewriteStrings(toResolve, rewrites); @@ -66,16 +47,38 @@ UnresolvedApp Installable::toApp(EvalState & state) auto type = cursor->getAttr("type")->getString(); - std::string expected = !attrPath.empty() && state.symbols[attrPath[0]] == "apps" ? "app" : "derivation"; + std::string expected = !attrPath.empty() && + (state.symbols[attrPath[0]] == "apps" || state.symbols[attrPath[0]] == "defaultApp") + ? "app" : "derivation"; if (type != expected) throw Error("attribute '%s' should have type '%s'", cursor->getAttrPathStr(), expected); if (type == "app") { auto [program, context] = cursor->getAttr("program")->getStringWithContext(); - std::vector<StorePathWithOutputs> context2; - for (auto & [path, name] : context) - context2.push_back({path, {name}}); + std::vector<DerivedPath> context2; + for (auto & c : context) { + context2.emplace_back(std::visit(overloaded { + [&](const NixStringContextElem::DrvDeep & d) -> DerivedPath { + /* We want all outputs of the drv */ + return DerivedPath::Built { + .drvPath = d.drvPath, + .outputs = OutputsSpec::All {}, + }; + }, + [&](const NixStringContextElem::Built & b) -> DerivedPath { + return DerivedPath::Built { + .drvPath = b.drvPath, + .outputs = OutputsSpec::Names { b.output }, + }; + }, + [&](const NixStringContextElem::Opaque & o) -> DerivedPath { + return DerivedPath::Opaque { + .path = o.path, + }; + }, + }, c.raw())); + } return UnresolvedApp{App { .context = std::move(context2), @@ -99,7 +102,10 @@ UnresolvedApp Installable::toApp(EvalState & state) : DrvName(name).name; auto program = outPath + "/bin/" + mainProgram; return UnresolvedApp { App { - .context = { { drvPath, {outputName} } }, + .context = { DerivedPath::Built { + .drvPath = drvPath, + .outputs = OutputsSpec::Names { outputName }, + } }, .program = program, }}; } @@ -113,11 +119,11 @@ App UnresolvedApp::resolve(ref<Store> evalStore, ref<Store> store) { auto res = unresolved; - std::vector<std::shared_ptr<Installable>> installableContext; + Installables installableContext; for (auto & ctxElt : unresolved.context) installableContext.push_back( - std::make_shared<InstallableDerivedPath>(store, ctxElt.toDerivedPath())); + make_ref<InstallableDerivedPath>(store, DerivedPath { ctxElt })); auto builtContext = Installable::build(evalStore, store, Realise::Outputs, installableContext); res.program = resolveString(*store, unresolved.program, builtContext); diff --git a/src/nix/build.cc b/src/nix/build.cc index 9c648d28e..bca20e97c 100644 --- a/src/nix/build.cc +++ b/src/nix/build.cc @@ -10,6 +10,60 @@ using namespace nix; +nlohmann::json derivedPathsToJSON(const DerivedPaths & paths, ref<Store> store) +{ + auto res = nlohmann::json::array(); + for (auto & t : paths) { + std::visit([&res, store](const auto & t) { + res.push_back(t.toJSON(store)); + }, t.raw()); + } + return res; +} + +nlohmann::json builtPathsWithResultToJSON(const std::vector<BuiltPathWithResult> & buildables, ref<Store> store) +{ + auto res = nlohmann::json::array(); + for (auto & b : buildables) { + std::visit([&](const auto & t) { + auto j = t.toJSON(store); + if (b.result) { + j["startTime"] = b.result->startTime; + j["stopTime"] = b.result->stopTime; + if (b.result->cpuUser) + j["cpuUser"] = ((double) b.result->cpuUser->count()) / 1000000; + if (b.result->cpuSystem) + j["cpuSystem"] = ((double) b.result->cpuSystem->count()) / 1000000; + } + res.push_back(j); + }, b.path.raw()); + } + return res; +} + +// TODO deduplicate with other code also setting such out links. +static void createOutLinks(const Path& outLink, const std::vector<BuiltPathWithResult>& buildables, LocalFSStore& store2) +{ + for (const auto & [_i, buildable] : enumerate(buildables)) { + auto i = _i; + std::visit(overloaded { + [&](const BuiltPath::Opaque & bo) { + std::string symlink = outLink; + if (i) symlink += fmt("-%d", i); + store2.addPermRoot(bo.path, absPath(symlink)); + }, + [&](const BuiltPath::Built & bfd) { + for (auto & output : bfd.outputs) { + std::string symlink = outLink; + if (i) symlink += fmt("-%d", i); + if (output.first != "out") symlink += fmt("-%s", output.first); + store2.addPermRoot(output.second, absPath(symlink)); + } + }, + }, buildable.path.raw()); + } +} + struct CmdBuild : InstallablesCommand, MixDryRun, MixJSON, MixProfile { Path outLink = "result"; @@ -58,18 +112,20 @@ struct CmdBuild : InstallablesCommand, MixDryRun, MixJSON, MixProfile ; } - void run(ref<Store> store) override + void run(ref<Store> store, Installables && installables) override { if (dryRun) { std::vector<DerivedPath> pathsToBuild; - for (auto & i : installables) { - auto b = i->toDerivedPaths(); - pathsToBuild.insert(pathsToBuild.end(), b.begin(), b.end()); - } + for (auto & i : installables) + for (auto & b : i->toDerivedPaths()) + pathsToBuild.push_back(b.path); + printMissing(store, pathsToBuild, lvlError); + if (json) logger->cout("%s", derivedPathsToJSON(pathsToBuild, store).dump()); + return; } @@ -78,46 +134,32 @@ struct CmdBuild : InstallablesCommand, MixDryRun, MixJSON, MixProfile Realise::Outputs, installables, buildMode); - if (json) logger->cout("%s", derivedPathsWithHintsToJSON(buildables, store).dump()); + if (json) logger->cout("%s", builtPathsWithResultToJSON(buildables, store).dump()); if (outLink != "") if (auto store2 = store.dynamic_pointer_cast<LocalFSStore>()) - for (const auto & [_i, buildable] : enumerate(buildables)) { - auto i = _i; - std::visit(overloaded { - [&](const BuiltPath::Opaque & bo) { - std::string symlink = outLink; - if (i) symlink += fmt("-%d", i); - store2->addPermRoot(bo.path, absPath(symlink)); - }, - [&](const BuiltPath::Built & bfd) { - for (auto & output : bfd.outputs) { - std::string symlink = outLink; - if (i) symlink += fmt("-%d", i); - if (output.first != "out") symlink += fmt("-%s", output.first); - store2->addPermRoot(output.second, absPath(symlink)); - } - }, - }, buildable.raw()); - } + createOutLinks(outLink, buildables, *store2); if (printOutputPaths) { stopProgressBar(); for (auto & buildable : buildables) { std::visit(overloaded { [&](const BuiltPath::Opaque & bo) { - std::cout << store->printStorePath(bo.path) << std::endl; + logger->cout(store->printStorePath(bo.path)); }, [&](const BuiltPath::Built & bfd) { for (auto & output : bfd.outputs) { - std::cout << store->printStorePath(output.second) << std::endl; + logger->cout(store->printStorePath(output.second)); } }, - }, buildable.raw()); + }, buildable.path.raw()); } } - updateProfile(buildables); + BuiltPaths buildables2; + for (auto & b : buildables) + buildables2.push_back(b.path); + updateProfile(buildables2); } }; diff --git a/src/nix/build.md b/src/nix/build.md index 6a79f308c..ee414dc86 100644 --- a/src/nix/build.md +++ b/src/nix/build.md @@ -82,7 +82,7 @@ R""( # Description -`nix build` builds the specified *installables*. Installables that +`nix build` builds the specified *installables*. [Installables](./nix.md#installables) that resolve to derivations are built (or substituted if possible). Store path installables are substituted. diff --git a/src/nix/bundle.cc b/src/nix/bundle.cc index 2e48e4c74..973bbd423 100644 --- a/src/nix/bundle.cc +++ b/src/nix/bundle.cc @@ -1,4 +1,5 @@ #include "command.hh" +#include "installable-flake.hh" #include "common-args.hh" #include "shared.hh" #include "store-api.hh" @@ -69,16 +70,16 @@ struct CmdBundle : InstallableCommand return res; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { auto evalState = getEvalState(); auto val = installable->toValue(*evalState).first; - auto [bundlerFlakeRef, bundlerName, outputsSpec] = parseFlakeRefWithFragmentAndOutputsSpec(bundler, absPath(".")); + auto [bundlerFlakeRef, bundlerName, extendedOutputsSpec] = parseFlakeRefWithFragmentAndExtendedOutputsSpec(bundler, absPath(".")); const flake::LockFlags lockFlags{ .writeLockFile = false }; InstallableFlake bundler{this, - evalState, std::move(bundlerFlakeRef), bundlerName, outputsSpec, + evalState, std::move(bundlerFlakeRef), bundlerName, extendedOutputsSpec, {"bundlers." + settings.thisSystem.get() + ".default", "defaultBundler." + settings.thisSystem.get() }, @@ -97,15 +98,20 @@ struct CmdBundle : InstallableCommand throw Error("the bundler '%s' does not produce a derivation", bundler.what()); PathSet context2; - auto drvPath = evalState->coerceToStorePath(attr1->pos, *attr1->value, context2); + auto drvPath = evalState->coerceToStorePath(attr1->pos, *attr1->value, context2, ""); auto attr2 = vRes->attrs->get(evalState->sOutPath); if (!attr2) throw Error("the bundler '%s' does not produce a derivation", bundler.what()); - auto outPath = evalState->coerceToStorePath(attr2->pos, *attr2->value, context2); + auto outPath = evalState->coerceToStorePath(attr2->pos, *attr2->value, context2, ""); - store->buildPaths({ DerivedPath::Built { drvPath } }); + store->buildPaths({ + DerivedPath::Built { + .drvPath = drvPath, + .outputs = OutputsSpec::All { }, + }, + }); auto outPathS = store->printStorePath(outPath); @@ -113,7 +119,7 @@ struct CmdBundle : InstallableCommand auto * attr = vRes->attrs->get(evalState->sName); if (!attr) throw Error("attribute 'name' missing"); - outLink = evalState->forceStringNoCtx(*attr->value, attr->pos); + outLink = evalState->forceStringNoCtx(*attr->value, attr->pos, ""); } // TODO: will crash if not a localFSStore? diff --git a/src/nix/bundle.md b/src/nix/bundle.md index 2bb70711f..89458aaaa 100644 --- a/src/nix/bundle.md +++ b/src/nix/bundle.md @@ -29,7 +29,7 @@ R""( # Description -`nix bundle`, by default, packs the closure of the *installable* into a single +`nix bundle`, by default, packs the closure of the [*installable*](./nix.md#installables) into a single self-extracting executable. See the [`bundlers` homepage](https://github.com/NixOS/bundlers) for more details. @@ -44,7 +44,7 @@ flake output attributes: * `bundlers.<system>.default` -If an attribute *name* is given, `nix run` tries the following flake +If an attribute *name* is given, `nix bundle` tries the following flake output attributes: * `bundlers.<system>.<name>` diff --git a/src/nix/cat.cc b/src/nix/cat.cc index 6420a0f79..60aa66ce0 100644 --- a/src/nix/cat.cc +++ b/src/nix/cat.cc @@ -17,7 +17,7 @@ struct MixCat : virtual Args if (st.type != FSAccessor::Type::tRegular) throw Error("path '%1%' is not a regular file", path); - std::cout << accessor->readFile(path); + writeFull(STDOUT_FILENO, accessor->readFile(path)); } }; diff --git a/src/nix/copy.cc b/src/nix/copy.cc index 8730a9a5c..151d28277 100644 --- a/src/nix/copy.cc +++ b/src/nix/copy.cc @@ -10,8 +10,6 @@ struct CmdCopy : virtual CopyCommand, virtual BuiltPathsCommand SubstituteFlag substitute = NoSubstitute; - using BuiltPathsCommand::run; - CmdCopy() : BuiltPathsCommand(true) { diff --git a/src/nix/daemon.cc b/src/nix/daemon.cc index 940923d3b..7e4a7ba86 100644 --- a/src/nix/daemon.cc +++ b/src/nix/daemon.cc @@ -34,6 +34,43 @@ using namespace nix; using namespace nix::daemon; +struct AuthorizationSettings : Config { + + Setting<Strings> trustedUsers{ + this, {"root"}, "trusted-users", + R"( + A list of names of users (separated by whitespace) that have + additional rights when connecting to the Nix daemon, such as the + ability to specify additional binary caches, or to import unsigned + NARs. You can also specify groups by prefixing them with `@`; for + instance, `@wheel` means all users in the `wheel` group. The default + is `root`. + + > **Warning** + > + > Adding a user to `trusted-users` is essentially equivalent to + > giving that user root access to the system. For example, the user + > can set `sandbox-paths` and thereby obtain read access to + > directories that are otherwise inacessible to them. + )"}; + + /* ?Who we trust to use the daemon in safe ways */ + Setting<Strings> allowedUsers{ + this, {"*"}, "allowed-users", + R"( + A list of names of users (separated by whitespace) that are allowed + to connect to the Nix daemon. As with the `trusted-users` option, + you can specify groups by prefixing them with `@`. Also, you can + allow all users by specifying `*`. The default is `*`. + + Note that trusted users are always allowed to connect. + )"}; +}; + +AuthorizationSettings authorizationSettings; + +static GlobalConfig::Register rSettings(&authorizationSettings); + #ifndef __linux__ #define SPLICE_F_MOVE 0 static ssize_t splice(int fd_in, void *off_in, int fd_out, void *off_out, size_t len, unsigned int flags) @@ -203,8 +240,8 @@ static void daemonLoop() struct group * gr = peer.gidKnown ? getgrgid(peer.gid) : 0; std::string group = gr ? gr->gr_name : std::to_string(peer.gid); - Strings trustedUsers = settings.trustedUsers; - Strings allowedUsers = settings.allowedUsers; + Strings trustedUsers = authorizationSettings.trustedUsers; + Strings allowedUsers = authorizationSettings.allowedUsers; if (matchUser(user, group, trustedUsers)) trusted = Trusted; @@ -212,9 +249,9 @@ static void daemonLoop() if ((!trusted && !matchUser(user, group, allowedUsers)) || group == settings.buildUsersGroup) throw Error("user '%1%' is not allowed to connect to the Nix daemon", user); - printInfo(format((std::string) "accepted connection from pid %1%, user %2%" + (trusted ? " (trusted)" : "")) - % (peer.pidKnown ? std::to_string(peer.pid) : "<unknown>") - % (peer.uidKnown ? user : "<unknown>")); + printInfo((std::string) "accepted connection from pid %1%, user %2%" + (trusted ? " (trusted)" : ""), + peer.pidKnown ? std::to_string(peer.pid) : "<unknown>", + peer.uidKnown ? user : "<unknown>"); // Fork a child to handle the connection. ProcessOptions options; @@ -241,15 +278,7 @@ static void daemonLoop() // Handle the connection. FdSource from(remote.get()); FdSink to(remote.get()); - processConnection(openUncachedStore(), from, to, trusted, NotRecursive, [&](Store & store) { -#if 0 - /* Prevent users from doing something very dangerous. */ - if (geteuid() == 0 && - querySetting("build-users-group", "") == "") - throw Error("if you run 'nix-daemon' as root, then you MUST set 'build-users-group'!"); -#endif - store.createUser(user, peer.uid); - }); + processConnection(openUncachedStore(), from, to, trusted, NotRecursive); exit(0); }, options); @@ -257,7 +286,7 @@ static void daemonLoop() } catch (Interrupted & e) { return; } catch (Error & error) { - ErrorInfo ei = error.info(); + auto ei = error.info(); // FIXME: add to trace? ei.msg = hintfmt("error processing connection: %1%", ei.msg.str()); logError(ei); @@ -302,7 +331,7 @@ static void runDaemon(bool stdio) /* Auth hook is empty because in this mode we blindly trust the standard streams. Limiting access to those is explicitly not `nix-daemon`'s responsibility. */ - processConnection(openUncachedStore(), from, to, Trusted, NotRecursive, [&](Store & _){}); + processConnection(openUncachedStore(), from, to, Trusted, NotRecursive); } } else daemonLoop(); diff --git a/src/nix/daemon.md b/src/nix/daemon.md index e97016a94..d5cdadf08 100644 --- a/src/nix/daemon.md +++ b/src/nix/daemon.md @@ -11,7 +11,7 @@ R""( # Description This command runs the Nix daemon, which is a required component in -multi-user Nix installations. It performs build actions and other +multi-user Nix installations. It runs build tasks and other operations on the Nix store on behalf of non-root users. Usually you don't run the daemon directly; instead it's managed by a service management framework such as `systemd`. diff --git a/src/nix/describe-stores.cc b/src/nix/describe-stores.cc index 1dd384c0e..eafcedd1f 100644 --- a/src/nix/describe-stores.cc +++ b/src/nix/describe-stores.cc @@ -25,7 +25,7 @@ struct CmdDescribeStores : Command, MixJSON res[storeName] = storeConfig->toJSON(); } if (json) { - std::cout << res; + logger->cout("%s", res); } else { for (auto & [storeName, storeConfig] : res.items()) { std::cout << "## " << storeName << std::endl << std::endl; diff --git a/src/nix/develop.cc b/src/nix/develop.cc index ba7ba7c25..f06ade008 100644 --- a/src/nix/develop.cc +++ b/src/nix/develop.cc @@ -1,9 +1,10 @@ #include "eval.hh" #include "command.hh" +#include "installable-flake.hh" #include "common-args.hh" #include "shared.hh" #include "store-api.hh" -#include "path-with-outputs.hh" +#include "outputs-spec.hh" #include "derivations.hh" #include "progress-bar.hh" #include "run.hh" @@ -164,6 +165,14 @@ struct BuildEnvironment { return vars == other.vars && bashFunctions == other.bashFunctions; } + + std::string getSystem() const + { + if (auto v = get(vars, "system")) + return getString(*v); + else + return settings.thisSystem; + } }; const static std::string getEnvSh = @@ -192,10 +201,12 @@ static StorePath getDerivationEnvironment(ref<Store> store, ref<Store> evalStore drv.env.erase("allowedRequisites"); drv.env.erase("disallowedReferences"); drv.env.erase("disallowedRequisites"); + drv.env.erase("name"); /* Rehash and write the derivation. FIXME: would be nice to use 'buildDerivation', but that's privileged. */ drv.name += "-env"; + drv.env.emplace("name", drv.name); drv.inputSrcs.insert(std::move(getEnvShPath)); if (settings.isExperimentalFeatureEnabled(Xp::CaDerivations)) { for (auto & output : drv.outputs) { @@ -222,7 +233,12 @@ static StorePath getDerivationEnvironment(ref<Store> store, ref<Store> evalStore auto shellDrvPath = writeDerivation(*evalStore, drv); /* Build the derivation. */ - store->buildPaths({DerivedPath::Built{shellDrvPath}}, bmNormal, evalStore); + store->buildPaths( + { DerivedPath::Built { + .drvPath = shellDrvPath, + .outputs = OutputsSpec::All { }, + }}, + bmNormal, evalStore); for (auto & [_0, optPath] : evalStore->queryPartialDerivationOutputMap(shellDrvPath)) { assert(optPath); @@ -246,6 +262,7 @@ struct Common : InstallableCommand, MixProfile "NIX_LOG_FD", "NIX_REMOTE", "PPID", + "SHELL", "SHELLOPTS", "SSL_CERT_FILE", // FIXME: only want to ignore /no-cert-file.crt "TEMP", @@ -296,7 +313,7 @@ struct Common : InstallableCommand, MixProfile buildEnvironment.toBash(out, ignoreVars); for (auto & var : savedVars) - out << fmt("%s=\"$%s:$nix_saved_%s\"\n", var, var, var); + out << fmt("%s=\"$%s${nix_saved_%s:+:$nix_saved_%s}\"\n", var, var, var, var); out << "export NIX_BUILD_TOP=\"$(mktemp -d -t nix-shell.XXXXXX)\"\n"; for (auto & i : {"TMP", "TMPDIR", "TEMP", "TEMPDIR"}) @@ -357,7 +374,7 @@ struct Common : InstallableCommand, MixProfile return res; } - StorePath getShellOutPath(ref<Store> store) + StorePath getShellOutPath(ref<Store> store, ref<Installable> installable) { auto path = installable->getStorePath(); if (path && hasSuffix(path->to_string(), "-env")) @@ -375,9 +392,10 @@ struct Common : InstallableCommand, MixProfile } } - std::pair<BuildEnvironment, std::string> getBuildEnvironment(ref<Store> store) + std::pair<BuildEnvironment, std::string> + getBuildEnvironment(ref<Store> store, ref<Installable> installable) { - auto shellOutPath = getShellOutPath(store); + auto shellOutPath = getShellOutPath(store, installable); auto strPath = store->printStorePath(shellOutPath); @@ -463,9 +481,9 @@ struct CmdDevelop : Common, MixEnvironment ; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { - auto [buildEnvironment, gcroot] = getBuildEnvironment(store); + auto [buildEnvironment, gcroot] = getBuildEnvironment(store, installable); auto [rcFileFd, rcFilePath] = createTempFile("nix-shell"); @@ -520,7 +538,7 @@ struct CmdDevelop : Common, MixEnvironment nixpkgsLockFlags.inputOverrides = {}; nixpkgsLockFlags.inputUpdates = {}; - auto bashInstallable = std::make_shared<InstallableFlake>( + auto bashInstallable = make_ref<InstallableFlake>( this, state, installable->nixpkgsFlakeRef(), @@ -556,7 +574,7 @@ struct CmdDevelop : Common, MixEnvironment // Need to chdir since phases assume in flake directory if (phase) { // chdir if installable is a flake of type git+file or path - auto installableFlake = std::dynamic_pointer_cast<InstallableFlake>(installable); + auto installableFlake = installable.dynamic_pointer_cast<InstallableFlake>(); if (installableFlake) { auto sourcePath = installableFlake->getLockedFlake()->flake.resolvedRef.input.getSourcePath(); if (sourcePath) { @@ -567,7 +585,7 @@ struct CmdDevelop : Common, MixEnvironment } } - runProgramInStore(store, shell, args); + runProgramInStore(store, shell, args, buildEnvironment.getSystem()); } }; @@ -587,9 +605,9 @@ struct CmdPrintDevEnv : Common, MixJSON Category category() override { return catUtility; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { - auto buildEnvironment = getBuildEnvironment(store).first; + auto buildEnvironment = getBuildEnvironment(store, installable).first; stopProgressBar(); diff --git a/src/nix/develop.md b/src/nix/develop.md index e036ec6b9..c49b39669 100644 --- a/src/nix/develop.md +++ b/src/nix/develop.md @@ -66,11 +66,17 @@ R""( `nixpkgs#glibc` in `~/my-glibc` and want to compile another package against it. +* Run a series of script commands: + + ```console + # nix develop --command bash -c "mkdir build && cmake .. && make" + ``` + # Description `nix develop` starts a `bash` shell that provides an interactive build environment nearly identical to what Nix would use to build -*installable*. Inside this shell, environment variables and shell +[*installable*](./nix.md#installables). Inside this shell, environment variables and shell functions are set up so that you can interactively and incrementally build your package. diff --git a/src/nix/diff-closures.cc b/src/nix/diff-closures.cc index 0621d662c..c7c37b66f 100644 --- a/src/nix/diff-closures.cc +++ b/src/nix/diff-closures.cc @@ -97,7 +97,7 @@ void printClosureDiff( items.push_back(fmt("%s → %s", showVersions(removed), showVersions(added))); if (showDelta) items.push_back(fmt("%s%+.1f KiB" ANSI_NORMAL, sizeDelta > 0 ? ANSI_RED : ANSI_GREEN, sizeDelta / 1024.0)); - std::cout << fmt("%s%s: %s\n", indent, name, concatStringsSep(", ", items)); + logger->cout("%s%s: %s", indent, name, concatStringsSep(", ", items)); } } } @@ -106,7 +106,7 @@ void printClosureDiff( using namespace nix; -struct CmdDiffClosures : SourceExprCommand +struct CmdDiffClosures : SourceExprCommand, MixOperateOnOptions { std::string _before, _after; diff --git a/src/nix/doctor.cc b/src/nix/doctor.cc index ea87e3d87..7da4549a1 100644 --- a/src/nix/doctor.cc +++ b/src/nix/doctor.cc @@ -18,7 +18,7 @@ std::string formatProtocol(unsigned int proto) if (proto) { auto major = GET_PROTOCOL_MAJOR(proto) >> 8; auto minor = GET_PROTOCOL_MINOR(proto); - return (format("%1%.%2%") % major % minor).str(); + return fmt("%1%.%2%", major, minor); } return "unknown"; } diff --git a/src/nix/edit.cc b/src/nix/edit.cc index 76a134b1f..c46c1c23c 100644 --- a/src/nix/edit.cc +++ b/src/nix/edit.cc @@ -3,6 +3,7 @@ #include "eval.hh" #include "attr-path.hh" #include "progress-bar.hh" +#include "editor-for.hh" #include <unistd.h> @@ -24,7 +25,7 @@ struct CmdEdit : InstallableCommand Category category() override { return catSecondary; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { auto state = getEvalState(); diff --git a/src/nix/eval.cc b/src/nix/eval.cc index 967dc8519..6c2b60427 100644 --- a/src/nix/eval.cc +++ b/src/nix/eval.cc @@ -4,19 +4,20 @@ #include "store-api.hh" #include "eval.hh" #include "eval-inline.hh" -#include "json.hh" #include "value-to-json.hh" #include "progress-bar.hh" +#include <nlohmann/json.hpp> + using namespace nix; -struct CmdEval : MixJSON, InstallableCommand +struct CmdEval : MixJSON, InstallableCommand, MixReadOnlyOption { bool raw = false; std::optional<std::string> apply; std::optional<Path> writeTo; - CmdEval() : InstallableCommand(true /* supportReadOnlyMode */) + CmdEval() : InstallableCommand() { addFlag({ .longName = "raw", @@ -53,7 +54,7 @@ struct CmdEval : MixJSON, InstallableCommand Category category() override { return catSecondary; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { if (raw && json) throw UsageError("--raw and --json are mutually exclusive"); @@ -111,12 +112,11 @@ struct CmdEval : MixJSON, InstallableCommand else if (raw) { stopProgressBar(); - std::cout << *state->coerceToString(noPos, *v, context); + writeFull(STDOUT_FILENO, *state->coerceToString(noPos, *v, context, "while generating the eval command output")); } else if (json) { - JSONPlaceholder jsonOut(std::cout); - printValueAsJSON(*state, true, *v, pos, jsonOut, context); + logger->cout("%s", printValueAsJSON(*state, true, *v, pos, context, false)); } else { diff --git a/src/nix/eval.md b/src/nix/eval.md index 61334cde1..3b510737a 100644 --- a/src/nix/eval.md +++ b/src/nix/eval.md @@ -50,7 +50,7 @@ R""( # Description -This command evaluates the Nix expression *installable* and prints the +This command evaluates the given Nix expression and prints the result on standard output. # Output format diff --git a/src/nix/flake-update.md b/src/nix/flake-update.md index 03b50e38e..8c6042d94 100644 --- a/src/nix/flake-update.md +++ b/src/nix/flake-update.md @@ -6,7 +6,7 @@ R""( lock file: ```console - # nix flake update + # nix flake update --commit-lock-file * Updated 'nix': 'github:NixOS/nix/9fab14adbc3810d5cc1f88672fde1eee4358405c' -> 'github:NixOS/nix/8927cba62f5afb33b01016d5c4f7f8b7d0adde3c' * Updated 'nixpkgs': 'github:NixOS/nixpkgs/3d2d8f281a27d466fa54b469b5993f7dde198375' -> 'github:NixOS/nixpkgs/a3a3dda3bacf61e8a39258a0ed9c924eeca8e293' … @@ -16,7 +16,7 @@ R""( # Description This command recreates the lock file of a flake (`flake.lock`), thus -updating the lock for every mutable input (like `nixpkgs`) to its +updating the lock for every unlocked input (like `nixpkgs`) to its current version. This is equivalent to passing `--recreate-lock-file` to any command that operates on a flake. That is, diff --git a/src/nix/flake.cc b/src/nix/flake.cc index e01bc6d10..0a6616396 100644 --- a/src/nix/flake.cc +++ b/src/nix/flake.cc @@ -1,4 +1,5 @@ #include "command.hh" +#include "installable-flake.hh" #include "common-args.hh" #include "shared.hh" #include "eval.hh" @@ -7,11 +8,10 @@ #include "get-drvs.hh" #include "store-api.hh" #include "derivations.hh" -#include "path-with-outputs.hh" +#include "outputs-spec.hh" #include "attr-path.hh" #include "fetchers.hh" #include "registry.hh" -#include "json.hh" #include "eval-cache.hh" #include "markdown.hh" @@ -21,6 +21,7 @@ using namespace nix; using namespace nix::flake; +using json = nlohmann::json; class FlakeCommand : virtual Args, public MixFlakeOptions { @@ -126,12 +127,12 @@ static void enumerateOutputs(EvalState & state, Value & vFlake, std::function<void(const std::string & name, Value & vProvide, const PosIdx pos)> callback) { auto pos = vFlake.determinePos(noPos); - state.forceAttrs(vFlake, pos); + state.forceAttrs(vFlake, pos, "while evaluating a flake to get its outputs"); auto aOutputs = vFlake.attrs->get(state.symbols.create("outputs")); assert(aOutputs); - state.forceAttrs(*aOutputs->value, pos); + state.forceAttrs(*aOutputs->value, pos, "while evaluating the outputs of a flake"); auto sHydraJobs = state.symbols.create("hydraJobs"); @@ -212,9 +213,10 @@ struct CmdFlakeMetadata : FlakeCommand, MixJSON ANSI_BOLD "Last modified:" ANSI_NORMAL " %s", std::put_time(std::localtime(&*lastModified), "%F %T")); - logger->cout(ANSI_BOLD "Inputs:" ANSI_NORMAL); + if (!lockedFlake.lockFile.root->inputs.empty()) + logger->cout(ANSI_BOLD "Inputs:" ANSI_NORMAL); - std::unordered_set<std::shared_ptr<Node>> visited; + std::set<ref<Node>> visited; std::function<void(const Node & node, const std::string & prefix)> recurse; @@ -226,7 +228,7 @@ struct CmdFlakeMetadata : FlakeCommand, MixJSON if (auto lockedNode = std::get_if<0>(&input.second)) { logger->cout("%s" ANSI_BOLD "%s" ANSI_NORMAL ": %s", prefix + (last ? treeLast : treeConn), input.first, - *lockedNode ? (*lockedNode)->lockedRef : flake.lockedRef); + (*lockedNode)->lockedRef); bool firstVisit = visited.insert(*lockedNode).second; @@ -347,7 +349,7 @@ struct CmdFlakeCheck : FlakeCommand // FIXME auto app = App(*state, v); for (auto & i : app.context) { - auto [drvPathS, outputName] = decodeContext(i); + auto [drvPathS, outputName] = NixStringContextElem::parse(i); store->parseStorePath(drvPathS); } #endif @@ -380,23 +382,6 @@ struct CmdFlakeCheck : FlakeCommand auto checkModule = [&](const std::string & attrPath, Value & v, const PosIdx pos) { try { state->forceValue(v, pos); - if (v.isLambda()) { - if (!v.lambda.fun->hasFormals() || !v.lambda.fun->formals->ellipsis) - throw Error("module must match an open attribute set ('{ config, ... }')"); - } else if (v.type() == nAttrs) { - for (auto & attr : *v.attrs) - try { - state->forceValue(*attr.value, attr.pos); - } catch (Error & e) { - e.addTrace( - state->positions[attr.pos], - hintfmt("while evaluating the option '%s'", state->symbols[attr.name])); - throw; - } - } else - throw Error("module must be a function or an attribute set"); - // FIXME: if we have a 'nixpkgs' input, use it to - // check the module. } catch (Error & e) { e.addTrace(resolve(pos), hintfmt("while checking the NixOS module '%s'", attrPath)); reportError(e); @@ -407,13 +392,13 @@ struct CmdFlakeCheck : FlakeCommand checkHydraJobs = [&](const std::string & attrPath, Value & v, const PosIdx pos) { try { - state->forceAttrs(v, pos); + state->forceAttrs(v, pos, ""); if (state->isDerivation(v)) throw Error("jobset should not be a derivation at top-level"); for (auto & attr : *v.attrs) { - state->forceAttrs(*attr.value, attr.pos); + state->forceAttrs(*attr.value, attr.pos, ""); auto attrPath2 = concatStrings(attrPath, ".", state->symbols[attr.name]); if (state->isDerivation(*attr.value)) { Activity act(*logger, lvlChatty, actUnknown, @@ -435,7 +420,7 @@ struct CmdFlakeCheck : FlakeCommand fmt("checking NixOS configuration '%s'", attrPath)); Bindings & bindings(*state->allocBindings(0)); auto vToplevel = findAlongAttrPath(*state, "config.system.build.toplevel", bindings, v).first; - state->forceAttrs(*vToplevel, pos); + state->forceValue(*vToplevel, pos); if (!state->isDerivation(*vToplevel)) throw Error("attribute 'config.system.build.toplevel' is not a derivation"); } catch (Error & e) { @@ -449,12 +434,12 @@ struct CmdFlakeCheck : FlakeCommand Activity act(*logger, lvlChatty, actUnknown, fmt("checking template '%s'", attrPath)); - state->forceAttrs(v, pos); + state->forceAttrs(v, pos, ""); if (auto attr = v.attrs->get(state->symbols.create("path"))) { if (attr->name == state->symbols.create("path")) { PathSet context; - auto path = state->coerceToPath(attr->pos, *attr->value, context); + auto path = state->coerceToPath(attr->pos, *attr->value, context, ""); if (!store->isInStore(path)) throw Error("template '%s' has a bad 'path' attribute"); // TODO: recursively check the flake in 'path'. @@ -463,7 +448,7 @@ struct CmdFlakeCheck : FlakeCommand throw Error("template '%s' lacks attribute 'path'", attrPath); if (auto attr = v.attrs->get(state->symbols.create("description"))) - state->forceStringNoCtx(*attr->value, attr->pos); + state->forceStringNoCtx(*attr->value, attr->pos, ""); else throw Error("template '%s' lacks attribute 'description'", attrPath); @@ -520,23 +505,27 @@ struct CmdFlakeCheck : FlakeCommand warn("flake output attribute '%s' is deprecated; use '%s' instead", name, replacement); if (name == "checks") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); - state->forceAttrs(*attr.value, attr.pos); + state->forceAttrs(*attr.value, attr.pos, ""); for (auto & attr2 : *attr.value->attrs) { auto drvPath = checkDerivation( fmt("%s.%s.%s", name, attr_name, state->symbols[attr2.name]), *attr2.value, attr2.pos); - if (drvPath && attr_name == settings.thisSystem.get()) - drvPaths.push_back(DerivedPath::Built{*drvPath}); + if (drvPath && attr_name == settings.thisSystem.get()) { + drvPaths.push_back(DerivedPath::Built { + .drvPath = *drvPath, + .outputs = OutputsSpec::All { }, + }); + } } } } else if (name == "formatter") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); @@ -547,11 +536,11 @@ struct CmdFlakeCheck : FlakeCommand } else if (name == "packages" || name == "devShells") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); - state->forceAttrs(*attr.value, attr.pos); + state->forceAttrs(*attr.value, attr.pos, ""); for (auto & attr2 : *attr.value->attrs) checkDerivation( fmt("%s.%s.%s", name, attr_name, state->symbols[attr2.name]), @@ -560,11 +549,11 @@ struct CmdFlakeCheck : FlakeCommand } else if (name == "apps") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); - state->forceAttrs(*attr.value, attr.pos); + state->forceAttrs(*attr.value, attr.pos, ""); for (auto & attr2 : *attr.value->attrs) checkApp( fmt("%s.%s.%s", name, attr_name, state->symbols[attr2.name]), @@ -573,7 +562,7 @@ struct CmdFlakeCheck : FlakeCommand } else if (name == "defaultPackage" || name == "devShell") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); @@ -584,7 +573,7 @@ struct CmdFlakeCheck : FlakeCommand } else if (name == "defaultApp") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); @@ -595,7 +584,7 @@ struct CmdFlakeCheck : FlakeCommand } else if (name == "legacyPackages") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { checkSystemName(state->symbols[attr.name], attr.pos); // FIXME: do getDerivations? @@ -606,7 +595,7 @@ struct CmdFlakeCheck : FlakeCommand checkOverlay(name, vOutput, pos); else if (name == "overlays") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) checkOverlay(fmt("%s.%s", name, state->symbols[attr.name]), *attr.value, attr.pos); @@ -616,14 +605,14 @@ struct CmdFlakeCheck : FlakeCommand checkModule(name, vOutput, pos); else if (name == "nixosModules") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) checkModule(fmt("%s.%s", name, state->symbols[attr.name]), *attr.value, attr.pos); } else if (name == "nixosConfigurations") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) checkNixOSConfiguration(fmt("%s.%s", name, state->symbols[attr.name]), *attr.value, attr.pos); @@ -636,14 +625,14 @@ struct CmdFlakeCheck : FlakeCommand checkTemplate(name, vOutput, pos); else if (name == "templates") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) checkTemplate(fmt("%s.%s", name, state->symbols[attr.name]), *attr.value, attr.pos); } else if (name == "defaultBundler") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); @@ -654,11 +643,11 @@ struct CmdFlakeCheck : FlakeCommand } else if (name == "bundlers") { - state->forceAttrs(vOutput, pos); + state->forceAttrs(vOutput, pos, ""); for (auto & attr : *vOutput.attrs) { const auto & attr_name = state->symbols[attr.name]; checkSystemName(attr_name, attr.pos); - state->forceAttrs(*attr.value, attr.pos); + state->forceAttrs(*attr.value, attr.pos, ""); for (auto & attr2 : *attr.value->attrs) { checkBundler( fmt("%s.%s.%s", name, attr_name, state->symbols[attr2.name]), @@ -667,6 +656,19 @@ struct CmdFlakeCheck : FlakeCommand } } + else if ( + name == "lib" + || name == "darwinConfigurations" + || name == "darwinModules" + || name == "flakeModule" + || name == "flakeModules" + || name == "herculesCI" + || name == "homeConfigurations" + || name == "nixopsConfigurations" + ) + // Known but unchecked community attribute + ; + else warn("unknown flake output '%s'", name); @@ -916,35 +918,44 @@ struct CmdFlakeArchive : FlakeCommand, MixJSON, MixDryRun { auto flake = lockFlake(); - auto jsonRoot = json ? std::optional<JSONObject>(std::cout) : std::nullopt; - StorePathSet sources; sources.insert(flake.flake.sourceInfo->storePath); - if (jsonRoot) - jsonRoot->attr("path", store->printStorePath(flake.flake.sourceInfo->storePath)); // FIXME: use graph output, handle cycles. - std::function<void(const Node & node, std::optional<JSONObject> & jsonObj)> traverse; - traverse = [&](const Node & node, std::optional<JSONObject> & jsonObj) + std::function<nlohmann::json(const Node & node)> traverse; + traverse = [&](const Node & node) { - auto jsonObj2 = jsonObj ? jsonObj->object("inputs") : std::optional<JSONObject>(); + nlohmann::json jsonObj2 = json ? json::object() : nlohmann::json(nullptr); for (auto & [inputName, input] : node.inputs) { if (auto inputNode = std::get_if<0>(&input)) { - auto jsonObj3 = jsonObj2 ? jsonObj2->object(inputName) : std::optional<JSONObject>(); auto storePath = dryRun ? (*inputNode)->lockedRef.input.computeStorePath(*store) : (*inputNode)->lockedRef.input.fetch(store).first.storePath; - if (jsonObj3) - jsonObj3->attr("path", store->printStorePath(storePath)); - sources.insert(std::move(storePath)); - traverse(**inputNode, jsonObj3); + if (json) { + auto& jsonObj3 = jsonObj2[inputName]; + jsonObj3["path"] = store->printStorePath(storePath); + sources.insert(std::move(storePath)); + jsonObj3["inputs"] = traverse(**inputNode); + } else { + sources.insert(std::move(storePath)); + traverse(**inputNode); + } } } + return jsonObj2; }; - traverse(*flake.lockFile.root, jsonRoot); + if (json) { + nlohmann::json jsonRoot = { + {"path", store->printStorePath(flake.flake.sourceInfo->storePath)}, + {"inputs", traverse(*flake.lockFile.root)}, + }; + logger->cout("%s", jsonRoot); + } else { + traverse(*flake.lockFile.root); + } if (!dryRun && !dstUri.empty()) { ref<Store> dstStore = dstUri.empty() ? openStore() : openStore(dstUri); @@ -956,6 +967,7 @@ struct CmdFlakeArchive : FlakeCommand, MixJSON, MixDryRun struct CmdFlakeShow : FlakeCommand, MixJSON { bool showLegacy = false; + bool showAllSystems = false; CmdFlakeShow() { @@ -964,6 +976,11 @@ struct CmdFlakeShow : FlakeCommand, MixJSON .description = "Show the contents of the `legacyPackages` output.", .handler = {&showLegacy, true} }); + addFlag({ + .longName = "all-systems", + .description = "Show the contents of outputs for all systems.", + .handler = {&showAllSystems, true} + }); } std::string description() override @@ -984,6 +1001,62 @@ struct CmdFlakeShow : FlakeCommand, MixJSON auto state = getEvalState(); auto flake = std::make_shared<LockedFlake>(lockFlake()); + auto localSystem = std::string(settings.thisSystem.get()); + + std::function<bool( + eval_cache::AttrCursor & visitor, + const std::vector<Symbol> &attrPath, + const Symbol &attr)> hasContent; + + // For frameworks it's important that structures are as lazy as possible + // to prevent infinite recursions, performance issues and errors that + // aren't related to the thing to evaluate. As a consequence, they have + // to emit more attributes than strictly (sic) necessary. + // However, these attributes with empty values are not useful to the user + // so we omit them. + hasContent = [&]( + eval_cache::AttrCursor & visitor, + const std::vector<Symbol> &attrPath, + const Symbol &attr) -> bool + { + auto attrPath2(attrPath); + attrPath2.push_back(attr); + auto attrPathS = state->symbols.resolve(attrPath2); + const auto & attrName = state->symbols[attr]; + + auto visitor2 = visitor.getAttr(attrName); + + if ((attrPathS[0] == "apps" + || attrPathS[0] == "checks" + || attrPathS[0] == "devShells" + || attrPathS[0] == "legacyPackages" + || attrPathS[0] == "packages") + && (attrPathS.size() == 1 || attrPathS.size() == 2)) { + for (const auto &subAttr : visitor2->getAttrs()) { + if (hasContent(*visitor2, attrPath2, subAttr)) { + return true; + } + } + return false; + } + + if ((attrPathS.size() == 1) + && (attrPathS[0] == "formatter" + || attrPathS[0] == "nixosConfigurations" + || attrPathS[0] == "nixosModules" + || attrPathS[0] == "overlays" + )) { + for (const auto &subAttr : visitor2->getAttrs()) { + if (hasContent(*visitor2, attrPath2, subAttr)) { + return true; + } + } + return false; + } + + // If we don't recognize it, it's probably content + return true; + }; std::function<nlohmann::json( eval_cache::AttrCursor & visitor, @@ -1010,7 +1083,12 @@ struct CmdFlakeShow : FlakeCommand, MixJSON { if (!json) logger->cout("%s", headerPrefix); - auto attrs = visitor.getAttrs(); + std::vector<Symbol> attrs; + for (const auto &attr : visitor.getAttrs()) { + if (hasContent(visitor, attrPath, attr)) + attrs.push_back(attr); + } + for (const auto & [i, attr] : enumerate(attrs)) { const auto & attrName = state->symbols[attr]; bool last = i + 1 == attrs.size(); @@ -1074,10 +1152,18 @@ struct CmdFlakeShow : FlakeCommand, MixJSON || (attrPath.size() == 3 && (attrPathS[0] == "checks" || attrPathS[0] == "packages" || attrPathS[0] == "devShells")) ) { - if (visitor.isDerivation()) - showDerivation(); - else - throw Error("expected a derivation"); + if (!showAllSystems && std::string(attrPathS[1]) != localSystem) { + if (!json) + logger->cout(fmt("%s " ANSI_WARNING "omitted" ANSI_NORMAL " (use '--all-systems' to show)", headerPrefix)); + else { + logger->warn(fmt("%s omitted (use '--all-systems' to show)", concatStringsSep(".", attrPathS))); + } + } else { + if (visitor.isDerivation()) + showDerivation(); + else + throw Error("expected a derivation"); + } } else if (attrPath.size() > 0 && attrPathS[0] == "hydraJobs") { @@ -1096,6 +1182,12 @@ struct CmdFlakeShow : FlakeCommand, MixJSON else { logger->warn(fmt("%s omitted (use '--legacy' to show)", concatStringsSep(".", attrPathS))); } + } else if (!showAllSystems && std::string(attrPathS[1]) != localSystem) { + if (!json) + logger->cout(fmt("%s " ANSI_WARNING "omitted" ANSI_NORMAL " (use '--all-systems' to show)", headerPrefix)); + else { + logger->warn(fmt("%s omitted (use '--all-systems' to show)", concatStringsSep(".", attrPathS))); + } } else { if (visitor.isDerivation()) showDerivation(); @@ -1237,7 +1329,6 @@ struct CmdFlake : NixMultiCommand if (!command) throw UsageError("'nix flake' requires a sub-command."); settings.requireExperimentalFeature(Xp::Flakes); - command->second->prepare(); command->second->run(); } }; diff --git a/src/nix/flake.md b/src/nix/flake.md index a1ab43281..8eaa41b96 100644 --- a/src/nix/flake.md +++ b/src/nix/flake.md @@ -18,51 +18,56 @@ values such as packages or NixOS modules provided by the flake). Flake references (*flakerefs*) are a way to specify the location of a flake. These have two different forms: -* An attribute set representation, e.g. - ```nix - { - type = "github"; - owner = "NixOS"; - repo = "nixpkgs"; - } - ``` +## Attribute set representation - The only required attribute is `type`. The supported types are - listed below. +Example: -* A URL-like syntax, e.g. +```nix +{ + type = "github"; + owner = "NixOS"; + repo = "nixpkgs"; +} +``` - ``` - github:NixOS/nixpkgs - ``` +The only required attribute is `type`. The supported types are +listed below. - These are used on the command line as a more convenient alternative - to the attribute set representation. For instance, in the command +## URL-like syntax - ```console - # nix build github:NixOS/nixpkgs#hello - ``` +Example: - `github:NixOS/nixpkgs` is a flake reference (while `hello` is an - output attribute). They are also allowed in the `inputs` attribute - of a flake, e.g. +``` +github:NixOS/nixpkgs +``` - ```nix - inputs.nixpkgs.url = github:NixOS/nixpkgs; - ``` +These are used on the command line as a more convenient alternative +to the attribute set representation. For instance, in the command - is equivalent to +```console +# nix build github:NixOS/nixpkgs#hello +``` - ```nix - inputs.nixpkgs = { - type = "github"; - owner = "NixOS"; - repo = "nixpkgs"; - }; - ``` +`github:NixOS/nixpkgs` is a flake reference (while `hello` is an +output attribute). They are also allowed in the `inputs` attribute +of a flake, e.g. + +```nix +inputs.nixpkgs.url = "github:NixOS/nixpkgs"; +``` + +is equivalent to + +```nix +inputs.nixpkgs = { + type = "github"; + owner = "NixOS"; + repo = "nixpkgs"; +}; +``` -## Examples +### Examples Here are some examples of flake references in their URL-like representation: @@ -270,14 +275,14 @@ Currently the `type` attribute can be one of the following: # Flake format As an example, here is a simple `flake.nix` that depends on the -Nixpkgs flake and provides a single package (i.e. an installable -derivation): +Nixpkgs flake and provides a single package (i.e. an +[installable](./nix.md#installables) derivation): ```nix { description = "A flake for building Hello World"; - inputs.nixpkgs.url = github:NixOS/nixpkgs/nixos-20.03; + inputs.nixpkgs.url = "github:NixOS/nixpkgs/nixos-20.03"; outputs = { self, nixpkgs }: { @@ -369,7 +374,7 @@ inputs.nixpkgs = { Alternatively, you can use the URL-like syntax: ```nix -inputs.import-cargo.url = github:edolstra/import-cargo; +inputs.import-cargo.url = "github:edolstra/import-cargo"; inputs.nixpkgs.url = "nixpkgs"; ``` diff --git a/src/nix/get-env.sh b/src/nix/get-env.sh index 42c806450..a7a8a01b9 100644 --- a/src/nix/get-env.sh +++ b/src/nix/get-env.sh @@ -43,6 +43,7 @@ __dumpEnv() { local __var_name="${BASH_REMATCH[2]}" if [[ $__var_name =~ ^BASH_ || \ + $__var_name =~ ^COMP_ || \ $__var_name = _ || \ $__var_name = DIRSTACK || \ $__var_name = EUID || \ @@ -54,7 +55,9 @@ __dumpEnv() { $__var_name = PWD || \ $__var_name = RANDOM || \ $__var_name = SHLVL || \ - $__var_name = SECONDS \ + $__var_name = SECONDS || \ + $__var_name = EPOCHREALTIME || \ + $__var_name = EPOCHSECONDS \ ]]; then continue; fi if [[ -z $__first ]]; then printf ',\n'; else __first=; fi diff --git a/src/nix/hash.cc b/src/nix/hash.cc index 60d9593a7..9feca9345 100644 --- a/src/nix/hash.cc +++ b/src/nix/hash.cc @@ -151,7 +151,6 @@ struct CmdHash : NixMultiCommand { if (!command) throw UsageError("'nix hash' requires a sub-command."); - command->second->prepare(); command->second->run(); } }; @@ -161,11 +160,11 @@ static auto rCmdHash = registerCommand<CmdHash>("hash"); /* Legacy nix-hash command. */ static int compatNixHash(int argc, char * * argv) { - HashType ht = htMD5; + std::optional<HashType> ht; bool flat = false; - bool base32 = false; + Base base = Base16; bool truncate = false; - enum { opHash, opTo32, opTo16 } op = opHash; + enum { opHash, opTo } op = opHash; std::vector<std::string> ss; parseCmdLine(argc, argv, [&](Strings::iterator & arg, const Strings::iterator & end) { @@ -174,14 +173,31 @@ static int compatNixHash(int argc, char * * argv) else if (*arg == "--version") printVersion("nix-hash"); else if (*arg == "--flat") flat = true; - else if (*arg == "--base32") base32 = true; + else if (*arg == "--base16") base = Base16; + else if (*arg == "--base32") base = Base32; + else if (*arg == "--base64") base = Base64; + else if (*arg == "--sri") base = SRI; else if (*arg == "--truncate") truncate = true; else if (*arg == "--type") { std::string s = getArg(*arg, arg, end); ht = parseHashType(s); } - else if (*arg == "--to-base16") op = opTo16; - else if (*arg == "--to-base32") op = opTo32; + else if (*arg == "--to-base16") { + op = opTo; + base = Base16; + } + else if (*arg == "--to-base32") { + op = opTo; + base = Base32; + } + else if (*arg == "--to-base64") { + op = opTo; + base = Base64; + } + else if (*arg == "--to-sri") { + op = opTo; + base = SRI; + } else if (*arg != "" && arg->at(0) == '-') return false; else @@ -191,17 +207,18 @@ static int compatNixHash(int argc, char * * argv) if (op == opHash) { CmdHashBase cmd(flat ? FileIngestionMethod::Flat : FileIngestionMethod::Recursive); - cmd.ht = ht; - cmd.base = base32 ? Base32 : Base16; + if (!ht.has_value()) ht = htMD5; + cmd.ht = ht.value(); + cmd.base = base; cmd.truncate = truncate; cmd.paths = ss; cmd.run(); } else { - CmdToBase cmd(op == opTo32 ? Base32 : Base16); + CmdToBase cmd(base); cmd.args = ss; - cmd.ht = ht; + if (ht.has_value()) cmd.ht = ht; cmd.run(); } diff --git a/src/nix/local.mk b/src/nix/local.mk index e4ec7634d..0f2f016ec 100644 --- a/src/nix/local.mk +++ b/src/nix/local.mk @@ -18,7 +18,7 @@ nix_CXXFLAGS += -I src/libutil -I src/libstore -I src/libfetchers -I src/libexpr nix_LIBS = libexpr libmain libfetchers libstore libutil libcmd -nix_LDFLAGS = -pthread $(SODIUM_LIBS) $(EDITLINE_LIBS) $(BOOST_LDFLAGS) -llowdown +nix_LDFLAGS = -pthread $(SODIUM_LIBS) $(EDITLINE_LIBS) $(BOOST_LDFLAGS) $(LOWDOWN_LIBS) $(foreach name, \ nix-build nix-channel nix-collect-garbage nix-copy-closure nix-daemon nix-env nix-hash nix-instantiate nix-prefetch-url nix-shell nix-store, \ diff --git a/src/nix/log.cc b/src/nix/log.cc index 72d02ef11..aaf829764 100644 --- a/src/nix/log.cc +++ b/src/nix/log.cc @@ -23,7 +23,7 @@ struct CmdLog : InstallableCommand Category category() override { return catSecondary; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { settings.readOnlyMode = true; @@ -49,11 +49,11 @@ struct CmdLog : InstallableCommand [&](const DerivedPath::Built & bfd) { return logSub.getBuildLog(bfd.drvPath); }, - }, b.raw()); + }, b.path.raw()); if (!log) continue; stopProgressBar(); printInfo("got build log for '%s' from '%s'", installable->what(), logSub.getUri()); - std::cout << *log; + writeFull(STDOUT_FILENO, *log); return; } diff --git a/src/nix/log.md b/src/nix/log.md index 1c76226a3..01e9801df 100644 --- a/src/nix/log.md +++ b/src/nix/log.md @@ -22,8 +22,7 @@ R""( # Description -This command prints the log of a previous build of the derivation -*installable* on standard output. +This command prints the log of a previous build of the [*installable*](./nix.md#installables) on standard output. Nix looks for build logs in two places: diff --git a/src/nix/ls.cc b/src/nix/ls.cc index 07554994b..c990a303c 100644 --- a/src/nix/ls.cc +++ b/src/nix/ls.cc @@ -3,7 +3,7 @@ #include "fs-accessor.hh" #include "nar-accessor.hh" #include "common-args.hh" -#include "json.hh" +#include <nlohmann/json.hpp> using namespace nix; @@ -91,10 +91,9 @@ struct MixLs : virtual Args, MixJSON if (path == "/") path = ""; if (json) { - JSONPlaceholder jsonRoot(std::cout); if (showDirectory) throw UsageError("'--directory' is useless with '--json'"); - listNar(jsonRoot, accessor, path, recursive); + logger->cout("%s", listNar(accessor, path, recursive)); } else listText(accessor); } diff --git a/src/nix/main.cc b/src/nix/main.cc index a8404a2ea..7b715f281 100644 --- a/src/nix/main.cc +++ b/src/nix/main.cc @@ -53,7 +53,6 @@ static bool haveInternet() } std::string programPath; -char * * savedArgv; struct HelpRequested { }; @@ -74,6 +73,7 @@ struct NixArgs : virtual MultiCommand, virtual MixCommonArgs addFlag({ .longName = "help", .description = "Show usage information.", + .category = miscCategory, .handler = {[&]() { throw HelpRequested(); }}, }); @@ -82,12 +82,13 @@ struct NixArgs : virtual MultiCommand, virtual MixCommonArgs .shortName = 'L', .description = "Print full build logs on standard error.", .category = loggingCategory, - .handler = {[&]() {setLogFormat(LogFormat::barWithLogs); }}, + .handler = {[&]() { logger->setPrintBuildLogs(true); }}, }); addFlag({ .longName = "version", .description = "Show version information.", + .category = miscCategory, .handler = {[&]() { showVersion = true; }}, }); @@ -95,12 +96,14 @@ struct NixArgs : virtual MultiCommand, virtual MixCommonArgs .longName = "offline", .aliases = {"no-net"}, // FIXME: remove .description = "Disable substituters and consider all previously downloaded files up-to-date.", + .category = miscCategory, .handler = {[&]() { useNet = false; }}, }); addFlag({ .longName = "refresh", .description = "Consider all previously downloaded files out-of-date.", + .category = miscCategory, .handler = {[&]() { refresh = true; }}, }); } @@ -187,7 +190,7 @@ static void showHelp(std::vector<std::string> subcommand, MultiCommand & topleve *vUtils); auto attrs = state.buildBindings(16); - attrs.alloc("command").mkString(toplevel.toJSON().dump()); + attrs.alloc("toplevel").mkString(toplevel.toJSON().dump()); auto vRes = state.allocValue(); state.callFunction(*vGenerateManpage, state.allocValue()->mkAttrs(attrs), *vRes, noPos); @@ -196,7 +199,7 @@ static void showHelp(std::vector<std::string> subcommand, MultiCommand & topleve if (!attr) throw UsageError("Nix has no subcommand '%s'", concatStringsSep("", subcommand)); - auto markdown = state.forceString(*attr->value); + auto markdown = state.forceString(*attr->value, noPos, "while evaluating the lowdown help text"); RunPager pager; std::cout << renderMarkdownToTerminal(markdown) << "\n"; @@ -266,7 +269,7 @@ void mainWrapped(int argc, char * * argv) programPath = argv[0]; auto programName = std::string(baseNameOf(programPath)); - if (argc > 0 && std::string_view(argv[0]) == "__build-remote") { + if (argc > 1 && std::string_view(argv[1]) == "__build-remote") { programName = "build-remote"; argv++; argc--; } @@ -289,7 +292,7 @@ void mainWrapped(int argc, char * * argv) NixArgs args; if (argc == 2 && std::string(argv[1]) == "__dump-args") { - std::cout << args.toJSON().dump() << "\n"; + logger->cout("%s", args.toJSON()); return; } @@ -309,7 +312,7 @@ void mainWrapped(int argc, char * * argv) b["doc"] = trim(stripIndentation(primOp->doc)); res[state.symbols[builtin.name]] = std::move(b); } - std::cout << res.dump() << "\n"; + logger->cout("%s", res); return; } @@ -318,14 +321,14 @@ void mainWrapped(int argc, char * * argv) if (completions) { switch (completionType) { case ctNormal: - std::cout << "normal\n"; break; + logger->cout("normal"); break; case ctFilenames: - std::cout << "filenames\n"; break; + logger->cout("filenames"); break; case ctAttrs: - std::cout << "attrs\n"; break; + logger->cout("attrs"); break; } for (auto & s : *completions) - std::cout << s.completion << "\t" << s.description << "\n"; + logger->cout(s.completion + "\t" + trim(s.description)); } }); @@ -391,7 +394,6 @@ void mainWrapped(int argc, char * * argv) if (args.command->second->forceImpureByDefault() && !evalSettings.pureEval.overridden) { evalSettings.pureEval = false; } - args.command->second->prepare(); args.command->second->run(); } diff --git a/src/nix/make-content-addressed.cc b/src/nix/make-content-addressed.cc index 34860c38f..d9c988a9f 100644 --- a/src/nix/make-content-addressed.cc +++ b/src/nix/make-content-addressed.cc @@ -2,10 +2,13 @@ #include "store-api.hh" #include "make-content-addressed.hh" #include "common-args.hh" -#include "json.hh" + +#include <nlohmann/json.hpp> using namespace nix; +using nlohmann::json; + struct CmdMakeContentAddressed : virtual CopyCommand, virtual StorePathsCommand, MixJSON { CmdMakeContentAddressed() @@ -33,13 +36,15 @@ struct CmdMakeContentAddressed : virtual CopyCommand, virtual StorePathsCommand, StorePathSet(storePaths.begin(), storePaths.end())); if (json) { - JSONObject jsonRoot(std::cout); - JSONObject jsonRewrites(jsonRoot.object("rewrites")); + auto jsonRewrites = json::object(); for (auto & path : storePaths) { auto i = remappings.find(path); assert(i != remappings.end()); - jsonRewrites.attr(srcStore->printStorePath(path), srcStore->printStorePath(i->second)); + jsonRewrites[srcStore->printStorePath(path)] = srcStore->printStorePath(i->second); } + auto json = json::object(); + json["rewrites"] = jsonRewrites; + logger->cout("%s", json); } else { for (auto & path : storePaths) { auto i = remappings.find(path); diff --git a/src/nix/make-content-addressed.md b/src/nix/make-content-addressed.md index 215683e6d..b1f7da525 100644 --- a/src/nix/make-content-addressed.md +++ b/src/nix/make-content-addressed.md @@ -22,7 +22,7 @@ R""( ```console # nix copy --to /tmp/nix --trusted-public-keys '' nixpkgs#hello - cannot add path '/nix/store/zy9wbxwcygrwnh8n2w9qbbcr6zk87m26-libunistring-0.9.10' because it lacks a valid signature + cannot add path '/nix/store/zy9wbxwcygrwnh8n2w9qbbcr6zk87m26-libunistring-0.9.10' because it lacks a signature by a trusted key ``` * Create a content-addressed representation of the current NixOS @@ -35,7 +35,9 @@ R""( # Description This command converts the closure of the store paths specified by -*installables* to content-addressed form. Nix store paths are usually +[*installables*](./nix.md#installables) to content-addressed form. + +Nix store paths are usually *input-addressed*, meaning that the hash part of the store path is computed from the contents of the derivation (i.e., the build-time dependency graph). Input-addressed paths need to be signed by a diff --git a/src/nix/nar.cc b/src/nix/nar.cc index dbb043d9b..9815410cf 100644 --- a/src/nix/nar.cc +++ b/src/nix/nar.cc @@ -25,7 +25,6 @@ struct CmdNar : NixMultiCommand { if (!command) throw UsageError("'nix nar' requires a sub-command."); - command->second->prepare(); command->second->run(); } }; diff --git a/src/nix/nix.md b/src/nix/nix.md index d48682a94..0a90fa6c9 100644 --- a/src/nix/nix.md +++ b/src/nix/nix.md @@ -48,103 +48,112 @@ manual](https://nixos.org/manual/nix/stable/). # Installables -Many `nix` subcommands operate on one or more *installables*. These are -command line arguments that represent something that can be built in -the Nix store. Here are the recognised types of installables: - -* **Flake output attributes**: `nixpkgs#hello` - - These have the form *flakeref*[`#`*attrpath*], where *flakeref* is a - flake reference and *attrpath* is an optional attribute path. For - more information on flakes, see [the `nix flake` manual - page](./nix3-flake.md). Flake references are most commonly a flake - identifier in the flake registry (e.g. `nixpkgs`), or a raw path - (e.g. `/path/to/my-flake` or `.` or `../foo`), or a full URL - (e.g. `github:nixos/nixpkgs` or `path:.`) - - When the flake reference is a raw path (a path without any URL - scheme), it is interpreted as a `path:` or `git+file:` url in the following - way: - - - If the path is within a Git repository, then the url will be of the form - `git+file://[GIT_REPO_ROOT]?dir=[RELATIVE_FLAKE_DIR_PATH]` - where `GIT_REPO_ROOT` is the path to the root of the git repository, - and `RELATIVE_FLAKE_DIR_PATH` is the path (relative to the directory - root) of the closest parent of the given path that contains a `flake.nix` within - the git repository. - If no such directory exists, then Nix will error-out. - - Note that the search will only include files indexed by git. In particular, files - which are matched by `.gitignore` or have never been `git add`-ed will not be - available in the flake. If this is undesirable, specify `path:<directory>` explicitly; - - For example, if `/foo/bar` is a git repository with the following structure: - ``` - . - └── baz - ├── blah - │ └── file.txt - └── flake.nix - ``` +Many `nix` subcommands operate on one or more *installables*. +These are command line arguments that represent something that can be realised in the Nix store. - Then `/foo/bar/baz/blah` will resolve to `git+file:///foo/bar?dir=baz` +The following types of installable are supported by most commands: - - If the supplied path is not a git repository, then the url will have the form - `path:FLAKE_DIR_PATH` where `FLAKE_DIR_PATH` is the closest parent - of the supplied path that contains a `flake.nix` file (within the same file-system). - If no such directory exists, then Nix will error-out. - - For example, if `/foo/bar/flake.nix` exists, then `/foo/bar/baz/` will resolve to - `path:/foo/bar` - - If *attrpath* is omitted, Nix tries some default values; for most - subcommands, the default is `packages.`*system*`.default` - (e.g. `packages.x86_64-linux.default`), but some subcommands have - other defaults. If *attrpath* *is* specified, *attrpath* is - interpreted as relative to one or more prefixes; for most - subcommands, these are `packages.`*system*, - `legacyPackages.*system*` and the empty prefix. Thus, on - `x86_64-linux` `nix build nixpkgs#hello` will try to build the - attributes `packages.x86_64-linux.hello`, - `legacyPackages.x86_64-linux.hello` and `hello`. - -* **Store paths**: `/nix/store/v5sv61sszx301i0x6xysaqzla09nksnd-hello-2.10` - - These are paths inside the Nix store, or symlinks that resolve to a - path in the Nix store. - -* **Store derivations**: `/nix/store/p7gp6lxdg32h4ka1q398wd9r2zkbbz2v-hello-2.10.drv` - - Store derivations are store paths with extension `.drv` and are a - low-level representation of a build-time dependency graph used - internally by Nix. By default, if you pass a store derivation to a - `nix` subcommand, it will operate on the *output paths* of the - derivation. For example, `nix path-info` prints information about - the output paths: +- [Flake output attribute](#flake-output-attribute) +- [Store path](#store-path) +- [Nix file](#nix-file), optionally qualified by an attribute path +- [Nix expression](#nix-expression), optionally qualified by an attribute path - ```console - # nix path-info --json /nix/store/p7gp6lxdg32h4ka1q398wd9r2zkbbz2v-hello-2.10.drv - [{"path":"/nix/store/v5sv61sszx301i0x6xysaqzla09nksnd-hello-2.10",…}] +For most commands, if no installable is specified, `.` as assumed. +That is, Nix will operate on the default flake output attribute of the flake in the current directory. + +### Flake output attribute + +Example: `nixpkgs#hello` + +These have the form *flakeref*[`#`*attrpath*], where *flakeref* is a +[flake reference](./nix3-flake.md#flake-references) and *attrpath* is an optional attribute path. For +more information on flakes, see [the `nix flake` manual +page](./nix3-flake.md). Flake references are most commonly a flake +identifier in the flake registry (e.g. `nixpkgs`), or a raw path +(e.g. `/path/to/my-flake` or `.` or `../foo`), or a full URL +(e.g. `github:nixos/nixpkgs` or `path:.`) + +When the flake reference is a raw path (a path without any URL +scheme), it is interpreted as a `path:` or `git+file:` url in the following +way: + +- If the path is within a Git repository, then the url will be of the form + `git+file://[GIT_REPO_ROOT]?dir=[RELATIVE_FLAKE_DIR_PATH]` + where `GIT_REPO_ROOT` is the path to the root of the git repository, + and `RELATIVE_FLAKE_DIR_PATH` is the path (relative to the directory + root) of the closest parent of the given path that contains a `flake.nix` within + the git repository. + If no such directory exists, then Nix will error-out. + + Note that the search will only include files indexed by git. In particular, files + which are matched by `.gitignore` or have never been `git add`-ed will not be + available in the flake. If this is undesirable, specify `path:<directory>` explicitly; + + For example, if `/foo/bar` is a git repository with the following structure: ``` + . + └── baz + ├── blah + │ └── file.txt + └── flake.nix + ``` + + Then `/foo/bar/baz/blah` will resolve to `git+file:///foo/bar?dir=baz` + +- If the supplied path is not a git repository, then the url will have the form + `path:FLAKE_DIR_PATH` where `FLAKE_DIR_PATH` is the closest parent + of the supplied path that contains a `flake.nix` file (within the same file-system). + If no such directory exists, then Nix will error-out. + + For example, if `/foo/bar/flake.nix` exists, then `/foo/bar/baz/` will resolve to + `path:/foo/bar` + +If *attrpath* is omitted, Nix tries some default values; for most +subcommands, the default is `packages.`*system*`.default` +(e.g. `packages.x86_64-linux.default`), but some subcommands have +other defaults. If *attrpath* *is* specified, *attrpath* is +interpreted as relative to one or more prefixes; for most +subcommands, these are `packages.`*system*, +`legacyPackages.*system*` and the empty prefix. Thus, on +`x86_64-linux` `nix build nixpkgs#hello` will try to build the +attributes `packages.x86_64-linux.hello`, +`legacyPackages.x86_64-linux.hello` and `hello`. + +### Store path + +Example: `/nix/store/v5sv61sszx301i0x6xysaqzla09nksnd-hello-2.10` + +These are paths inside the Nix store, or symlinks that resolve to a path in the Nix store. - If you want to operate on the store derivation itself, pass the - `--derivation` flag. +A [store derivation] is also addressed by store path. -* **Nix attributes**: `--file /path/to/nixpkgs hello` +Example: `/nix/store/p7gp6lxdg32h4ka1q398wd9r2zkbbz2v-hello-2.10.drv` - When the `-f` / `--file` *path* option is given, installables are - interpreted as attribute paths referencing a value returned by - evaluating the Nix file *path*. +If you want to refer to an output path of that store derivation, add the output name preceded by a caret (`^`). -* **Nix expressions**: `--expr '(import <nixpkgs> {}).hello.overrideDerivation (prev: { name = "my-hello"; })'`. +Example: `/nix/store/p7gp6lxdg32h4ka1q398wd9r2zkbbz2v-hello-2.10.drv^out` - When the `--expr` option is given, all installables are interpreted - as Nix expressions. You may need to specify `--impure` if the - expression references impure inputs (such as `<nixpkgs>`). +All outputs can be referred to at once with the special syntax `^*`. -For most commands, if no installable is specified, the default is `.`, -i.e. Nix will operate on the default flake output attribute of the -flake in the current directory. +Example: `/nix/store/p7gp6lxdg32h4ka1q398wd9r2zkbbz2v-hello-2.10.drv^*` + +### Nix file + +Example: `--file /path/to/nixpkgs hello` + +When the option `-f` / `--file` *path* \[*attrpath*...\] is given, installables are interpreted as the value of the expression in the Nix file at *path*. +If attribute paths are provided, commands will operate on the corresponding values accessible at these paths. +The Nix expression in that file, or any selected attribute, must evaluate to a derivation. + +### Nix expression + +Example: `--expr 'import <nixpkgs> {}' hello` + +When the option `--expr` *expression* \[*attrpath*...\] is given, installables are interpreted as the value of the of the Nix expression. +If attribute paths are provided, commands will operate on the corresponding values accessible at these paths. +The Nix expression, or any selected attribute, must evaluate to a derivation. + +You may need to specify `--impure` if the expression references impure inputs (such as `<nixpkgs>`). ## Derivation output selection @@ -164,6 +173,13 @@ operate are determined as follows: … ``` + and likewise, using a store path to a "drv" file to specify the derivation: + + ```console + # nix build '/nix/store/gzaflydcr6sb3567hap9q6srzx8ggdgg-glibc-2.33-78.drv^dev,static' + … + ``` + * You can also specify that *all* outputs should be used using the syntax *installable*`^*`. For example, the following shows the size of all outputs of the `glibc` package in the binary cache: @@ -177,6 +193,12 @@ operate are determined as follows: /nix/store/q6580lr01jpcsqs4r5arlh4ki2c1m9rv-glibc-2.33-123-dev 44200560 ``` + and likewise, using a store path to a "drv" file to specify the derivation: + + ```console + # nix path-info -S '/nix/store/gzaflydcr6sb3567hap9q6srzx8ggdgg-glibc-2.33-78.drv^*' + … + ``` * If you didn't specify the desired outputs, but the derivation has an attribute `meta.outputsToInstall`, Nix will use those outputs. For example, since the package `nixpkgs#libxml2` has this attribute: @@ -189,6 +211,11 @@ operate are determined as follows: a command like `nix shell nixpkgs#libxml2` will provide only those two outputs by default. + Note that a [store derivation] (given by its `.drv` file store path) doesn't have + any attributes like `meta`, and thus this case doesn't apply to it. + + [store derivation]: ../../glossary.md#gloss-store-derivation + * Otherwise, Nix will use all outputs of the derivation. # Nix stores diff --git a/src/nix/path-from-hash-part.cc b/src/nix/path-from-hash-part.cc new file mode 100644 index 000000000..7f7cda8d3 --- /dev/null +++ b/src/nix/path-from-hash-part.cc @@ -0,0 +1,39 @@ +#include "command.hh" +#include "store-api.hh" + +using namespace nix; + +struct CmdPathFromHashPart : StoreCommand +{ + std::string hashPart; + + CmdPathFromHashPart() + { + expectArgs({ + .label = "hash-part", + .handler = {&hashPart}, + }); + } + + std::string description() override + { + return "get a store path from its hash part"; + } + + std::string doc() override + { + return + #include "path-from-hash-part.md" + ; + } + + void run(ref<Store> store) override + { + if (auto storePath = store->queryPathFromHashPart(hashPart)) + logger->cout(store->printStorePath(*storePath)); + else + throw Error("there is no store path corresponding to '%s'", hashPart); + } +}; + +static auto rCmdPathFromHashPart = registerCommand2<CmdPathFromHashPart>({"store", "path-from-hash-part"}); diff --git a/src/nix/path-from-hash-part.md b/src/nix/path-from-hash-part.md new file mode 100644 index 000000000..788e13ab6 --- /dev/null +++ b/src/nix/path-from-hash-part.md @@ -0,0 +1,20 @@ +R""( + +# Examples + +* Return the full store path with the given hash part: + + ```console + # nix store path-from-hash-part --store https://cache.nixos.org/ 0i2jd68mp5g6h2sa5k9c85rb80sn8hi9 + /nix/store/0i2jd68mp5g6h2sa5k9c85rb80sn8hi9-hello-2.10 + ``` + +# Description + +Given the hash part of a store path (that is, the 32 characters +following `/nix/store/`), return the full store path. This is +primarily useful in the implementation of binary caches, where a +request for a `.narinfo` file only supplies the hash part +(e.g. `https://cache.nixos.org/0i2jd68mp5g6h2sa5k9c85rb80sn8hi9.narinfo`). + +)"" diff --git a/src/nix/path-info.cc b/src/nix/path-info.cc index d690fe594..613c5b191 100644 --- a/src/nix/path-info.cc +++ b/src/nix/path-info.cc @@ -1,12 +1,13 @@ #include "command.hh" #include "shared.hh" #include "store-api.hh" -#include "json.hh" #include "common-args.hh" #include <algorithm> #include <array> +#include <nlohmann/json.hpp> + using namespace nix; struct CmdPathInfo : StorePathsCommand, MixJSON @@ -86,11 +87,10 @@ struct CmdPathInfo : StorePathsCommand, MixJSON pathLen = std::max(pathLen, store->printStorePath(storePath).size()); if (json) { - JSONPlaceholder jsonRoot(std::cout); - store->pathInfoToJSON(jsonRoot, + std::cout << store->pathInfoToJSON( // FIXME: preserve order? StorePathSet(storePaths.begin(), storePaths.end()), - true, showClosureSize, SRI, AllowInvalid); + true, showClosureSize, SRI, AllowInvalid).dump(); } else { diff --git a/src/nix/path-info.md b/src/nix/path-info.md index 7a1714ba4..6ad23a02e 100644 --- a/src/nix/path-info.md +++ b/src/nix/path-info.md @@ -68,7 +68,9 @@ R""( ] ``` -* Print the path of the store derivation produced by `nixpkgs#hello`: +* Print the path of the [store derivation] produced by `nixpkgs#hello`: + + [store derivation]: ../../glossary.md#gloss-store-derivation ```console # nix path-info --derivation nixpkgs#hello @@ -78,7 +80,7 @@ R""( # Description This command shows information about the store paths produced by -*installables*, or about all paths in the store if you pass `--all`. +[*installables*](./nix.md#installables), or about all paths in the store if you pass `--all`. By default, this command only prints the store paths. You can get additional information by passing flags such as `--closure-size`, diff --git a/src/nix/ping-store.cc b/src/nix/ping-store.cc index 3c3b7bb45..5c44510ab 100644 --- a/src/nix/ping-store.cc +++ b/src/nix/ping-store.cc @@ -1,10 +1,13 @@ #include "command.hh" #include "shared.hh" #include "store-api.hh" +#include "finally.hh" + +#include <nlohmann/json.hpp> using namespace nix; -struct CmdPingStore : StoreCommand +struct CmdPingStore : StoreCommand, MixJSON { std::string description() override { @@ -20,10 +23,21 @@ struct CmdPingStore : StoreCommand void run(ref<Store> store) override { - notice("Store URL: %s", store->getUri()); - store->connect(); - if (auto version = store->getVersion()) - notice("Version: %s", *version); + if (!json) { + notice("Store URL: %s", store->getUri()); + store->connect(); + if (auto version = store->getVersion()) + notice("Version: %s", *version); + } else { + nlohmann::json res; + Finally printRes([&]() { + logger->cout("%s", res); + }); + res["url"] = store->getUri(); + store->connect(); + if (auto version = store->getVersion()) + res["version"] = *version; + } } }; diff --git a/src/nix/prefetch.cc b/src/nix/prefetch.cc index ce3288dc1..51c8a3319 100644 --- a/src/nix/prefetch.cc +++ b/src/nix/prefetch.cc @@ -28,17 +28,17 @@ std::string resolveMirrorUrl(EvalState & state, const std::string & url) Value vMirrors; // FIXME: use nixpkgs flake state.eval(state.parseExprFromString("import <nixpkgs/pkgs/build-support/fetchurl/mirrors.nix>", "."), vMirrors); - state.forceAttrs(vMirrors, noPos); + state.forceAttrs(vMirrors, noPos, "while evaluating the set of all mirrors"); auto mirrorList = vMirrors.attrs->find(state.symbols.create(mirrorName)); if (mirrorList == vMirrors.attrs->end()) throw Error("unknown mirror name '%s'", mirrorName); - state.forceList(*mirrorList->value, noPos); + state.forceList(*mirrorList->value, noPos, "while evaluating one mirror configuration"); if (mirrorList->value->listSize() < 1) throw Error("mirror URL '%s' did not expand to anything", url); - std::string mirror(state.forceString(*mirrorList->value->listElems()[0])); + std::string mirror(state.forceString(*mirrorList->value->listElems()[0], noPos, "while evaluating the first available mirror")); return mirror + (hasSuffix(mirror, "/") ? "" : "/") + s.substr(p + 1); } @@ -196,29 +196,29 @@ static int main_nix_prefetch_url(int argc, char * * argv) Value vRoot; state->evalFile(path, vRoot); Value & v(*findAlongAttrPath(*state, attrPath, autoArgs, vRoot).first); - state->forceAttrs(v, noPos); + state->forceAttrs(v, noPos, "while evaluating the source attribute to prefetch"); /* Extract the URL. */ auto * attr = v.attrs->get(state->symbols.create("urls")); if (!attr) throw Error("attribute 'urls' missing"); - state->forceList(*attr->value, noPos); + state->forceList(*attr->value, noPos, "while evaluating the urls to prefetch"); if (attr->value->listSize() < 1) throw Error("'urls' list is empty"); - url = state->forceString(*attr->value->listElems()[0]); + url = state->forceString(*attr->value->listElems()[0], noPos, "while evaluating the first url from the urls list"); /* Extract the hash mode. */ auto attr2 = v.attrs->get(state->symbols.create("outputHashMode")); if (!attr2) printInfo("warning: this does not look like a fetchurl call"); else - unpack = state->forceString(*attr2->value) == "recursive"; + unpack = state->forceString(*attr2->value, noPos, "while evaluating the outputHashMode of the source to prefetch") == "recursive"; /* Extract the name. */ if (!name) { auto attr3 = v.attrs->get(state->symbols.create("name")); if (!attr3) - name = state->forceString(*attr3->value); + name = state->forceString(*attr3->value, noPos, "while evaluating the name of the source to prefetch"); } } @@ -234,9 +234,9 @@ static int main_nix_prefetch_url(int argc, char * * argv) if (!printPath) printInfo("path is '%s'", store->printStorePath(storePath)); - std::cout << printHash16or32(hash) << std::endl; + logger->cout(printHash16or32(hash)); if (printPath) - std::cout << store->printStorePath(storePath) << std::endl; + logger->cout(store->printStorePath(storePath)); return 0; } diff --git a/src/nix/print-dev-env.md b/src/nix/print-dev-env.md index 2aad491de..a8ce9d36a 100644 --- a/src/nix/print-dev-env.md +++ b/src/nix/print-dev-env.md @@ -40,7 +40,7 @@ R""( This command prints a shell script that can be sourced by `bash` and that sets the variables and shell functions defined by the build -process of *installable*. This allows you to get a similar build +process of [*installable*](./nix.md#installables). This allows you to get a similar build environment in your current shell rather than in a subshell (as with `nix develop`). diff --git a/src/nix/profile-install.md b/src/nix/profile-install.md index aed414963..4c0f82c09 100644 --- a/src/nix/profile-install.md +++ b/src/nix/profile-install.md @@ -29,6 +29,6 @@ R""( # Description -This command adds *installables* to a Nix profile. +This command adds [*installables*](./nix.md#installables) to a Nix profile. )"" diff --git a/src/nix/profile-list.md b/src/nix/profile-list.md index bdab9a208..fa786162f 100644 --- a/src/nix/profile-list.md +++ b/src/nix/profile-list.md @@ -20,11 +20,11 @@ following fields: * An integer that can be used to unambiguously identify the package in invocations of `nix profile remove` and `nix profile upgrade`. -* The original ("mutable") flake reference and output attribute path +* The original ("unlocked") flake reference and output attribute path used at installation time. -* The immutable flake reference to which the mutable flake reference - was resolved. +* The locked flake reference to which the unlocked flake reference was + resolved. * The store path(s) of the package. diff --git a/src/nix/profile-upgrade.md b/src/nix/profile-upgrade.md index e06e74abe..39cca428b 100644 --- a/src/nix/profile-upgrade.md +++ b/src/nix/profile-upgrade.md @@ -2,7 +2,7 @@ R""( # Examples -* Upgrade all packages that were installed using a mutable flake +* Upgrade all packages that were installed using an unlocked flake reference: ```console @@ -32,9 +32,9 @@ the package was installed. > **Warning** > -> This only works if you used a *mutable* flake reference at +> This only works if you used an *unlocked* flake reference at > installation time, e.g. `nixpkgs#hello`. It does not work if you -> used an *immutable* flake reference +> used a *locked* flake reference > (e.g. `github:NixOS/nixpkgs/13d0c311e3ae923a00f734b43fd1d35b47d8943a#hello`), > since in that case the "latest version" is always the same. diff --git a/src/nix/profile.cc b/src/nix/profile.cc index 3814e7d5a..d72dd1a13 100644 --- a/src/nix/profile.cc +++ b/src/nix/profile.cc @@ -1,4 +1,5 @@ #include "command.hh" +#include "installable-flake.hh" #include "common-args.hh" #include "shared.hh" #include "store-api.hh" @@ -22,7 +23,7 @@ struct ProfileElementSource // FIXME: record original attrpath. FlakeRef resolvedRef; std::string attrPath; - OutputsSpec outputs; + ExtendedOutputsSpec outputs; bool operator < (const ProfileElementSource & other) const { @@ -32,17 +33,19 @@ struct ProfileElementSource } }; +const int defaultPriority = 5; + struct ProfileElement { StorePathSet storePaths; std::optional<ProfileElementSource> source; bool active = true; - int priority = 5; + int priority = defaultPriority; std::string describe() const { if (source) - return fmt("%s#%s%s", source->originalRef, source->attrPath, printOutputsSpec(source->outputs)); + return fmt("%s#%s%s", source->originalRef, source->attrPath, source->outputs.to_string()); StringSet names; for (auto & path : storePaths) names.insert(DrvName(path.name()).name); @@ -124,7 +127,7 @@ struct ProfileManifest parseFlakeRef(e[sOriginalUrl]), parseFlakeRef(e[sUrl]), e["attrPath"], - e["outputs"].get<OutputsSpec>() + e["outputs"].get<ExtendedOutputsSpec>() }; } elements.emplace_back(std::move(element)); @@ -225,12 +228,12 @@ struct ProfileManifest while (i != prevElems.end() || j != curElems.end()) { if (j != curElems.end() && (i == prevElems.end() || i->describe() > j->describe())) { - std::cout << fmt("%s%s: ∅ -> %s\n", indent, j->describe(), j->versions()); + logger->cout("%s%s: ∅ -> %s", indent, j->describe(), j->versions()); changes = true; ++j; } else if (i != prevElems.end() && (j == curElems.end() || i->describe() < j->describe())) { - std::cout << fmt("%s%s: %s -> ∅\n", indent, i->describe(), i->versions()); + logger->cout("%s%s: %s -> ∅", indent, i->describe(), i->versions()); changes = true; ++i; } @@ -238,7 +241,7 @@ struct ProfileManifest auto v1 = i->versions(); auto v2 = j->versions(); if (v1 != v2) { - std::cout << fmt("%s%s: %s -> %s\n", indent, i->describe(), v1, v2); + logger->cout("%s%s: %s -> %s", indent, i->describe(), v1, v2); changes = true; } ++i; @@ -247,17 +250,24 @@ struct ProfileManifest } if (!changes) - std::cout << fmt("%sNo changes.\n", indent); + logger->cout("%sNo changes.", indent); } }; -static std::map<Installable *, BuiltPaths> +static std::map<Installable *, std::pair<BuiltPaths, ExtraPathInfo>> builtPathsPerInstallable( - const std::vector<std::pair<std::shared_ptr<Installable>, BuiltPath>> & builtPaths) + const std::vector<std::pair<ref<Installable>, BuiltPathWithResult>> & builtPaths) { - std::map<Installable *, BuiltPaths> res; - for (auto & [installable, builtPath] : builtPaths) - res[installable.get()].push_back(builtPath); + std::map<Installable *, std::pair<BuiltPaths, ExtraPathInfo>> res; + for (auto & [installable, builtPath] : builtPaths) { + auto & r = res[&*installable]; + /* Note that there could be conflicting info + (e.g. meta.priority fields) if the installable returned + multiple derivations. So pick one arbitrarily. FIXME: + print a warning? */ + r.first.push_back(builtPath.path); + r.second = builtPath.info; + } return res; } @@ -286,7 +296,7 @@ struct CmdProfileInstall : InstallablesCommand, MixDefaultProfile ; } - void run(ref<Store> store) override + void run(ref<Store> store, Installables && installables) override { ProfileManifest manifest(*getEvalState(), *profile); @@ -297,33 +307,86 @@ struct CmdProfileInstall : InstallablesCommand, MixDefaultProfile for (auto & installable : installables) { ProfileElement element; + auto & [res, info] = builtPaths[&*installable]; - - if (auto installable2 = std::dynamic_pointer_cast<InstallableFlake>(installable)) { - // FIXME: make build() return this? - auto [attrPath, resolvedRef, drv] = installable2->toDerivation(); + if (info.originalRef && info.resolvedRef && info.attrPath && info.extendedOutputsSpec) { element.source = ProfileElementSource { - installable2->flakeRef, - resolvedRef, - attrPath, - installable2->outputsSpec + .originalRef = *info.originalRef, + .resolvedRef = *info.resolvedRef, + .attrPath = *info.attrPath, + .outputs = *info.extendedOutputsSpec, }; - - if(drv.priority) { - element.priority = *drv.priority; - } } - if(priority) { // if --priority was specified we want to override the priority of the installable - element.priority = *priority; - }; + // If --priority was specified we want to override the + // priority of the installable. + element.priority = + priority + ? *priority + : info.priority.value_or(defaultPriority); - element.updateStorePaths(getEvalStore(), store, builtPaths[installable.get()]); + element.updateStorePaths(getEvalStore(), store, res); manifest.elements.push_back(std::move(element)); } - updateProfile(manifest.build(store)); + try { + updateProfile(manifest.build(store)); + } catch (BuildEnvFileConflictError & conflictError) { + // FIXME use C++20 std::ranges once macOS has it + // See https://github.com/NixOS/nix/compare/3efa476c5439f8f6c1968a6ba20a31d1239c2f04..1fe5d172ece51a619e879c4b86f603d9495cc102 + auto findRefByFilePath = [&]<typename Iterator>(Iterator begin, Iterator end) { + for (auto it = begin; it != end; it++) { + auto profileElement = *it; + for (auto & storePath : profileElement.storePaths) { + if (conflictError.fileA.starts_with(store->printStorePath(storePath))) { + return std::pair(conflictError.fileA, profileElement.source->originalRef); + } + if (conflictError.fileB.starts_with(store->printStorePath(storePath))) { + return std::pair(conflictError.fileB, profileElement.source->originalRef); + } + } + } + throw conflictError; + }; + // There are 2 conflicting files. We need to find out which one is from the already installed package and + // which one is the package that is the new package that is being installed. + // The first matching package is the one that was already installed (original). + auto [originalConflictingFilePath, originalConflictingRef] = findRefByFilePath(manifest.elements.begin(), manifest.elements.end()); + // The last matching package is the one that was going to be installed (new). + auto [newConflictingFilePath, newConflictingRef] = findRefByFilePath(manifest.elements.rbegin(), manifest.elements.rend()); + + throw Error( + "An existing package already provides the following file:\n" + "\n" + " %1%\n" + "\n" + "This is the conflicting file from the new package:\n" + "\n" + " %2%\n" + "\n" + "To remove the existing package:\n" + "\n" + " nix profile remove %3%\n" + "\n" + "The new package can also be installed next to the existing one by assigning a different priority.\n" + "The conflicting packages have a priority of %5%.\n" + "To prioritise the new package:\n" + "\n" + " nix profile install %4% --priority %6%\n" + "\n" + "To prioritise the existing package:\n" + "\n" + " nix profile install %4% --priority %7%\n", + originalConflictingFilePath, + newConflictingFilePath, + originalConflictingRef.to_string(), + newConflictingRef.to_string(), + conflictError.priority, + conflictError.priority - 1, + conflictError.priority + 1 + ); + } } }; @@ -450,7 +513,7 @@ struct CmdProfileUpgrade : virtual SourceExprCommand, MixDefaultProfile, MixProf auto matchers = getMatchers(store); - std::vector<std::shared_ptr<Installable>> installables; + Installables installables; std::vector<size_t> indices; auto upgradedCount = 0; @@ -466,7 +529,7 @@ struct CmdProfileUpgrade : virtual SourceExprCommand, MixDefaultProfile, MixProf Activity act(*logger, lvlChatty, actUnknown, fmt("checking '%s' for updates", element.source->attrPath)); - auto installable = std::make_shared<InstallableFlake>( + auto installable = make_ref<InstallableFlake>( this, getEvalState(), FlakeRef(element.source->originalRef), @@ -476,18 +539,22 @@ struct CmdProfileUpgrade : virtual SourceExprCommand, MixDefaultProfile, MixProf Strings{}, lockFlags); - auto [attrPath, resolvedRef, drv] = installable->toDerivation(); + auto derivedPaths = installable->toDerivedPaths(); + if (derivedPaths.empty()) continue; + auto & info = derivedPaths[0].info; + + assert(info.resolvedRef && info.attrPath); - if (element.source->resolvedRef == resolvedRef) continue; + if (element.source->resolvedRef == info.resolvedRef) continue; printInfo("upgrading '%s' from flake '%s' to '%s'", - element.source->attrPath, element.source->resolvedRef, resolvedRef); + element.source->attrPath, element.source->resolvedRef, *info.resolvedRef); element.source = ProfileElementSource { - installable->flakeRef, - resolvedRef, - attrPath, - installable->outputsSpec + .originalRef = installable->flakeRef, + .resolvedRef = *info.resolvedRef, + .attrPath = *info.attrPath, + .outputs = installable->extendedOutputsSpec, }; installables.push_back(installable); @@ -515,7 +582,7 @@ struct CmdProfileUpgrade : virtual SourceExprCommand, MixDefaultProfile, MixProf for (size_t i = 0; i < installables.size(); ++i) { auto & installable = installables.at(i); auto & element = manifest.elements[indices.at(i)]; - element.updateStorePaths(getEvalStore(), store, builtPaths[installable.get()]); + element.updateStorePaths(getEvalStore(), store, builtPaths[&*installable].first); } updateProfile(manifest.build(store)); @@ -543,8 +610,8 @@ struct CmdProfileList : virtual EvalCommand, virtual StoreCommand, MixDefaultPro for (size_t i = 0; i < manifest.elements.size(); ++i) { auto & element(manifest.elements[i]); logger->cout("%d %s %s %s", i, - element.source ? element.source->originalRef.to_string() + "#" + element.source->attrPath + printOutputsSpec(element.source->outputs) : "-", - element.source ? element.source->resolvedRef.to_string() + "#" + element.source->attrPath + printOutputsSpec(element.source->outputs) : "-", + element.source ? element.source->originalRef.to_string() + "#" + element.source->attrPath + element.source->outputs.to_string() : "-", + element.source ? element.source->resolvedRef.to_string() + "#" + element.source->attrPath + element.source->outputs.to_string() : "-", concatStringsSep(" ", store->printStorePathSet(element.storePaths))); } } @@ -573,9 +640,9 @@ struct CmdProfileDiffClosures : virtual StoreCommand, MixDefaultProfile for (auto & gen : gens) { if (prevGen) { - if (!first) std::cout << "\n"; + if (!first) logger->cout(""); first = false; - std::cout << fmt("Version %d -> %d:\n", prevGen->number, gen.number); + logger->cout("Version %d -> %d:", prevGen->number, gen.number); printClosureDiff(store, store->followLinksToStorePath(prevGen->path), store->followLinksToStorePath(gen.path), @@ -611,10 +678,10 @@ struct CmdProfileHistory : virtual StoreCommand, EvalCommand, MixDefaultProfile for (auto & gen : gens) { ProfileManifest manifest(*getEvalState(), gen.path); - if (!first) std::cout << "\n"; + if (!first) logger->cout(""); first = false; - std::cout << fmt("Version %s%d" ANSI_NORMAL " (%s)%s:\n", + logger->cout("Version %s%d" ANSI_NORMAL " (%s)%s:", gen.number == curGen ? ANSI_GREEN : ANSI_BOLD, gen.number, std::put_time(std::gmtime(&gen.creationTime), "%Y-%m-%d"), @@ -731,7 +798,6 @@ struct CmdProfile : NixMultiCommand { if (!command) throw UsageError("'nix profile' requires a sub-command."); - command->second->prepare(); command->second->run(); } }; diff --git a/src/nix/profile.md b/src/nix/profile.md index 8dade051d..273e02280 100644 --- a/src/nix/profile.md +++ b/src/nix/profile.md @@ -11,7 +11,7 @@ them to be rolled back easily. The default profile used by `nix profile` is `$HOME/.nix-profile`, which, if it does not exist, is created as a symlink to -`/nix/var/nix/profiles/per-user/default` if Nix is invoked by the +`/nix/var/nix/profiles/default` if Nix is invoked by the `root` user, or `/nix/var/nix/profiles/per-user/`*username* otherwise. You can specify another profile location using `--profile` *path*. @@ -88,8 +88,7 @@ has the following fields: the user at the time of installation (e.g. `nixpkgs`). This is also the flake reference that will be used by `nix profile upgrade`. -* `uri`: The immutable flake reference to which `originalUrl` - resolved. +* `uri`: The locked flake reference to which `originalUrl` resolved. * `attrPath`: The flake output attribute that provided this package. Note that this is not necessarily the attribute that the diff --git a/src/nix/realisation.cc b/src/nix/realisation.cc index c9a7157cd..13db80282 100644 --- a/src/nix/realisation.cc +++ b/src/nix/realisation.cc @@ -21,7 +21,6 @@ struct CmdRealisation : virtual NixMultiCommand { if (!command) throw UsageError("'nix realisation' requires a sub-command."); - command->second->prepare(); command->second->run(); } }; @@ -65,18 +64,16 @@ struct CmdRealisationInfo : BuiltPathsCommand, MixJSON res.push_back(currentPath); } - std::cout << res.dump(); + logger->cout("%s", res); } else { for (auto & path : realisations) { if (auto realisation = std::get_if<Realisation>(&path.raw)) { - std::cout << - realisation->id.to_string() << " " << - store->printStorePath(realisation->outPath); + logger->cout("%s %s", + realisation->id.to_string(), + store->printStorePath(realisation->outPath)); } else - std::cout << store->printStorePath(path.path()); - - std::cout << std::endl; + logger->cout("%s", store->printStorePath(path.path())); } } } diff --git a/src/nix/registry.cc b/src/nix/registry.cc index c496f94f8..1f4f820ac 100644 --- a/src/nix/registry.cc +++ b/src/nix/registry.cc @@ -183,14 +183,12 @@ struct CmdRegistryPin : RegistryCommand, EvalCommand void run(nix::ref<nix::Store> store) override { - if (locked.empty()) { - locked = url; - } + if (locked.empty()) locked = url; auto registry = getRegistry(); auto ref = parseFlakeRef(url); - auto locked_ref = parseFlakeRef(locked); + auto lockedRef = parseFlakeRef(locked); registry->remove(ref.input); - auto [tree, resolved] = locked_ref.resolve(store).input.fetch(store); + auto [tree, resolved] = lockedRef.resolve(store).input.fetch(store); fetchers::Attrs extraAttrs; if (ref.subdir != "") extraAttrs["dir"] = ref.subdir; registry->add(ref.input, resolved, extraAttrs); @@ -229,7 +227,6 @@ struct CmdRegistry : virtual NixMultiCommand settings.requireExperimentalFeature(Xp::Flakes); if (!command) throw UsageError("'nix registry' requires a sub-command."); - command->second->prepare(); command->second->run(); } }; diff --git a/src/nix/repl.cc b/src/nix/repl.cc new file mode 100644 index 000000000..51d3074b4 --- /dev/null +++ b/src/nix/repl.cc @@ -0,0 +1,92 @@ +#include "eval.hh" +#include "globals.hh" +#include "command.hh" +#include "repl.hh" + +namespace nix { + +struct CmdRepl : RawInstallablesCommand +{ + CmdRepl() { + evalSettings.pureEval = false; + } + + std::vector<std::string> files; + + Strings getDefaultFlakeAttrPaths() override + { + return {""}; + } + + bool forceImpureByDefault() override + { + return true; + } + + std::string description() override + { + return "start an interactive environment for evaluating Nix expressions"; + } + + std::string doc() override + { + return + #include "repl.md" + ; + } + + void applyDefaultInstallables(std::vector<std::string> & rawInstallables) override + { + if (!settings.isExperimentalFeatureEnabled(Xp::ReplFlake) && !(file) && rawInstallables.size() >= 1) { + warn("future versions of Nix will require using `--file` to load a file"); + if (rawInstallables.size() > 1) + warn("more than one input file is not currently supported"); + auto filePath = rawInstallables[0].data(); + file = std::optional(filePath); + rawInstallables.front() = rawInstallables.back(); + rawInstallables.pop_back(); + } + if (rawInstallables.empty() && (file.has_value() || expr.has_value())) { + rawInstallables.push_back("."); + } + } + + void run(ref<Store> store, std::vector<std::string> && rawInstallables) override + { + auto state = getEvalState(); + auto getValues = [&]()->AbstractNixRepl::AnnotatedValues{ + auto installables = parseInstallables(store, rawInstallables); + AbstractNixRepl::AnnotatedValues values; + for (auto & installable: installables){ + auto what = installable->what(); + if (file){ + auto [val, pos] = installable->toValue(*state); + auto what = installable->what(); + state->forceValue(*val, pos); + auto autoArgs = getAutoArgs(*state); + auto valPost = state->allocValue(); + state->autoCallFunction(*autoArgs, *val, *valPost); + state->forceValue(*valPost, pos); + values.push_back( {valPost, what }); + } else { + auto [val, pos] = installable->toValue(*state); + values.push_back( {val, what} ); + } + } + return values; + }; + auto repl = AbstractNixRepl::create( + searchPath, + openStore(), + state, + getValues + ); + repl->autoArgs = getAutoArgs(*repl->state); + repl->initEnv(); + repl->mainLoop(); + } +}; + +static auto rCmdRepl = registerCommand<CmdRepl>("repl"); + +} diff --git a/src/nix/repl.md b/src/nix/repl.md index 23ef0f4e6..c5113be61 100644 --- a/src/nix/repl.md +++ b/src/nix/repl.md @@ -36,7 +36,7 @@ R""( Loading Installable ''... Added 1 variables. - # nix repl --extra_experimental_features 'flakes repl-flake' nixpkgs + # nix repl --extra-experimental-features 'flakes repl-flake' nixpkgs Loading Installable 'flake:nixpkgs#'... Added 5 variables. diff --git a/src/nix/run.cc b/src/nix/run.cc index 45d2dfd0d..56605d9d5 100644 --- a/src/nix/run.cc +++ b/src/nix/run.cc @@ -9,6 +9,7 @@ #include "fs-accessor.hh" #include "progress-bar.hh" #include "eval.hh" +#include "build/personality.hh" #if __linux__ #include <sys/mount.h> @@ -24,7 +25,8 @@ namespace nix { void runProgramInStore(ref<Store> store, const std::string & program, - const Strings & args) + const Strings & args, + std::optional<std::string_view> system) { stopProgressBar(); @@ -44,7 +46,7 @@ void runProgramInStore(ref<Store> store, throw Error("store '%s' is not a local store so it does not support command execution", store->getUri()); if (store->storeDir != store2->getRealStoreDir()) { - Strings helperArgs = { chrootHelperName, store->storeDir, store2->getRealStoreDir(), program }; + Strings helperArgs = { chrootHelperName, store->storeDir, store2->getRealStoreDir(), std::string(system.value_or("")), program }; for (auto & arg : args) helperArgs.push_back(arg); execv(getSelfExe().value_or("nix").c_str(), stringsToCharPtrs(helperArgs).data()); @@ -52,6 +54,9 @@ void runProgramInStore(ref<Store> store, throw SysError("could not execute chroot helper"); } + if (system) + setPersonality(*system); + execvp(program.c_str(), stringsToCharPtrs(args).data()); throw SysError("unable to execute '%s'", program); @@ -92,7 +97,7 @@ struct CmdShell : InstallablesCommand, MixEnvironment ; } - void run(ref<Store> store) override + void run(ref<Store> store, Installables && installables) override { auto outPaths = Installable::toStorePaths(getEvalStore(), store, Realise::Outputs, OperateOn::Output, installables); @@ -178,7 +183,7 @@ struct CmdRun : InstallableCommand return res; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { auto state = getEvalState(); @@ -199,6 +204,7 @@ void chrootHelper(int argc, char * * argv) int p = 1; std::string storeDir = argv[p++]; std::string realStoreDir = argv[p++]; + std::string system = argv[p++]; std::string cmd = argv[p++]; Strings args; while (p < argc) @@ -262,6 +268,9 @@ void chrootHelper(int argc, char * * argv) writeFile("/proc/self/uid_map", fmt("%d %d %d", uid, uid, 1)); writeFile("/proc/self/gid_map", fmt("%d %d %d", gid, gid, 1)); + if (system != "") + setPersonality(system); + execvp(cmd.c_str(), stringsToCharPtrs(args).data()); throw SysError("unable to exec '%s'", cmd); diff --git a/src/nix/run.hh b/src/nix/run.hh index 6180a87dd..fed360158 100644 --- a/src/nix/run.hh +++ b/src/nix/run.hh @@ -6,6 +6,7 @@ namespace nix { void runProgramInStore(ref<Store> store, const std::string & program, - const Strings & args); + const Strings & args, + std::optional<std::string_view> system = std::nullopt); } diff --git a/src/nix/run.md b/src/nix/run.md index a0f362076..250ea65aa 100644 --- a/src/nix/run.md +++ b/src/nix/run.md @@ -35,7 +35,7 @@ R""( # Description -`nix run` builds and runs *installable*, which must evaluate to an +`nix run` builds and runs [*installable*](./nix.md#installables), which must evaluate to an *app* or a regular Nix derivation. If *installable* evaluates to an *app* (see below), it executes the diff --git a/src/nix/search.cc b/src/nix/search.cc index bdd45cbed..994ec44c2 100644 --- a/src/nix/search.cc +++ b/src/nix/search.cc @@ -5,7 +5,6 @@ #include "names.hh" #include "get-drvs.hh" #include "common-args.hh" -#include "json.hh" #include "shared.hh" #include "eval-cache.hh" #include "attr-path.hh" @@ -13,8 +12,10 @@ #include <regex> #include <fstream> +#include <nlohmann/json.hpp> using namespace nix; +using json = nlohmann::json; std::string wrap(std::string prefix, std::string s) { @@ -55,12 +56,12 @@ struct CmdSearch : InstallableCommand, MixJSON Strings getDefaultFlakeAttrPaths() override { return { - "packages." + settings.thisSystem.get() + ".", - "legacyPackages." + settings.thisSystem.get() + "." + "packages." + settings.thisSystem.get(), + "legacyPackages." + settings.thisSystem.get() }; } - void run(ref<Store> store) override + void run(ref<Store> store, ref<Installable> installable) override { settings.readOnlyMode = true; evalSettings.enableImportFromDerivation.setDefault(false); @@ -84,7 +85,8 @@ struct CmdSearch : InstallableCommand, MixJSON auto state = getEvalState(); - auto jsonOut = json ? std::make_unique<JSONObject>(std::cout) : nullptr; + std::optional<nlohmann::json> jsonOut; + if (json) jsonOut = json::object(); uint64_t results = 0; @@ -151,10 +153,11 @@ struct CmdSearch : InstallableCommand, MixJSON { results++; if (json) { - auto jsonElem = jsonOut->object(attrPath2); - jsonElem.attr("pname", name.name); - jsonElem.attr("version", name.version); - jsonElem.attr("description", description); + (*jsonOut)[attrPath2] = { + {"pname", name.name}, + {"version", name.version}, + {"description", description}, + }; } else { auto name2 = hiliteMatches(name.name, nameMatches, ANSI_GREEN, "\e[0;2m"); if (results > 1) logger->cout(""); @@ -193,6 +196,9 @@ struct CmdSearch : InstallableCommand, MixJSON for (auto & cursor : installable->getCursors(*state)) visit(*cursor, cursor->getAttrPath(), true); + if (json) + logger->cout("%s", *jsonOut); + if (!json && !results) throw Error("no results for the given search term(s)!"); } diff --git a/src/nix/search.md b/src/nix/search.md index 5a5b5ae05..4caa90654 100644 --- a/src/nix/search.md +++ b/src/nix/search.md @@ -62,10 +62,10 @@ R""( # Description -`nix search` searches *installable* (which must be evaluatable, e.g. a -flake) for packages whose name or description matches all of the +`nix search` searches [*installable*](./nix.md#installables) (which can be evaluated, that is, a +flake or Nix expression, but not a store path or store derivation path) for packages whose name or description matches all of the regular expressions *regex*. For each matching package, It prints the -full attribute name (from the root of the installable), the version +full attribute name (from the root of the [installable](./nix.md#installables)), the version and the `meta.description` field, highlighting the substrings that were matched by the regular expressions. If no regular expressions are specified, all packages are shown. diff --git a/src/nix/shell.md b/src/nix/shell.md index 90b81fb2f..13a389103 100644 --- a/src/nix/shell.md +++ b/src/nix/shell.md @@ -23,6 +23,12 @@ R""( Hi everybody! ``` +* Run multiple commands in a shell environment: + + ```console + # nix shell nixpkgs#gnumake -c sh -c "cd src && make" + ``` + * Run GNU Hello in a chroot store: ```console @@ -42,7 +48,7 @@ R""( # Description `nix shell` runs a command in an environment in which the `$PATH` variable -provides the specified *installables*. If no command is specified, it starts the +provides the specified [*installables*](./nix.md#installable). If no command is specified, it starts the default shell of your user account specified by `$SHELL`. )"" diff --git a/src/nix/show-config.cc b/src/nix/show-config.cc index 29944e748..3530584f9 100644 --- a/src/nix/show-config.cc +++ b/src/nix/show-config.cc @@ -9,15 +9,44 @@ using namespace nix; struct CmdShowConfig : Command, MixJSON { + std::optional<std::string> name; + + CmdShowConfig() { + expectArgs({ + .label = {"name"}, + .optional = true, + .handler = {&name}, + }); + } + std::string description() override { - return "show the Nix configuration"; + return "show the Nix configuration or the value of a specific setting"; } Category category() override { return catUtility; } void run() override { + if (name) { + if (json) { + throw UsageError("'--json' is not supported when specifying a setting name"); + } + + std::map<std::string, Config::SettingInfo> settings; + globalConfig.getSettings(settings); + auto setting = settings.find(*name); + + if (setting == settings.end()) { + throw Error("could not find setting '%1%'", *name); + } else { + const auto & value = setting->second.value; + logger->cout("%s", value); + } + + return; + } + if (json) { // FIXME: use appropriate JSON types (bool, ints, etc). logger->cout("%s", globalConfig.toJSON().dump()); diff --git a/src/nix/show-derivation.cc b/src/nix/show-derivation.cc index fb46b4dbf..4a406ae08 100644 --- a/src/nix/show-derivation.cc +++ b/src/nix/show-derivation.cc @@ -5,10 +5,11 @@ #include "common-args.hh" #include "store-api.hh" #include "archive.hh" -#include "json.hh" #include "derivations.hh" +#include <nlohmann/json.hpp> using namespace nix; +using json = nlohmann::json; struct CmdShowDerivation : InstallablesCommand { @@ -38,7 +39,7 @@ struct CmdShowDerivation : InstallablesCommand Category category() override { return catUtility; } - void run(ref<Store> store) override + void run(ref<Store> store, Installables && installables) override { auto drvPaths = Installable::toDerivations(store, installables, true); @@ -48,77 +49,15 @@ struct CmdShowDerivation : InstallablesCommand drvPaths = std::move(closure); } - { - - JSONObject jsonRoot(std::cout, true); + json jsonRoot = json::object(); for (auto & drvPath : drvPaths) { if (!drvPath.isDerivation()) continue; - auto drvObj(jsonRoot.object(store->printStorePath(drvPath))); - - auto drv = store->readDerivation(drvPath); - - { - auto outputsObj(drvObj.object("outputs")); - for (auto & [_outputName, output] : drv.outputs) { - auto & outputName = _outputName; // work around clang bug - auto outputObj { outputsObj.object(outputName) }; - std::visit(overloaded { - [&](const DerivationOutput::InputAddressed & doi) { - outputObj.attr("path", store->printStorePath(doi.path)); - }, - [&](const DerivationOutput::CAFixed & dof) { - outputObj.attr("path", store->printStorePath(dof.path(*store, drv.name, outputName))); - outputObj.attr("hashAlgo", dof.hash.printMethodAlgo()); - outputObj.attr("hash", dof.hash.hash.to_string(Base16, false)); - }, - [&](const DerivationOutput::CAFloating & dof) { - outputObj.attr("hashAlgo", makeFileIngestionPrefix(dof.method) + printHashType(dof.hashType)); - }, - [&](const DerivationOutput::Deferred &) {}, - [&](const DerivationOutput::Impure & doi) { - outputObj.attr("hashAlgo", makeFileIngestionPrefix(doi.method) + printHashType(doi.hashType)); - outputObj.attr("impure", true); - }, - }, output.raw()); - } - } - - { - auto inputsList(drvObj.list("inputSrcs")); - for (auto & input : drv.inputSrcs) - inputsList.elem(store->printStorePath(input)); - } - - { - auto inputDrvsObj(drvObj.object("inputDrvs")); - for (auto & input : drv.inputDrvs) { - auto inputList(inputDrvsObj.list(store->printStorePath(input.first))); - for (auto & outputId : input.second) - inputList.elem(outputId); - } - } - - drvObj.attr("system", drv.platform); - drvObj.attr("builder", drv.builder); - - { - auto argsList(drvObj.list("args")); - for (auto & arg : drv.args) - argsList.elem(arg); - } - - { - auto envObj(drvObj.object("env")); - for (auto & var : drv.env) - envObj.attr(var.first, var.second); - } + jsonRoot[store->printStorePath(drvPath)] = + store->readDerivation(drvPath).toJSON(*store); } - - } - - std::cout << "\n"; + logger->cout(jsonRoot.dump(2)); } }; diff --git a/src/nix/show-derivation.md b/src/nix/show-derivation.md index aa863899c..1d37c6f5a 100644 --- a/src/nix/show-derivation.md +++ b/src/nix/show-derivation.md @@ -2,9 +2,11 @@ R""( # Examples -* Show the store derivation that results from evaluating the Hello +* Show the [store derivation] that results from evaluating the Hello package: + [store derivation]: ../../glossary.md#gloss-store-derivation + ```console # nix show-derivation nixpkgs#hello { @@ -37,7 +39,7 @@ R""( # Description This command prints on standard output a JSON representation of the -store derivations to which *installables* evaluate. Store derivations +[store derivation]s to which [*installables*](./nix.md#installables) evaluate. Store derivations are used internally by Nix. They are store paths with extension `.drv` that represent the build-time dependency graph to which a Nix expression evaluates. diff --git a/src/nix/sigs.cc b/src/nix/sigs.cc index 3d659d6d2..45cd2e1a6 100644 --- a/src/nix/sigs.cc +++ b/src/nix/sigs.cc @@ -45,7 +45,7 @@ struct CmdCopySigs : StorePathsCommand //logger->setExpected(doneLabel, storePaths.size()); auto doPath = [&](const Path & storePathS) { - //Activity act(*logger, lvlInfo, format("getting signatures for '%s'") % storePath); + //Activity act(*logger, lvlInfo, "getting signatures for '%s'", storePath); checkInterrupt(); @@ -173,7 +173,7 @@ struct CmdKeyGenerateSecret : Command if (!keyName) throw UsageError("required argument '--key-name' is missing"); - std::cout << SecretKey::generate(*keyName).to_string(); + writeFull(STDOUT_FILENO, SecretKey::generate(*keyName).to_string()); } }; @@ -194,7 +194,7 @@ struct CmdKeyConvertSecretToPublic : Command void run() override { SecretKey secretKey(drainFD(STDIN_FILENO)); - std::cout << secretKey.toPublicKey().to_string(); + writeFull(STDOUT_FILENO, secretKey.toPublicKey().to_string()); } }; @@ -219,7 +219,6 @@ struct CmdKey : NixMultiCommand { if (!command) throw UsageError("'nix key' requires a sub-command."); - command->second->prepare(); command->second->run(); } }; diff --git a/src/nix/store-copy-log.cc b/src/nix/store-copy-log.cc index 2e288f743..a6e8aeff7 100644 --- a/src/nix/store-copy-log.cc +++ b/src/nix/store-copy-log.cc @@ -24,22 +24,14 @@ struct CmdCopyLog : virtual CopyCommand, virtual InstallablesCommand ; } - Category category() override { return catUtility; } - - void run(ref<Store> srcStore) override + void run(ref<Store> srcStore, Installables && installables) override { auto & srcLogStore = require<LogStore>(*srcStore); auto dstStore = getDstStore(); auto & dstLogStore = require<LogStore>(*dstStore); - StorePathSet drvPaths; - - for (auto & i : installables) - for (auto & drvPath : i->toDrvPaths(getEvalStore())) - drvPaths.insert(drvPath); - - for (auto & drvPath : drvPaths) { + for (auto & drvPath : Installable::toDerivations(getEvalStore(), installables, true)) { if (auto log = srcLogStore.getBuildLog(drvPath)) dstLogStore.addBuildLog(drvPath, *log); else diff --git a/src/nix/store-copy-log.md b/src/nix/store-copy-log.md index 19ae57079..0937250f2 100644 --- a/src/nix/store-copy-log.md +++ b/src/nix/store-copy-log.md @@ -18,7 +18,9 @@ R""( (The flag `--substituters ''` avoids querying `https://cache.nixos.org` for the log.) -* To copy the log for a specific store derivation via SSH: +* To copy the log for a specific [store derivation] via SSH: + + [store derivation]: ../../glossary.md#gloss-store-derivation ```console # nix store copy-log --to ssh-ng://machine /nix/store/ilgm50plpmcgjhcp33z6n4qbnpqfhxym-glibc-2.33-59.drv diff --git a/src/nix/store-delete.cc b/src/nix/store-delete.cc index ca43f1530..6719227df 100644 --- a/src/nix/store-delete.cc +++ b/src/nix/store-delete.cc @@ -32,7 +32,7 @@ struct CmdStoreDelete : StorePathsCommand ; } - void run(ref<Store> store, std::vector<StorePath> && storePaths) override + void run(ref<Store> store, StorePaths && storePaths) override { auto & gcStore = require<GcStore>(*store); diff --git a/src/nix/store-delete.md b/src/nix/store-delete.md index db535f87c..431bc5f5e 100644 --- a/src/nix/store-delete.md +++ b/src/nix/store-delete.md @@ -10,7 +10,7 @@ R""( # Description -This command deletes the store paths specified by *installables*. , +This command deletes the store paths specified by [*installables*](./nix.md#installables), but only if it is safe to do so; that is, when the path is not reachable from a root of the garbage collector. This means that you can only delete paths that would also be deleted by `nix store diff --git a/src/nix/store-dump-path.md b/src/nix/store-dump-path.md index 4ef563526..56e2174b6 100644 --- a/src/nix/store-dump-path.md +++ b/src/nix/store-dump-path.md @@ -18,6 +18,6 @@ R""( # Description This command generates a NAR file containing the serialisation of the -store path *installable*. The NAR is written to standard output. +store path [*installable*](./nix.md#installables). The NAR is written to standard output. )"" diff --git a/src/nix/store-repair.cc b/src/nix/store-repair.cc index 8fcb3639a..895e39685 100644 --- a/src/nix/store-repair.cc +++ b/src/nix/store-repair.cc @@ -17,7 +17,7 @@ struct CmdStoreRepair : StorePathsCommand ; } - void run(ref<Store> store, std::vector<StorePath> && storePaths) override + void run(ref<Store> store, StorePaths && storePaths) override { for (auto & path : storePaths) store->repairPath(path); diff --git a/src/nix/store-repair.md b/src/nix/store-repair.md index 92d2205a9..180c577ac 100644 --- a/src/nix/store-repair.md +++ b/src/nix/store-repair.md @@ -17,7 +17,7 @@ R""( # Description This command attempts to "repair" the store paths specified by -*installables* by redownloading them using the available +[*installables*](./nix.md#installables) by redownloading them using the available substituters. If no substitutes are available, then repair is not possible. diff --git a/src/nix/store.cc b/src/nix/store.cc index 44e53c7c7..2879e03b3 100644 --- a/src/nix/store.cc +++ b/src/nix/store.cc @@ -18,7 +18,6 @@ struct CmdStore : virtual NixMultiCommand { if (!command) throw UsageError("'nix store' requires a sub-command."); - command->second->prepare(); command->second->run(); } }; diff --git a/src/nix/upgrade-nix.cc b/src/nix/upgrade-nix.cc index 2d2453395..17796d6b8 100644 --- a/src/nix/upgrade-nix.cc +++ b/src/nix/upgrade-nix.cc @@ -144,7 +144,7 @@ struct CmdUpgradeNix : MixDryRun, StoreCommand Bindings & bindings(*state->allocBindings(0)); auto v2 = findAlongAttrPath(*state, settings.thisSystem, bindings, *v).first; - return store->parseStorePath(state->forceString(*v2)); + return store->parseStorePath(state->forceString(*v2, noPos, "while evaluating the path tho latest nix version")); } }; diff --git a/src/nix/verify.cc b/src/nix/verify.cc index e92df1303..0b306cc11 100644 --- a/src/nix/verify.cc +++ b/src/nix/verify.cc @@ -41,7 +41,7 @@ struct CmdVerify : StorePathsCommand addFlag({ .longName = "sigs-needed", .shortName = 'n', - .description = "Require that each path has at least *n* valid signatures.", + .description = "Require that each path is signed by at least *n* different keys.", .labels = {"n"}, .handler = {&sigsNeeded} }); @@ -81,14 +81,14 @@ struct CmdVerify : StorePathsCommand ThreadPool pool; - auto doPath = [&](const Path & storePath) { + auto doPath = [&](const StorePath & storePath) { try { checkInterrupt(); MaintainCount<std::atomic<size_t>> mcActive(active); update(); - auto info = store->queryPathInfo(store->parseStorePath(storePath)); + auto info = store->queryPathInfo(storePath); // Note: info->path can be different from storePath // for binary cache stores when using --all (since we @@ -173,7 +173,7 @@ struct CmdVerify : StorePathsCommand }; for (auto & storePath : storePaths) - pool.enqueue(std::bind(doPath, store->printStorePath(storePath))); + pool.enqueue(std::bind(doPath, storePath)); pool.process(); diff --git a/src/nix/verify.md b/src/nix/verify.md index 1c43792e7..cc1122c02 100644 --- a/src/nix/verify.md +++ b/src/nix/verify.md @@ -24,7 +24,7 @@ R""( # Description -This command verifies the integrity of the store paths *installables*, +This command verifies the integrity of the store paths [*installables*](./nix.md#installables), or, if `--all` is given, the entire Nix store. For each path, it checks that diff --git a/src/nix/why-depends.cc b/src/nix/why-depends.cc index 1d9ab28ba..a3a9dc698 100644 --- a/src/nix/why-depends.cc +++ b/src/nix/why-depends.cc @@ -27,7 +27,7 @@ static std::string filterPrintable(const std::string & s) return res; } -struct CmdWhyDepends : SourceExprCommand +struct CmdWhyDepends : SourceExprCommand, MixOperateOnOptions { std::string _package, _dependency; bool all = false; @@ -83,20 +83,37 @@ struct CmdWhyDepends : SourceExprCommand { auto package = parseInstallable(store, _package); auto packagePath = Installable::toStorePath(getEvalStore(), store, Realise::Outputs, operateOn, package); + + /* We don't need to build `dependency`. We try to get the store + * path if it's already known, and if not, then it's not a dependency. + * + * Why? If `package` does depends on `dependency`, then getting the + * store path of `package` above necessitated having the store path + * of `dependency`. The contrapositive is, if the store path of + * `dependency` is not already known at this point (i.e. it's a CA + * derivation which hasn't been built), then `package` did not need it + * to build. + */ auto dependency = parseInstallable(store, _dependency); - auto dependencyPath = Installable::toStorePath(getEvalStore(), store, Realise::Derivation, operateOn, dependency); - auto dependencyPathHash = dependencyPath.hashPart(); + auto optDependencyPath = [&]() -> std::optional<StorePath> { + try { + return {Installable::toStorePath(getEvalStore(), store, Realise::Derivation, operateOn, dependency)}; + } catch (MissingRealisation &) { + return std::nullopt; + } + }(); StorePathSet closure; store->computeFSClosure({packagePath}, closure, false, false); - if (!closure.count(dependencyPath)) { - printError("'%s' does not depend on '%s'", - store->printStorePath(packagePath), - store->printStorePath(dependencyPath)); + if (!optDependencyPath.has_value() || !closure.count(*optDependencyPath)) { + printError("'%s' does not depend on '%s'", package->what(), dependency->what()); return; } + auto dependencyPath = *optDependencyPath; + auto dependencyPathHash = dependencyPath.hashPart(); + stopProgressBar(); // FIXME auto accessor = store->getFSAccessor(); diff --git a/src/resolve-system-dependencies/resolve-system-dependencies.cc b/src/resolve-system-dependencies/resolve-system-dependencies.cc index c6023eb03..4ea268d24 100644 --- a/src/resolve-system-dependencies/resolve-system-dependencies.cc +++ b/src/resolve-system-dependencies/resolve-system-dependencies.cc @@ -157,13 +157,9 @@ int main(int argc, char ** argv) uname(&_uname); - auto cacheParentDir = (format("%1%/dependency-maps") % settings.nixStateDir).str(); + auto cacheParentDir = fmt("%1%/dependency-maps", settings.nixStateDir); - cacheDir = (format("%1%/%2%-%3%-%4%") - % cacheParentDir - % _uname.machine - % _uname.sysname - % _uname.release).str(); + cacheDir = fmt("%1%/%2%-%3%-%4%", cacheParentDir, _uname.machine, _uname.sysname, _uname.release); mkdir(cacheParentDir.c_str(), 0755); mkdir(cacheDir.c_str(), 0755); diff --git a/tests/binary-cache.sh b/tests/binary-cache.sh index 0361ac6a8..7c64a115c 100644 --- a/tests/binary-cache.sh +++ b/tests/binary-cache.sh @@ -15,15 +15,15 @@ outPath=$(nix-build dependencies.nix --no-out-link) nix copy --to file://$cacheDir $outPath # Test copying build logs to the binary cache. -nix log --store file://$cacheDir $outPath 2>&1 | grep 'is not available' +expect 1 nix log --store file://$cacheDir $outPath 2>&1 | grep 'is not available' nix store copy-log --to file://$cacheDir $outPath nix log --store file://$cacheDir $outPath | grep FOO rm -rf $TEST_ROOT/var/log/nix -nix log $outPath 2>&1 | grep 'is not available' +expect 1 nix log $outPath 2>&1 | grep 'is not available' nix log --substituters file://$cacheDir $outPath | grep FOO # Test copying build logs from the binary cache. -nix store copy-log --from file://$cacheDir $(nix-store -qd $outPath) +nix store copy-log --from file://$cacheDir $(nix-store -qd $outPath)^'*' nix log $outPath | grep FOO basicDownloadTests() { @@ -78,8 +78,8 @@ mv $nar $nar.good mkdir -p $TEST_ROOT/empty nix-store --dump $TEST_ROOT/empty | xz > $nar -nix-build --substituters "file://$cacheDir" --no-require-sigs dependencies.nix -o $TEST_ROOT/result 2>&1 | tee $TEST_ROOT/log -grep -q "hash mismatch" $TEST_ROOT/log +expect 1 nix-build --substituters "file://$cacheDir" --no-require-sigs dependencies.nix -o $TEST_ROOT/result 2>&1 | tee $TEST_ROOT/log +grepQuiet "hash mismatch" $TEST_ROOT/log mv $nar.good $nar @@ -126,9 +126,9 @@ clearStore rm -v $(grep -l "StorePath:.*dependencies-input-2" $cacheDir/*.narinfo) nix-build --substituters "file://$cacheDir" --no-require-sigs dependencies.nix -o $TEST_ROOT/result 2>&1 | tee $TEST_ROOT/log -grep -q "copying path.*input-0" $TEST_ROOT/log -grep -q "copying path.*input-2" $TEST_ROOT/log -grep -q "copying path.*top" $TEST_ROOT/log +grepQuiet "copying path.*input-0" $TEST_ROOT/log +grepQuiet "copying path.*input-2" $TEST_ROOT/log +grepQuiet "copying path.*top" $TEST_ROOT/log # Idem, but without cached .narinfo. @@ -136,11 +136,11 @@ clearStore clearCacheCache nix-build --substituters "file://$cacheDir" --no-require-sigs dependencies.nix -o $TEST_ROOT/result 2>&1 | tee $TEST_ROOT/log -grep -q "don't know how to build" $TEST_ROOT/log -grep -q "building.*input-1" $TEST_ROOT/log -grep -q "building.*input-2" $TEST_ROOT/log -grep -q "copying path.*input-0" $TEST_ROOT/log -grep -q "copying path.*top" $TEST_ROOT/log +grepQuiet "don't know how to build" $TEST_ROOT/log +grepQuiet "building.*input-1" $TEST_ROOT/log +grepQuiet "building.*input-2" $TEST_ROOT/log +grepQuiet "copying path.*input-0" $TEST_ROOT/log +grepQuiet "copying path.*top" $TEST_ROOT/log # Create a signed binary cache. diff --git a/tests/build-delete.sh b/tests/build-delete.sh new file mode 100644 index 000000000..9c56b00e8 --- /dev/null +++ b/tests/build-delete.sh @@ -0,0 +1,54 @@ +source common.sh + +clearStore + +# https://github.com/NixOS/nix/issues/6572 +issue_6572_independent_outputs() { + nix build -f multiple-outputs.nix --json independent --no-link > $TEST_ROOT/independent.json + + # Make sure that 'nix build' can build a derivation that depends on both outputs of another derivation. + p=$(nix build -f multiple-outputs.nix use-independent --no-link --print-out-paths) + nix-store --delete "$p" # Clean up for next test + + # Make sure that 'nix build' tracks input-outputs correctly when a single output is already present. + nix-store --delete "$(jq -r <$TEST_ROOT/independent.json .[0].outputs.first)" + p=$(nix build -f multiple-outputs.nix use-independent --no-link --print-out-paths) + cmp $p <<EOF +first +second +EOF + nix-store --delete "$p" # Clean up for next test + + # Make sure that 'nix build' tracks input-outputs correctly when a single output is already present. + nix-store --delete "$(jq -r <$TEST_ROOT/independent.json .[0].outputs.second)" + p=$(nix build -f multiple-outputs.nix use-independent --no-link --print-out-paths) + cmp $p <<EOF +first +second +EOF + nix-store --delete "$p" # Clean up for next test +} +issue_6572_independent_outputs + + +# https://github.com/NixOS/nix/issues/6572 +issue_6572_dependent_outputs() { + + nix build -f multiple-outputs.nix --json a --no-link > $TEST_ROOT/a.json + + # # Make sure that 'nix build' can build a derivation that depends on both outputs of another derivation. + p=$(nix build -f multiple-outputs.nix use-a --no-link --print-out-paths) + nix-store --delete "$p" # Clean up for next test + + # Make sure that 'nix build' tracks input-outputs correctly when a single output is already present. + nix-store --delete "$(jq -r <$TEST_ROOT/a.json .[0].outputs.second)" + p=$(nix build -f multiple-outputs.nix use-a --no-link --print-out-paths) + cmp $p <<EOF +first +second +EOF + nix-store --delete "$p" # Clean up for next test +} +if isDaemonNewer "2.12pre0"; then + issue_6572_dependent_outputs +fi diff --git a/tests/build-dry.sh b/tests/build-dry.sh index f0f38e9a0..6d1754af5 100644 --- a/tests/build-dry.sh +++ b/tests/build-dry.sh @@ -18,9 +18,6 @@ nix-build --no-out-link dependencies.nix --dry-run 2>&1 | grep "will be built" # Now new command: nix build -f dependencies.nix --dry-run 2>&1 | grep "will be built" -# TODO: XXX: FIXME: #1793 -# Disable this part of the test until the problem is resolved: -if [ -n "$ISSUE_1795_IS_FIXED" ]; then clearStore clearCache @@ -28,7 +25,6 @@ clearCache nix build -f dependencies.nix --dry-run 2>&1 | grep "will be built" # Now old command: nix-build --no-out-link dependencies.nix --dry-run 2>&1 | grep "will be built" -fi ################################################### # Check --dry-run doesn't create links with --dry-run @@ -58,7 +54,7 @@ clearCache RES=$(nix build -f dependencies.nix --dry-run --json) -if [[ -z "$NIX_TESTS_CA_BY_DEFAULT" ]]; then +if [[ -z "${NIX_TESTS_CA_BY_DEFAULT-}" ]]; then echo "$RES" | jq '.[0] | [ (.drvPath | test("'$NIX_STORE_DIR'.*\\.drv")), (.outputs.out | test("'$NIX_STORE_DIR'")) diff --git a/tests/build-hook-ca-floating.nix b/tests/build-hook-ca-floating.nix index 67295985f..dfdbb82da 100644 --- a/tests/build-hook-ca-floating.nix +++ b/tests/build-hook-ca-floating.nix @@ -1,53 +1,6 @@ { busybox }: -with import ./config.nix; - -let - - mkDerivation = args: - derivation ({ - inherit system; - builder = busybox; - args = ["sh" "-e" args.builder or (builtins.toFile "builder-${args.name}.sh" "if [ -e .attrs.sh ]; then source .attrs.sh; fi; eval \"$buildCommand\"")]; - outputHashMode = "recursive"; - outputHashAlgo = "sha256"; - __contentAddressed = true; - } // removeAttrs args ["builder" "meta"]) - // { meta = args.meta or {}; }; - - input1 = mkDerivation { - shell = busybox; - name = "build-remote-input-1"; - buildCommand = "echo FOO > $out"; - requiredSystemFeatures = ["foo"]; - }; - - input2 = mkDerivation { - shell = busybox; - name = "build-remote-input-2"; - buildCommand = "echo BAR > $out"; - requiredSystemFeatures = ["bar"]; - }; - - input3 = mkDerivation { - shell = busybox; - name = "build-remote-input-3"; - buildCommand = '' - read x < ${input2} - echo $x BAZ > $out - ''; - requiredSystemFeatures = ["baz"]; - }; - -in - - mkDerivation { - shell = busybox; - name = "build-remote"; - buildCommand = - '' - read x < ${input1} - read y < ${input3} - echo "$x $y" > $out - ''; - } +import ./build-hook.nix { + inherit busybox; + contentAddressed = true; +} diff --git a/tests/build-hook.nix b/tests/build-hook.nix index 643334caa..7effd7903 100644 --- a/tests/build-hook.nix +++ b/tests/build-hook.nix @@ -1,15 +1,22 @@ -{ busybox }: +{ busybox, contentAddressed ? false }: with import ./config.nix; let + caArgs = if contentAddressed then { + outputHashMode = "recursive"; + outputHashAlgo = "sha256"; + __contentAddressed = true; + } else {}; + mkDerivation = args: derivation ({ inherit system; builder = busybox; args = ["sh" "-e" args.builder or (builtins.toFile "builder-${args.name}.sh" "if [ -e .attrs.sh ]; then source .attrs.sh; fi; eval \"$buildCommand\"")]; - } // removeAttrs args ["builder" "meta" "passthru"]) + } // removeAttrs args ["builder" "meta" "passthru"] + // caArgs) // { meta = args.meta or {}; passthru = args.passthru or {}; }; input1 = mkDerivation { diff --git a/tests/build-remote.sh b/tests/build-remote.sh index e73c37ea4..78e12b477 100644 --- a/tests/build-remote.sh +++ b/tests/build-remote.sh @@ -1,5 +1,5 @@ -if ! canUseSandbox; then exit 99; fi -if ! [[ $busybox =~ busybox ]]; then exit 99; fi +requireSandboxSupport +[[ $busybox =~ busybox ]] || skipTest "no busybox" unset NIX_STORE_DIR unset NIX_STATE_DIR @@ -7,7 +7,7 @@ unset NIX_STATE_DIR function join_by { local d=$1; shift; echo -n "$1"; shift; printf "%s" "${@/#/$d}"; } EXTRA_SYSTEM_FEATURES=() -if [[ -n "$CONTENT_ADDRESSED" ]]; then +if [[ -n "${CONTENT_ADDRESSED-}" ]]; then EXTRA_SYSTEM_FEATURES=("ca-derivations") fi @@ -42,33 +42,31 @@ testPrintOutPath=$(nix build -L -v -f $file --no-link --print-out-paths --max-jo [[ $testPrintOutPath =~ store.*build-remote ]] -set -o pipefail - # Ensure that input1 was built on store1 due to the required feature. -nix path-info --store $TEST_ROOT/machine1 --all \ - | grep builder-build-remote-input-1.sh \ - | grep -v builder-build-remote-input-2.sh \ - | grep -v builder-build-remote-input-3.sh +output=$(nix path-info --store $TEST_ROOT/machine1 --all) +echo "$output" | grepQuiet builder-build-remote-input-1.sh +echo "$output" | grepQuietInverse builder-build-remote-input-2.sh +echo "$output" | grepQuietInverse builder-build-remote-input-3.sh +unset output # Ensure that input2 was built on store2 due to the required feature. -nix path-info --store $TEST_ROOT/machine2 --all \ - | grep -v builder-build-remote-input-1.sh \ - | grep builder-build-remote-input-2.sh \ - | grep -v builder-build-remote-input-3.sh +output=$(nix path-info --store $TEST_ROOT/machine2 --all) +echo "$output" | grepQuietInverse builder-build-remote-input-1.sh +echo "$output" | grepQuiet builder-build-remote-input-2.sh +echo "$output" | grepQuietInverse builder-build-remote-input-3.sh +unset output # Ensure that input3 was built on store3 due to the required feature. -nix path-info --store $TEST_ROOT/machine3 --all \ - | grep -v builder-build-remote-input-1.sh \ - | grep -v builder-build-remote-input-2.sh \ - | grep builder-build-remote-input-3.sh +output=$(nix path-info --store $TEST_ROOT/machine3 --all) +echo "$output" | grepQuietInverse builder-build-remote-input-1.sh +echo "$output" | grepQuietInverse builder-build-remote-input-2.sh +echo "$output" | grepQuiet builder-build-remote-input-3.sh +unset output -# Temporarily disabled because of https://github.com/NixOS/nix/issues/6209 -if [[ -z "$CONTENT_ADDRESSED" ]]; then - for i in input1 input3; do - nix log --store $TEST_ROOT/machine0 --file "$file" --arg busybox $busybox passthru."$i" | grep hi-$i - done -fi +for i in input1 input3; do +nix log --store $TEST_ROOT/machine0 --file "$file" --arg busybox $busybox passthru."$i" | grep hi-$i +done # Behavior of keep-failed out="$(nix-build 2>&1 failing.nix \ diff --git a/tests/build.sh b/tests/build.sh index fc6825e25..b579fc374 100644 --- a/tests/build.sh +++ b/tests/build.sh @@ -2,19 +2,19 @@ source common.sh clearStore -set -o pipefail - # Make sure that 'nix build' returns all outputs by default. nix build -f multiple-outputs.nix --json a b --no-link | jq --exit-status ' (.[0] | (.drvPath | match(".*multiple-outputs-a.drv")) and - (.outputs | keys | length == 2) and - (.outputs.first | match(".*multiple-outputs-a-first")) and - (.outputs.second | match(".*multiple-outputs-a-second"))) + (.outputs | + (keys | length == 2) and + (.first | match(".*multiple-outputs-a-first")) and + (.second | match(".*multiple-outputs-a-second")))) and (.[1] | (.drvPath | match(".*multiple-outputs-b.drv")) and - (.outputs | keys | length == 1) and - (.outputs.out | match(".*multiple-outputs-b"))) + (.outputs | + (keys | length == 1) and + (.out | match(".*multiple-outputs-b")))) ' # Test output selection using the '^' syntax. @@ -40,33 +40,68 @@ nix build -f multiple-outputs.nix --json 'a^*' --no-link | jq --exit-status ' nix build -f multiple-outputs.nix --json e --no-link | jq --exit-status ' (.[0] | (.drvPath | match(".*multiple-outputs-e.drv")) and - (.outputs | keys == ["a", "b"])) + (.outputs | keys == ["a_a", "b"])) ' # But not when it's overriden. -nix build -f multiple-outputs.nix --json e^a --no-link | jq --exit-status ' +nix build -f multiple-outputs.nix --json e^a_a --no-link +nix build -f multiple-outputs.nix --json e^a_a --no-link | jq --exit-status ' (.[0] | (.drvPath | match(".*multiple-outputs-e.drv")) and - (.outputs | keys == ["a"])) + (.outputs | keys == ["a_a"])) ' nix build -f multiple-outputs.nix --json 'e^*' --no-link | jq --exit-status ' (.[0] | (.drvPath | match(".*multiple-outputs-e.drv")) and - (.outputs | keys == ["a", "b", "c"])) + (.outputs | keys == ["a_a", "b", "c"])) +' + +# Test building from raw store path to drv not expression. + +drv=$(nix eval -f multiple-outputs.nix --raw a.drvPath) +if nix build "$drv^not-an-output" --no-link --json; then + fail "'not-an-output' should fail to build" +fi + +if nix build "$drv^" --no-link --json; then + fail "'empty outputs list' should fail to build" +fi + +if nix build "$drv^*nope" --no-link --json; then + fail "'* must be entire string' should fail to build" +fi + +nix build "$drv^first" --no-link --json | jq --exit-status ' + (.[0] | + (.drvPath | match(".*multiple-outputs-a.drv")) and + (.outputs | + (keys | length == 1) and + (.first | match(".*multiple-outputs-a-first")) and + (has("second") | not))) +' + +nix build "$drv^first,second" --no-link --json | jq --exit-status ' + (.[0] | + (.drvPath | match(".*multiple-outputs-a.drv")) and + (.outputs | + (keys | length == 2) and + (.first | match(".*multiple-outputs-a-first")) and + (.second | match(".*multiple-outputs-a-second")))) +' + +nix build "$drv^*" --no-link --json | jq --exit-status ' + (.[0] | + (.drvPath | match(".*multiple-outputs-a.drv")) and + (.outputs | + (keys | length == 2) and + (.first | match(".*multiple-outputs-a-first")) and + (.second | match(".*multiple-outputs-a-second")))) ' # Make sure that `--impure` works (regression test for https://github.com/NixOS/nix/issues/6488) nix build --impure -f multiple-outputs.nix --json e --no-link | jq --exit-status ' (.[0] | (.drvPath | match(".*multiple-outputs-e.drv")) and - (.outputs | keys == ["a", "b"])) + (.outputs | keys == ["a_a", "b"])) ' - -testNormalization () { - clearStore - outPath=$(nix-build ./simple.nix --no-out-link) - test "$(stat -c %Y $outPath)" -eq 1 -} - -testNormalization diff --git a/tests/ca/build.sh b/tests/ca/build.sh index 92f8b429a..98e1c5125 100644 --- a/tests/ca/build.sh +++ b/tests/ca/build.sh @@ -2,14 +2,14 @@ source common.sh -drv=$(nix-instantiate --experimental-features ca-derivations ./content-addressed.nix -A rootCA --arg seed 1) -nix --experimental-features 'nix-command ca-derivations' show-derivation --derivation "$drv" --arg seed 1 +drv=$(nix-instantiate ./content-addressed.nix -A rootCA --arg seed 1) +nix show-derivation "$drv" --arg seed 1 buildAttr () { local derivationPath=$1 local seedValue=$2 shift; shift - local args=("--experimental-features" "ca-derivations" "./content-addressed.nix" "-A" "$derivationPath" --arg seed "$seedValue" "--no-out-link") + local args=("./content-addressed.nix" "-A" "$derivationPath" --arg seed "$seedValue" "--no-out-link") args+=("$@") nix-build "${args[@]}" } @@ -19,7 +19,7 @@ testRemoteCache () { local outPath=$(buildAttr dependentNonCA 1) nix copy --to file://$cacheDir $outPath clearStore - buildAttr dependentNonCA 1 --option substituters file://$cacheDir --no-require-sigs |& (! grep "building dependent-non-ca") + buildAttr dependentNonCA 1 --option substituters file://$cacheDir --no-require-sigs |& grepQuietInverse "building dependent-non-ca" } testDeterministicCA () { @@ -46,17 +46,17 @@ testCutoff () { } testGC () { - nix-instantiate --experimental-features ca-derivations ./content-addressed.nix -A rootCA --arg seed 5 - nix-collect-garbage --experimental-features ca-derivations --option keep-derivations true + nix-instantiate ./content-addressed.nix -A rootCA --arg seed 5 + nix-collect-garbage --option keep-derivations true clearStore buildAttr rootCA 1 --out-link $TEST_ROOT/rootCA - nix-collect-garbage --experimental-features ca-derivations + nix-collect-garbage buildAttr rootCA 1 -j0 } testNixCommand () { clearStore - nix build --experimental-features 'nix-command ca-derivations' --file ./content-addressed.nix --no-link + nix build --file ./content-addressed.nix --no-link } # Regression test for https://github.com/NixOS/nix/issues/4775 diff --git a/tests/ca/new-build-cmd.sh b/tests/ca/new-build-cmd.sh new file mode 100644 index 000000000..432d4d132 --- /dev/null +++ b/tests/ca/new-build-cmd.sh @@ -0,0 +1,5 @@ +source common.sh + +export NIX_TESTS_CA_BY_DEFAULT=1 +cd .. +source ./build.sh diff --git a/tests/ca/recursive.sh b/tests/ca/recursive.sh index 0354d23b4..cd6736b24 100755 --- a/tests/ca/recursive.sh +++ b/tests/ca/recursive.sh @@ -7,5 +7,3 @@ requireDaemonNewerThan "2.4pre20210623" export NIX_TESTS_CA_BY_DEFAULT=1 cd .. source ./recursive.sh - - diff --git a/tests/ca/substitute.sh b/tests/ca/substitute.sh index 819f3fd85..ea981adc4 100644 --- a/tests/ca/substitute.sh +++ b/tests/ca/substitute.sh @@ -28,6 +28,12 @@ nix realisation info --file ./content-addressed.nix transitivelyDependentCA nix realisation info --file ./content-addressed.nix dependentCA # nix realisation info --file ./content-addressed.nix rootCA --outputs out +if isDaemonNewer "2.13"; then + pushToStore="../push-to-store.sh" +else + pushToStore="../push-to-store-old.sh" +fi + # Same thing, but # 1. With non-ca derivations # 2. Erasing the realisations on the remote store @@ -37,7 +43,7 @@ nix realisation info --file ./content-addressed.nix dependentCA # # Regression test for #4725 clearStore -nix build --file ../simple.nix -L --no-link --post-build-hook ../push-to-store.sh +nix build --file ../simple.nix -L --no-link --post-build-hook "$pushToStore" clearStore rm -r "$REMOTE_STORE_DIR/realisations" nix build --file ../simple.nix -L --no-link --substitute --substituters "$REMOTE_STORE" --no-require-sigs -j0 @@ -52,7 +58,7 @@ if [[ -z "$(ls "$REMOTE_STORE_DIR/realisations")" ]]; then fi # Test the local realisation disk cache -buildDrvs --post-build-hook ../push-to-store.sh +buildDrvs --post-build-hook "$pushToStore" clearStore # Add the realisations of rootCA to the cachecache clearCacheCache diff --git a/tests/ca/why-depends.sh b/tests/ca/why-depends.sh new file mode 100644 index 000000000..0c079f63b --- /dev/null +++ b/tests/ca/why-depends.sh @@ -0,0 +1,5 @@ +source common.sh + +export NIX_TESTS_CA_BY_DEFAULT=1 + +cd .. && source why-depends.sh diff --git a/tests/check-refs.nix b/tests/check-refs.nix index 9d90b0920..99d69a226 100644 --- a/tests/check-refs.nix +++ b/tests/check-refs.nix @@ -67,4 +67,11 @@ rec { disallowedReferences = [test5]; }; + test11 = makeTest 11 { + __structuredAttrs = true; + unsafeDiscardReferences.out = true; + outputChecks.out.allowedReferences = []; + buildCommand = ''echo ${dep} > "''${outputs[out]}"''; + }; + } diff --git a/tests/check-refs.sh b/tests/check-refs.sh index 16bbabc40..2778e491d 100644 --- a/tests/check-refs.sh +++ b/tests/check-refs.sh @@ -8,14 +8,14 @@ dep=$(nix-build -o $RESULT check-refs.nix -A dep) # test1 references dep, not itself. test1=$(nix-build -o $RESULT check-refs.nix -A test1) -(! nix-store -q --references $test1 | grep -q $test1) -nix-store -q --references $test1 | grep -q $dep +nix-store -q --references $test1 | grepQuietInverse $test1 +nix-store -q --references $test1 | grepQuiet $dep # test2 references src, not itself nor dep. test2=$(nix-build -o $RESULT check-refs.nix -A test2) -(! nix-store -q --references $test2 | grep -q $test2) -(! nix-store -q --references $test2 | grep -q $dep) -nix-store -q --references $test2 | grep -q aux-ref +nix-store -q --references $test2 | grepQuietInverse $test2 +nix-store -q --references $test2 | grepQuietInverse $dep +nix-store -q --references $test2 | grepQuiet aux-ref # test3 should fail (unallowed ref). (! nix-build -o $RESULT check-refs.nix -A test3) @@ -40,3 +40,12 @@ nix-build -o $RESULT check-refs.nix -A test7 # test10 should succeed (no disallowed references). nix-build -o $RESULT check-refs.nix -A test10 + +if isDaemonNewer 2.12pre20230103; then + enableFeatures discard-references + restartDaemon + + # test11 should succeed. + test11=$(nix-build -o $RESULT check-refs.nix -A test11) + [[ -z $(nix-store -q --references "$test11") ]] +fi diff --git a/tests/check-reqs.sh b/tests/check-reqs.sh index e9f65fc2a..856c94cec 100644 --- a/tests/check-reqs.sh +++ b/tests/check-reqs.sh @@ -8,8 +8,8 @@ nix-build -o $RESULT check-reqs.nix -A test1 (! nix-build -o $RESULT check-reqs.nix -A test2) (! nix-build -o $RESULT check-reqs.nix -A test3) -(! nix-build -o $RESULT check-reqs.nix -A test4) 2>&1 | grep -q 'check-reqs-dep1' -(! nix-build -o $RESULT check-reqs.nix -A test4) 2>&1 | grep -q 'check-reqs-dep2' +(! nix-build -o $RESULT check-reqs.nix -A test4) 2>&1 | grepQuiet 'check-reqs-dep1' +(! nix-build -o $RESULT check-reqs.nix -A test4) 2>&1 | grepQuiet 'check-reqs-dep2' (! nix-build -o $RESULT check-reqs.nix -A test5) (! nix-build -o $RESULT check-reqs.nix -A test6) diff --git a/tests/check.nix b/tests/check.nix index ed91ff845..ddab8eea9 100644 --- a/tests/check.nix +++ b/tests/check.nix @@ -44,7 +44,7 @@ with import ./config.nix; }; hashmismatch = import <nix/fetchurl.nix> { - url = "file://" + builtins.getEnv "TMPDIR" + "/dummy"; + url = "file://" + builtins.getEnv "TEST_ROOT" + "/dummy"; sha256 = "0mdqa9w1p6cmli6976v4wi0sw9r4p5prkj7lzfd1877wk11c9c73"; }; diff --git a/tests/check.sh b/tests/check.sh index ab48ff865..645b90222 100644 --- a/tests/check.sh +++ b/tests/check.sh @@ -37,7 +37,7 @@ checkBuildTempDirRemoved $TEST_ROOT/log nix-build check.nix -A deterministic --argstr checkBuildId $checkBuildId \ --no-out-link --check --keep-failed 2> $TEST_ROOT/log -if grep -q 'may not be deterministic' $TEST_ROOT/log; then false; fi +if grepQuiet 'may not be deterministic' $TEST_ROOT/log; then false; fi checkBuildTempDirRemoved $TEST_ROOT/log nix-build check.nix -A nondeterministic --argstr checkBuildId $checkBuildId \ @@ -58,12 +58,6 @@ if checkBuildTempDirRemoved $TEST_ROOT/log; then false; fi clearStore -nix-build dependencies.nix --no-out-link --repeat 3 - -nix-build check.nix -A nondeterministic --no-out-link --repeat 1 2> $TEST_ROOT/log || status=$? -[ "$status" = "1" ] -grep 'differs from previous round' $TEST_ROOT/log - path=$(nix-build check.nix -A fetchurl --no-out-link) chmod +w $path @@ -77,13 +71,13 @@ nix-build check.nix -A fetchurl --no-out-link --check nix-build check.nix -A fetchurl --no-out-link --repair [[ $(cat $path) != foo ]] -echo 'Hello World' > $TMPDIR/dummy +echo 'Hello World' > $TEST_ROOT/dummy nix-build check.nix -A hashmismatch --no-out-link || status=$? [ "$status" = "102" ] -echo -n > $TMPDIR/dummy +echo -n > $TEST_ROOT/dummy nix-build check.nix -A hashmismatch --no-out-link -echo 'Hello World' > $TMPDIR/dummy +echo 'Hello World' > $TEST_ROOT/dummy nix-build check.nix -A hashmismatch --no-out-link --check || status=$? [ "$status" = "102" ] diff --git a/tests/common.sh b/tests/common.sh new file mode 100644 index 000000000..8941671d6 --- /dev/null +++ b/tests/common.sh @@ -0,0 +1,12 @@ +set -eu -o pipefail + +if [[ -z "${COMMON_SH_SOURCED-}" ]]; then + +COMMON_SH_SOURCED=1 + +source "$(readlink -f "$(dirname "${BASH_SOURCE[0]}")")/common/vars-and-functions.sh" +if [[ -n "${NIX_DAEMON_PACKAGE:-}" ]]; then + startDaemon +fi + +fi # COMMON_SH_SOURCED diff --git a/tests/common.sh.in b/tests/common/vars-and-functions.sh.in index 79da10199..a9e6c802f 100644 --- a/tests/common.sh.in +++ b/tests/common/vars-and-functions.sh.in @@ -1,8 +1,10 @@ -set -e +set -eu -o pipefail -if [[ -z "$COMMON_SH_SOURCED" ]]; then +if [[ -z "${COMMON_VARS_AND_FUNCTIONS_SH_SOURCED-}" ]]; then -COMMON_SH_SOURCED=1 +COMMON_VARS_AND_FUNCTIONS_SH_SOURCED=1 + +export PS4='+(${BASH_SOURCE[0]}:$LINENO) ' export TEST_ROOT=$(realpath ${TMPDIR:-/tmp}/nix-test)/${TEST_NAME:-default} export NIX_STORE_DIR @@ -23,10 +25,12 @@ if [[ -n $NIX_STORE ]]; then fi export _NIX_IN_TEST=$TEST_ROOT/shared export _NIX_TEST_NO_LSOF=1 -export NIX_REMOTE=$NIX_REMOTE_ +export NIX_REMOTE=${NIX_REMOTE_-} unset NIX_PATH export TEST_HOME=$TEST_ROOT/test-home export HOME=$TEST_HOME +unset XDG_STATE_HOME +unset XDG_DATA_HOME unset XDG_CONFIG_HOME unset XDG_CONFIG_DIRS unset XDG_CACHE_HOME @@ -62,8 +66,8 @@ readLink() { } clearProfiles() { - profiles="$NIX_STATE_DIR"/profiles - rm -rf $profiles + profiles="$HOME"/.local/state/nix/profiles + rm -rf "$profiles" } clearStore() { @@ -86,13 +90,14 @@ clearCacheCache() { startDaemon() { # Don’t start the daemon twice, as this would just make it loop indefinitely - if [[ "$NIX_REMOTE" == daemon ]]; then - return + if [[ "${_NIX_TEST_DAEMON_PID-}" != '' ]]; then + return fi # Start the daemon, wait for the socket to appear. rm -f $NIX_DAEMON_SOCKET_PATH - PATH=$DAEMON_PATH nix-daemon& - pidDaemon=$! + PATH=$DAEMON_PATH nix-daemon & + _NIX_TEST_DAEMON_PID=$! + export _NIX_TEST_DAEMON_PID for ((i = 0; i < 300; i++)); do if [[ -S $NIX_DAEMON_SOCKET_PATH ]]; then DAEMON_STARTED=1 @@ -104,25 +109,35 @@ startDaemon() { fail "Didn’t manage to start the daemon" fi trap "killDaemon" EXIT + # Save for if daemon is killed + NIX_REMOTE_OLD=$NIX_REMOTE export NIX_REMOTE=daemon } killDaemon() { - kill $pidDaemon + # Don’t fail trying to stop a non-existant daemon twice + if [[ "${_NIX_TEST_DAEMON_PID-}" == '' ]]; then + return + fi + kill $_NIX_TEST_DAEMON_PID for i in {0..100}; do - kill -0 $pidDaemon 2> /dev/null || break + kill -0 $_NIX_TEST_DAEMON_PID 2> /dev/null || break sleep 0.1 done - kill -9 $pidDaemon 2> /dev/null || true - wait $pidDaemon || true + kill -9 $_NIX_TEST_DAEMON_PID 2> /dev/null || true + wait $_NIX_TEST_DAEMON_PID || true + rm -f $NIX_DAEMON_SOCKET_PATH + # Indicate daemon is stopped + unset _NIX_TEST_DAEMON_PID + # Restore old nix remote + NIX_REMOTE=$NIX_REMOTE_OLD trap "" EXIT } restartDaemon() { - [[ -z "${pidDaemon:-}" ]] && return 0 + [[ -z "${_NIX_TEST_DAEMON_PID:-}" ]] && return 0 killDaemon - unset NIX_REMOTE startDaemon } @@ -137,31 +152,64 @@ isDaemonNewer () { [[ $(nix eval --expr "builtins.compareVersions ''$daemonVersion'' ''$requiredVersion''") -ge 0 ]] } +skipTest () { + echo "$1, skipping this test..." >&2 + exit 99 +} + requireDaemonNewerThan () { - isDaemonNewer "$1" || exit 99 + isDaemonNewer "$1" || skipTest "Daemon is too old" } canUseSandbox() { - if [[ ! $_canUseSandbox ]]; then - echo "Sandboxing not supported, skipping this test..." - return 1 - fi + [[ ${_canUseSandbox-} ]] +} - return 0 +requireSandboxSupport () { + canUseSandbox || skipTest "Sandboxing not supported" +} + +requireGit() { + [[ $(type -p git) ]] || skipTest "Git not installed" } fail() { - echo "$1" + echo "$1" >&2 exit 1 } +# Run a command failing if it didn't exit with the expected exit code. +# +# Has two advantages over the built-in `!`: +# +# 1. `!` conflates all non-0 codes. `expect` allows testing for an exact +# code. +# +# 2. `!` unexpectedly negates `set -e`, and cannot be used on individual +# pipeline stages with `set -o pipefail`. It only works on the entire +# pipeline, which is useless if we want, say, `nix ...` invocation to +# *fail*, but a grep on the error message it outputs to *succeed*. expect() { local expected res expected="$1" shift - "$@" || res="$?" + "$@" && res=0 || res="$?" + if [[ $res -ne $expected ]]; then + echo "Expected '$expected' but got '$res' while running '${*@Q}'" >&2 + return 1 + fi + return 0 +} + +# Better than just doing `expect ... >&2` because the "Expected..." +# message below will *not* be redirected. +expectStderr() { + local expected res + expected="$1" + shift + "$@" 2>&1 && res=0 || res="$?" if [[ $res -ne $expected ]]; then - echo "Expected '$expected' but got '$res' while running '$*'" + echo "Expected '$expected' but got '$res' while running '${*@Q}'" >&2 return 1 fi return 0 @@ -169,14 +217,13 @@ expect() { needLocalStore() { if [[ "$NIX_REMOTE" == "daemon" ]]; then - echo "Can’t run through the daemon ($1), skipping this test..." - return 99 + skipTest "Can’t run through the daemon ($1)" fi } # Just to make it easy to find which tests should be fixed buggyNeedLocalStore() { - needLocalStore + needLocalStore "$1" } enableFeatures() { @@ -186,19 +233,44 @@ enableFeatures() { set -x -if [[ -n "${NIX_DAEMON_PACKAGE:-}" ]]; then - startDaemon -fi - onError() { set +x echo "$0: test failed at:" >&2 - for ((i = 1; i < 16; i++)); do + for ((i = 1; i < ${#BASH_SOURCE[@]}; i++)); do if [[ -z ${BASH_SOURCE[i]} ]]; then break; fi echo " ${FUNCNAME[i]} in ${BASH_SOURCE[i]}:${BASH_LINENO[i-1]}" >&2 done } +# `grep -v` doesn't work well for exit codes. We want `!(exist line l. l +# matches)`. It gives us `exist line l. !(l matches)`. +# +# `!` normally doesn't work well with `set -e`, but when we wrap in a +# function it *does*. +grepInverse() { + ! grep "$@" +} + +# A shorthand, `> /dev/null` is a bit noisy. +# +# `grep -q` would seem to do this, no function necessary, but it is a +# bad fit with pipes and `set -o pipefail`: `-q` will exit after the +# first match, and then subsequent writes will result in broken pipes. +# +# Note that reproducing the above is a bit tricky as it depends on +# non-deterministic properties such as the timing between the match and +# the closing of the pipe, the buffering of the pipe, and the speed of +# the producer into the pipe. But rest assured we've seen it happen in +# CI reliably. +grepQuiet() { + grep "$@" > /dev/null +} + +# The previous two, combined +grepQuietInverse() { + ! grep "$@" > /dev/null +} + trap onError ERR -fi # COMMON_SH_SOURCED +fi # COMMON_VARS_AND_FUNCTIONS_SH_SOURCED diff --git a/tests/completions.sh b/tests/completions.sh index 522aa1c86..19dc61098 100644 --- a/tests/completions.sh +++ b/tests/completions.sh @@ -28,6 +28,10 @@ cat <<EOF > bar/flake.nix }; } EOF +mkdir -p err +cat <<EOF > err/flake.nix +throw "error" +EOF # Test the completion of a subcommand [[ "$(NIX_GET_COMPLETIONS=1 nix buil)" == $'normal\nbuild\t' ]] @@ -60,3 +64,5 @@ NIX_GET_COMPLETIONS=3 nix build --option allow-import-from | grep -- "allow-impo # Attr path completions [[ "$(NIX_GET_COMPLETIONS=2 nix eval ./foo\#sam)" == $'attrs\n./foo#sampleOutput\t' ]] [[ "$(NIX_GET_COMPLETIONS=4 nix eval --file ./foo/flake.nix outp)" == $'attrs\noutputs\t' ]] +[[ "$(NIX_GET_COMPLETIONS=4 nix eval --file ./err/flake.nix outp 2>&1)" == $'attrs' ]] +[[ "$(NIX_GET_COMPLETIONS=2 nix eval ./err\# 2>&1)" == $'attrs' ]] diff --git a/tests/compute-levels.sh b/tests/compute-levels.sh index e4322dfa1..de3da2ebd 100644 --- a/tests/compute-levels.sh +++ b/tests/compute-levels.sh @@ -3,5 +3,5 @@ source common.sh if [[ $(uname -ms) = "Linux x86_64" ]]; then # x86_64 CPUs must always support the baseline # microarchitecture level. - nix -vv --version | grep -q "x86_64-v1-linux" + nix -vv --version | grepQuiet "x86_64-v1-linux" fi diff --git a/tests/config.sh b/tests/config.sh index 3d0da3cef..723f575ed 100644 --- a/tests/config.sh +++ b/tests/config.sh @@ -51,3 +51,8 @@ exp_features=$(nix show-config | grep '^experimental-features' | cut -d '=' -f 2 [[ $prev != $exp_cores ]] [[ $exp_cores == "4242" ]] [[ $exp_features == "flakes nix-command" ]] + +# Test that it's possible to retrieve a single setting's value +val=$(nix show-config | grep '^warn-dirty' | cut -d '=' -f 2 | xargs) +val2=$(nix show-config warn-dirty) +[[ $val == $val2 ]] diff --git a/tests/db-migration.sh b/tests/db-migration.sh index 3f9dc8972..44cd16bc0 100644 --- a/tests/db-migration.sh +++ b/tests/db-migration.sh @@ -1,19 +1,18 @@ # Test that we can successfully migrate from an older db schema +source common.sh + # Only run this if we have an older Nix available # XXX: This assumes that the `daemon` package is older than the `client` one -if [[ -z "$NIX_DAEMON_PACKAGE" ]]; then - exit 99 +if [[ -z "${NIX_DAEMON_PACKAGE-}" ]]; then + skipTest "not using the Nix daemon" fi -source common.sh - killDaemon -unset NIX_REMOTE # Fill the db using the older Nix PATH_WITH_NEW_NIX="$PATH" -export PATH="$NIX_DAEMON_PACKAGE/bin:$PATH" +export PATH="${NIX_DAEMON_PACKAGE}/bin:$PATH" clearStore nix-build simple.nix --no-out-link nix-store --generate-binary-cache-key cache1.example.org $TEST_ROOT/sk1 $TEST_ROOT/pk1 diff --git a/tests/dependencies.sh b/tests/dependencies.sh index 092950aa7..f9da0c6bc 100644 --- a/tests/dependencies.sh +++ b/tests/dependencies.sh @@ -36,10 +36,10 @@ deps=$(nix-store -quR "$drvPath") echo "output closure contains $deps" # The output path should be in the closure. -echo "$deps" | grep -q "$outPath" +echo "$deps" | grepQuiet "$outPath" # Input-1 is not retained. -if echo "$deps" | grep -q "dependencies-input-1"; then exit 1; fi +if echo "$deps" | grepQuiet "dependencies-input-1"; then exit 1; fi # Input-2 is retained. input2OutPath=$(echo "$deps" | grep "dependencies-input-2") @@ -49,4 +49,4 @@ nix-store -q --referrers-closure "$input2OutPath" | grep "$outPath" # Check that the derivers are set properly. test $(nix-store -q --deriver "$outPath") = "$drvPath" -nix-store -q --deriver "$input2OutPath" | grep -q -- "-input-2.drv" +nix-store -q --deriver "$input2OutPath" | grepQuiet -- "-input-2.drv" diff --git a/tests/eval-store.sh b/tests/eval-store.sh index 679da5741..8fc859730 100644 --- a/tests/eval-store.sh +++ b/tests/eval-store.sh @@ -2,7 +2,7 @@ source common.sh # Using `--eval-store` with the daemon will eventually copy everything # to the build store, invalidating most of the tests here -needLocalStore +needLocalStore "“--eval-store” doesn't achieve much with the daemon" eval_store=$TEST_ROOT/eval-store diff --git a/tests/eval.sh b/tests/eval.sh index d74976019..ffae08a6a 100644 --- a/tests/eval.sh +++ b/tests/eval.sh @@ -29,3 +29,7 @@ nix-instantiate --eval -E 'assert 1 + 2 == 3; true' [[ $(nix-instantiate -A attr --eval "./eval.nix") == '{ foo = "bar"; }' ]] [[ $(nix-instantiate -A attr --eval --json "./eval.nix") == '{"foo":"bar"}' ]] [[ $(nix-instantiate -A int --eval - < "./eval.nix") == 123 ]] + +# Check that symlink cycles don't cause a hang. +ln -sfn cycle.nix $TEST_ROOT/cycle.nix +(! nix eval --file $TEST_ROOT/cycle.nix) diff --git a/tests/export-graph.sh b/tests/export-graph.sh index a1449b34e..1f6232a40 100644 --- a/tests/export-graph.sh +++ b/tests/export-graph.sh @@ -4,7 +4,7 @@ clearStore clearProfiles checkRef() { - nix-store -q --references $TEST_ROOT/result | grep -q "$1" || fail "missing reference $1" + nix-store -q --references $TEST_ROOT/result | grepQuiet "$1"'$' || fail "missing reference $1" } # Test the export of the runtime dependency graph. diff --git a/tests/fetchClosure.sh b/tests/fetchClosure.sh index 44050c878..a207f647c 100644 --- a/tests/fetchClosure.sh +++ b/tests/fetchClosure.sh @@ -1,7 +1,6 @@ source common.sh enableFeatures "fetch-closure" -needLocalStore "'--no-require-sigs' can’t be used with the daemon" clearStore clearCacheCache @@ -28,24 +27,28 @@ clearStore [ ! -e $nonCaPath ] [ -e $caPath ] -# In impure mode, we can use non-CA paths. -[[ $(nix eval --raw --no-require-sigs --impure --expr " - builtins.fetchClosure { - fromStore = \"file://$cacheDir\"; - fromPath = $nonCaPath; - } -") = $nonCaPath ]] +if [[ "$NIX_REMOTE" != "daemon" ]]; then + + # In impure mode, we can use non-CA paths. + [[ $(nix eval --raw --no-require-sigs --impure --expr " + builtins.fetchClosure { + fromStore = \"file://$cacheDir\"; + fromPath = $nonCaPath; + } + ") = $nonCaPath ]] + + [ -e $nonCaPath ] -[ -e $nonCaPath ] +fi # 'toPath' set to empty string should fail but print the expected path. -nix eval -v --json --expr " +expectStderr 1 nix eval -v --json --expr " builtins.fetchClosure { fromStore = \"file://$cacheDir\"; fromPath = $nonCaPath; toPath = \"\"; } -" 2>&1 | grep "error: rewriting.*$nonCaPath.*yielded.*$caPath" +" | grep "error: rewriting.*$nonCaPath.*yielded.*$caPath" # If fromPath is CA, then toPath isn't needed. nix copy --to file://$cacheDir $caPath diff --git a/tests/fetchGit.sh b/tests/fetchGit.sh index 166bccfc7..e2ccb0e97 100644 --- a/tests/fetchGit.sh +++ b/tests/fetchGit.sh @@ -1,9 +1,6 @@ source common.sh -if [[ -z $(type -p git) ]]; then - echo "Git not installed; skipping Git tests" - exit 99 -fi +requireGit clearStore @@ -24,12 +21,14 @@ touch $repo/.gitignore git -C $repo add hello .gitignore git -C $repo commit -m 'Bla1' rev1=$(git -C $repo rev-parse HEAD) +git -C $repo tag -a tag1 -m tag1 echo world > $repo/hello git -C $repo commit -m 'Bla2' -a git -C $repo worktree add $TEST_ROOT/worktree echo hello >> $TEST_ROOT/worktree/hello rev2=$(git -C $repo rev-parse HEAD) +git -C $repo tag -a tag2 -m tag2 # Fetch a worktree unset _NIX_FORCE_HTTP @@ -120,6 +119,7 @@ git -C $repo commit -m 'Bla3' -a path4=$(nix eval --impure --refresh --raw --expr "(builtins.fetchGit file://$repo).outPath") [[ $path2 = $path4 ]] +status=0 nix eval --impure --raw --expr "(builtins.fetchGit { url = $repo; rev = \"$rev2\"; narHash = \"sha256-B5yIPHhEm0eysJKEsO7nqxprh9vcblFxpJG11gXJus1=\"; }).outPath" || status=$? [[ "$status" = "102" ]] @@ -217,6 +217,16 @@ rev4_nix=$(nix eval --impure --raw --expr "(builtins.fetchGit { url = \"file://$ path9=$(nix eval --impure --raw --expr "(builtins.fetchGit { url = \"file://$repo\"; ref = \"HEAD\"; name = \"foo\"; }).outPath") [[ $path9 =~ -foo$ ]] +# Specifying a ref without a rev shouldn't pick a cached rev for a different ref +export _NIX_FORCE_HTTP=1 +rev_tag1_nix=$(nix eval --impure --raw --expr "(builtins.fetchGit { url = \"file://$repo\"; ref = \"refs/tags/tag1\"; }).rev") +rev_tag1=$(git -C $repo rev-parse refs/tags/tag1) +[[ $rev_tag1_nix = $rev_tag1 ]] +rev_tag2_nix=$(nix eval --impure --raw --expr "(builtins.fetchGit { url = \"file://$repo\"; ref = \"refs/tags/tag2\"; }).rev") +rev_tag2=$(git -C $repo rev-parse refs/tags/tag2) +[[ $rev_tag2_nix = $rev_tag2 ]] +unset _NIX_FORCE_HTTP + # should fail if there is no repo rm -rf $repo/.git (! nix eval --impure --raw --expr "(builtins.fetchGit \"file://$repo\").outPath") @@ -224,3 +234,17 @@ rm -rf $repo/.git # should succeed for a repo without commits git init $repo path10=$(nix eval --impure --raw --expr "(builtins.fetchGit \"file://$repo\").outPath") + +# should succeed for a path with a space +# regression test for #7707 +repo="$TEST_ROOT/a b" +git init "$repo" +git -C "$repo" config user.email "foobar@example.com" +git -C "$repo" config user.name "Foobar" + +echo utrecht > "$repo/hello" +touch "$repo/.gitignore" +git -C "$repo" add hello .gitignore +git -C "$repo" commit -m 'Bla1' +cd "$repo" +path11=$(nix eval --impure --raw --expr "(builtins.fetchGit ./.).outPath") diff --git a/tests/fetchGitRefs.sh b/tests/fetchGitRefs.sh index 52926040b..d643fea04 100644 --- a/tests/fetchGitRefs.sh +++ b/tests/fetchGitRefs.sh @@ -1,9 +1,6 @@ source common.sh -if [[ -z $(type -p git) ]]; then - echo "Git not installed; skipping Git tests" - exit 99 -fi +requireGit clearStore @@ -56,7 +53,7 @@ invalid_ref() { else (! git check-ref-format --branch "$1" >/dev/null 2>&1) fi - nix --debug eval --raw --impure --expr "(builtins.fetchGit { url = $repo; ref = ''$1''; }).outPath" 2>&1 | grep 'invalid Git branch/tag name' >/dev/null + expect 1 nix --debug eval --raw --impure --expr "(builtins.fetchGit { url = $repo; ref = ''$1''; }).outPath" 2>&1 | grep 'invalid Git branch/tag name' >/dev/null } diff --git a/tests/fetchGitSubmodules.sh b/tests/fetchGitSubmodules.sh index 5f104355f..df81232e5 100644 --- a/tests/fetchGitSubmodules.sh +++ b/tests/fetchGitSubmodules.sh @@ -2,10 +2,7 @@ source common.sh set -u -if [[ -z $(type -p git) ]]; then - echo "Git not installed; skipping Git submodule tests" - exit 99 -fi +requireGit clearStore @@ -14,6 +11,15 @@ subRepo=$TEST_ROOT/gitSubmodulesSub rm -rf ${rootRepo} ${subRepo} $TEST_HOME/.cache/nix +# Submodules can't be fetched locally by default, which can cause +# information leakage vulnerabilities, but for these tests our +# submodule is intentionally local and it's all trusted, so we +# disable this restriction. Setting it per repo is not sufficient, as +# the repo-local config does not apply to the commands run from +# outside the repos by Nix. +export XDG_CONFIG_HOME=$TEST_HOME/.config +git config --global protocol.file.allow always + initGitRepo() { git init $1 git -C $1 config user.email "foobar@example.com" @@ -95,3 +101,28 @@ noSubmoduleRepoBaseline=$(nix eval --raw --expr "(builtins.fetchGit { url = file noSubmoduleRepo=$(nix eval --raw --expr "(builtins.fetchGit { url = file://$subRepo; rev = \"$subRev\"; submodules = true; }).outPath") [[ $noSubmoduleRepoBaseline == $noSubmoduleRepo ]] + +# Test relative submodule URLs. +rm $TEST_HOME/.cache/nix/fetcher-cache* +rm -rf $rootRepo/.git $rootRepo/.gitmodules $rootRepo/sub +initGitRepo $rootRepo +git -C $rootRepo submodule add ../gitSubmodulesSub sub +git -C $rootRepo commit -m "Add submodule" +rev2=$(git -C $rootRepo rev-parse HEAD) +pathWithRelative=$(nix eval --raw --expr "(builtins.fetchGit { url = file://$rootRepo; rev = \"$rev2\"; submodules = true; }).outPath") +diff -r -x .gitmodules $pathWithSubmodules $pathWithRelative + +# Test clones that have an upstream with relative submodule URLs. +rm $TEST_HOME/.cache/nix/fetcher-cache* +cloneRepo=$TEST_ROOT/a/b/gitSubmodulesClone # NB /a/b to make the relative path not work relative to $cloneRepo +git clone $rootRepo $cloneRepo +pathIndirect=$(nix eval --raw --expr "(builtins.fetchGit { url = file://$cloneRepo; rev = \"$rev2\"; submodules = true; }).outPath") +[[ $pathIndirect = $pathWithRelative ]] + +# Test that if the clone has the submodule already, we're not fetching +# it again. +git -C $cloneRepo submodule update --init +rm $TEST_HOME/.cache/nix/fetcher-cache* +rm -rf $subRepo +pathSubmoduleGone=$(nix eval --raw --expr "(builtins.fetchGit { url = file://$cloneRepo; rev = \"$rev2\"; submodules = true; }).outPath") +[[ $pathSubmoduleGone = $pathWithRelative ]] diff --git a/tests/fetchMercurial.sh b/tests/fetchMercurial.sh index 5c64ffd26..e6f8525c6 100644 --- a/tests/fetchMercurial.sh +++ b/tests/fetchMercurial.sh @@ -1,9 +1,6 @@ source common.sh -if [[ -z $(type -p hg) ]]; then - echo "Mercurial not installed; skipping Mercurial tests" - exit 99 -fi +[[ $(type -p hq) ]] || skipTest "Mercurial not installed" clearStore diff --git a/tests/fetchTree-file.sh b/tests/fetchTree-file.sh index f0c530466..fe569cfb8 100644 --- a/tests/fetchTree-file.sh +++ b/tests/fetchTree-file.sh @@ -79,7 +79,7 @@ EOF EOF popd - [[ -z "${NIX_DAEMON_PACKAGE}" ]] && return 0 + [[ -z "${NIX_DAEMON_PACKAGE-}" ]] && return 0 # Ensure that a lockfile generated by the current Nix for tarball inputs # can still be read by an older Nix @@ -91,7 +91,7 @@ EOF flake = false; }; outputs = { self, tarball }: { - foo = builtins.readFile "${tarball}/test_input_file"; + foo = builtins.readFile "\${tarball}/test_input_file"; }; } nix flake update diff --git a/tests/fetchurl.sh b/tests/fetchurl.sh index b41d8c4b7..8cd40c09f 100644 --- a/tests/fetchurl.sh +++ b/tests/fetchurl.sh @@ -62,7 +62,7 @@ hash=$(nix-hash --flat --type sha256 $nar) outPath=$(nix-build -vvvvv --expr 'import <nix/fetchurl.nix>' --argstr url file://$nar --argstr sha256 $hash \ --arg unpack true --argstr name xyzzy --no-out-link) -echo $outPath | grep -q 'xyzzy' +echo $outPath | grepQuiet 'xyzzy' test -x $outPath/fetchurl.sh test -L $outPath/symlink diff --git a/tests/flakes/absolute-paths.sh b/tests/flakes/absolute-paths.sh new file mode 100644 index 000000000..e7bfba12d --- /dev/null +++ b/tests/flakes/absolute-paths.sh @@ -0,0 +1,17 @@ +source ./common.sh + +requireGit + +flake1Dir=$TEST_ROOT/flake1 +flake2Dir=$TEST_ROOT/flake2 + +createGitRepo $flake1Dir +cat > $flake1Dir/flake.nix <<EOF +{ + outputs = { self }: { x = builtins.readFile $(pwd)/absolute-paths.sh; }; +} +EOF +git -C $flake1Dir add flake.nix +git -C $flake1Dir commit -m Initial + +nix eval --impure --json $flake1Dir#x diff --git a/tests/flakes/build-paths.sh b/tests/flakes/build-paths.sh new file mode 100644 index 000000000..b399a066e --- /dev/null +++ b/tests/flakes/build-paths.sh @@ -0,0 +1,66 @@ +source ./common.sh + +flake1Dir=$TEST_ROOT/flake1 +flake2Dir=$TEST_ROOT/flake2 + +mkdir -p $flake1Dir $flake2Dir + +writeSimpleFlake $flake2Dir +tar cfz $TEST_ROOT/flake.tar.gz -C $TEST_ROOT flake2 +hash=$(nix hash path $flake2Dir) + +dep=$(nix store add-path ./common.sh) + +cat > $flake1Dir/flake.nix <<EOF +{ + inputs.flake2.url = "file://$TEST_ROOT/flake.tar.gz"; + + outputs = { self, flake2 }: { + + a1 = builtins.fetchTarball { + #type = "tarball"; + url = "file://$TEST_ROOT/flake.tar.gz"; + sha256 = "$hash"; + }; + + a2 = ./foo; + + a3 = ./.; + + a4 = self.outPath; + + # FIXME + a5 = self; + + a6 = flake2.outPath; + + # FIXME + a7 = "\${flake2}/config.nix"; + + # This is only allowed in impure mode. + a8 = builtins.storePath $dep; + + a9 = "$dep"; + }; +} +EOF + +echo bar > $flake1Dir/foo + +nix build --json --out-link $TEST_ROOT/result $flake1Dir#a1 +[[ -e $TEST_ROOT/result/simple.nix ]] + +nix build --json --out-link $TEST_ROOT/result $flake1Dir#a2 +[[ $(cat $TEST_ROOT/result) = bar ]] + +nix build --json --out-link $TEST_ROOT/result $flake1Dir#a3 + +nix build --json --out-link $TEST_ROOT/result $flake1Dir#a4 + +nix build --json --out-link $TEST_ROOT/result $flake1Dir#a6 +[[ -e $TEST_ROOT/result/simple.nix ]] + +nix build --impure --json --out-link $TEST_ROOT/result $flake1Dir#a8 +diff common.sh $TEST_ROOT/result + +(! nix build --impure --json --out-link $TEST_ROOT/result $flake1Dir#a9) diff --git a/tests/flakes/check.sh b/tests/flakes/check.sh index f572aa75c..865ca61b4 100644 --- a/tests/flakes/check.sh +++ b/tests/flakes/check.sh @@ -41,9 +41,9 @@ nix flake check $flakeDir cat > $flakeDir/flake.nix <<EOF { outputs = { self }: { - nixosModules.foo = { + nixosModules.foo = assert false; { a.b.c = 123; - foo = assert false; true; + foo = true; }; }; } @@ -66,18 +66,6 @@ nix flake check $flakeDir cat > $flakeDir/flake.nix <<EOF { outputs = { self }: { - nixosModule = { config, pkgs }: { - a.b.c = 123; - }; - }; -} -EOF - -(! nix flake check $flakeDir) - -cat > $flakeDir/flake.nix <<EOF -{ - outputs = { self }: { packages.system-1.default = "foo"; packages.system-2.default = "bar"; }; @@ -85,5 +73,5 @@ cat > $flakeDir/flake.nix <<EOF EOF checkRes=$(nix flake check --keep-going $flakeDir 2>&1 && fail "nix flake check should have failed" || true) -echo "$checkRes" | grep -q "packages.system-1.default" -echo "$checkRes" | grep -q "packages.system-2.default" +echo "$checkRes" | grepQuiet "packages.system-1.default" +echo "$checkRes" | grepQuiet "packages.system-2.default" diff --git a/tests/flakes/common.sh b/tests/flakes/common.sh index c333733c2..427abcdde 100644 --- a/tests/flakes/common.sh +++ b/tests/flakes/common.sh @@ -2,13 +2,6 @@ source ../common.sh registry=$TEST_ROOT/registry.json -requireGit() { - if [[ -z $(type -p git) ]]; then - echo "Git not installed; skipping flake tests" - exit 99 - fi -} - writeSimpleFlake() { local flakeDir="$1" cat > $flakeDir/flake.nix <<EOF @@ -20,9 +13,13 @@ writeSimpleFlake() { foo = import ./simple.nix; default = foo; }; + packages.someOtherSystem = rec { + foo = import ./simple.nix; + default = foo; + }; # To test "nix flake init". - legacyPackages.x86_64-linux.hello = import ./simple.nix; + legacyPackages.$system.hello = import ./simple.nix; }; } EOF @@ -62,7 +59,7 @@ EOF createGitRepo() { local repo="$1" - local extraArgs="$2" + local extraArgs="${2-}" rm -rf $repo $repo.tmp mkdir -p $repo diff --git a/tests/flakes/flake-in-submodule.sh b/tests/flakes/flake-in-submodule.sh new file mode 100644 index 000000000..21a4b52de --- /dev/null +++ b/tests/flakes/flake-in-submodule.sh @@ -0,0 +1,52 @@ +source common.sh + +# Tests that: +# - flake.nix may reside inside of a git submodule +# - the flake can access content outside of the submodule +# +# rootRepo +# ├── root.nix +# └── submodule +# ├── flake.nix +# └── sub.nix + + +requireGit + +clearStore + +# Submodules can't be fetched locally by default. +# See fetchGitSubmodules.sh +export XDG_CONFIG_HOME=$TEST_HOME/.config +git config --global protocol.file.allow always + + +rootRepo=$TEST_ROOT/rootRepo +subRepo=$TEST_ROOT/submodule + + +createGitRepo $subRepo +cat > $subRepo/flake.nix <<EOF +{ + outputs = { self }: { + sub = import ./sub.nix; + root = import ../root.nix; + }; +} +EOF +echo '"expression in submodule"' > $subRepo/sub.nix +git -C $subRepo add flake.nix sub.nix +git -C $subRepo commit -m Initial + +createGitRepo $rootRepo + +git -C $rootRepo submodule init +git -C $rootRepo submodule add $subRepo submodule +echo '"expression in root repo"' > $rootRepo/root.nix +git -C $rootRepo add root.nix +git -C $rootRepo commit -m "Add root.nix" + +# Flake can live inside a submodule and can be accessed via ?dir=submodule +[[ $(nix eval --json git+file://$rootRepo\?submodules=1\&dir=submodule#sub ) = '"expression in submodule"' ]] +# The flake can access content outside of the submodule +[[ $(nix eval --json git+file://$rootRepo\?submodules=1\&dir=submodule#root ) = '"expression in root repo"' ]] diff --git a/tests/flakes/flakes.sh b/tests/flakes/flakes.sh index 267e2cd6f..5c922d7c5 100644 --- a/tests/flakes/flakes.sh +++ b/tests/flakes/flakes.sh @@ -53,7 +53,11 @@ cat > $flake3Dir/flake.nix <<EOF } EOF -git -C $flake3Dir add flake.nix +cat > $flake3Dir/default.nix <<EOF +{ x = 123; } +EOF + +git -C $flake3Dir add flake.nix default.nix git -C $flake3Dir commit -m 'Initial' cat > $nonFlakeDir/README.md <<EOF @@ -70,17 +74,19 @@ nix registry add --registry $registry flake3 git+file://$flake3Dir nix registry add --registry $registry flake4 flake3 nix registry add --registry $registry nixpkgs flake1 -# Test 'nix flake list'. +# Test 'nix registry list'. [[ $(nix registry list | wc -l) == 5 ]] +nix registry list | grep '^global' +nix registry list | grepInverse '^user' # nothing in user registry # Test 'nix flake metadata'. nix flake metadata flake1 -nix flake metadata flake1 | grep -q 'Locked URL:.*flake1.*' +nix flake metadata flake1 | grepQuiet 'Locked URL:.*flake1.*' # Test 'nix flake metadata' on a local flake. -(cd $flake1Dir && nix flake metadata) | grep -q 'URL:.*flake1.*' -(cd $flake1Dir && nix flake metadata .) | grep -q 'URL:.*flake1.*' -nix flake metadata $flake1Dir | grep -q 'URL:.*flake1.*' +(cd $flake1Dir && nix flake metadata) | grepQuiet 'URL:.*flake1.*' +(cd $flake1Dir && nix flake metadata .) | grepQuiet 'URL:.*flake1.*' +nix flake metadata $flake1Dir | grepQuiet 'URL:.*flake1.*' # Test 'nix flake metadata --json'. json=$(nix flake metadata flake1 --json | jq .) @@ -109,11 +115,12 @@ nix build -o $TEST_ROOT/result git+file://$flake1Dir nix build -o $flake1Dir/result git+file://$flake1Dir nix path-info $flake1Dir/result -# 'getFlake' on a mutable flakeref should fail in pure mode, but succeed in impure mode. +# 'getFlake' on an unlocked flakeref should fail in pure mode, but +# succeed in impure mode. (! nix build -o $TEST_ROOT/result --expr "(builtins.getFlake \"$flake1Dir\").packages.$system.default") nix build -o $TEST_ROOT/result --expr "(builtins.getFlake \"$flake1Dir\").packages.$system.default" --impure -# 'getFlake' on an immutable flakeref should succeed even in pure mode. +# 'getFlake' on a locked flakeref should succeed even in pure mode. nix build -o $TEST_ROOT/result --expr "(builtins.getFlake \"git+file://$flake1Dir?rev=$hash2\").packages.$system.default" # Building a flake with an unlocked dependency should fail in pure mode. @@ -127,11 +134,11 @@ nix build -o $TEST_ROOT/result flake2#bar --impure --no-write-lock-file nix eval --expr "builtins.getFlake \"$flake2Dir\"" --impure # Building a local flake with an unlocked dependency should fail with --no-update-lock-file. -nix build -o $TEST_ROOT/result $flake2Dir#bar --no-update-lock-file 2>&1 | grep 'requires lock file changes' +expect 1 nix build -o $TEST_ROOT/result $flake2Dir#bar --no-update-lock-file 2>&1 | grep 'requires lock file changes' # But it should succeed without that flag. nix build -o $TEST_ROOT/result $flake2Dir#bar --no-write-lock-file -nix build -o $TEST_ROOT/result $flake2Dir#bar --no-update-lock-file 2>&1 | grep 'requires lock file changes' +expect 1 nix build -o $TEST_ROOT/result $flake2Dir#bar --no-update-lock-file 2>&1 | grep 'requires lock file changes' nix build -o $TEST_ROOT/result $flake2Dir#bar --commit-lock-file [[ -e $flake2Dir/flake.lock ]] [[ -z $(git -C $flake2Dir diff main || echo failed) ]] @@ -189,10 +196,10 @@ git -C $flake3Dir add flake.lock git -C $flake3Dir commit -m 'Add lockfile' # Test whether registry caching works. -nix registry list --flake-registry file://$registry | grep -q flake3 +nix registry list --flake-registry file://$registry | grepQuiet flake3 mv $registry $registry.tmp nix store gc -nix registry list --flake-registry file://$registry --refresh | grep -q flake3 +nix registry list --flake-registry file://$registry --refresh | grepQuiet flake3 mv $registry.tmp $registry # Test whether flakes are registered as GC roots for offline use. @@ -335,6 +342,16 @@ nix registry pin flake1 flake3 nix registry remove flake1 [[ $(nix registry list | wc -l) == 5 ]] +# Test 'nix registry list' with a disabled global registry. +nix registry add user-flake1 git+file://$flake1Dir +nix registry add user-flake2 git+file://$flake2Dir +[[ $(nix --flake-registry "" registry list | wc -l) == 2 ]] +nix --flake-registry "" registry list | grepQuietInverse '^global' # nothing in global registry +nix --flake-registry "" registry list | grepQuiet '^user' +nix registry remove user-flake1 +nix registry remove user-flake2 +[[ $(nix registry list | wc -l) == 5 ]] + # Test 'nix flake clone'. rm -rf $TEST_ROOT/flake1-v2 nix flake clone flake1 --dest $TEST_ROOT/flake1-v2 @@ -437,7 +454,7 @@ url=$(nix flake metadata --json file://$TEST_ROOT/flake.tar.gz | jq -r .url) nix build -o $TEST_ROOT/result $url # Building with an incorrect SRI hash should fail. -nix build -o $TEST_ROOT/result "file://$TEST_ROOT/flake.tar.gz?narHash=sha256-qQ2Zz4DNHViCUrp6gTS7EE4+RMqFQtUfWF2UNUtJKS0=" 2>&1 | grep 'NAR hash mismatch' +expectStderr 102 nix build -o $TEST_ROOT/result "file://$TEST_ROOT/flake.tar.gz?narHash=sha256-qQ2Zz4DNHViCUrp6gTS7EE4+RMqFQtUfWF2UNUtJKS0=" | grep 'NAR hash mismatch' # Test --override-input. git -C $flake3Dir reset --hard @@ -460,7 +477,7 @@ nix flake lock $flake3Dir --update-input flake2/flake1 # Test 'nix flake metadata --json'. nix flake metadata $flake3Dir --json | jq . -# Test flake in store does not evaluate +# Test flake in store does not evaluate. rm -rf $badFlakeDir mkdir $badFlakeDir echo INVALID > $badFlakeDir/flake.nix @@ -468,3 +485,9 @@ nix store delete $(nix store add-path $badFlakeDir) [[ $(nix path-info $(nix store add-path $flake1Dir)) =~ flake1 ]] [[ $(nix path-info path:$(nix store add-path $flake1Dir)) =~ simple ]] + +# Test fetching flakerefs in the legacy CLI. +[[ $(nix-instantiate --eval flake:flake3 -A x) = 123 ]] +[[ $(nix-instantiate --eval flake:git+file://$flake3Dir -A x) = 123 ]] +[[ $(nix-instantiate -I flake3=flake:flake3 --eval '<flake3>' -A x) = 123 ]] +[[ $(NIX_PATH=flake3=flake:flake3 nix-instantiate --eval '<flake3>' -A x) = 123 ]] diff --git a/tests/flakes/follow-paths.sh b/tests/flakes/follow-paths.sh index 19cc1bafa..fe9b51c65 100644 --- a/tests/flakes/follow-paths.sh +++ b/tests/flakes/follow-paths.sh @@ -128,7 +128,7 @@ EOF git -C $flakeFollowsA add flake.nix -nix flake lock $flakeFollowsA 2>&1 | grep 'points outside' +expect 1 nix flake lock $flakeFollowsA 2>&1 | grep 'points outside' # Non-existant follows should print a warning. cat >$flakeFollowsA/flake.nix <<EOF diff --git a/tests/flakes/init.sh b/tests/flakes/init.sh index 36cb9956a..2d4c77ba1 100644 --- a/tests/flakes/init.sh +++ b/tests/flakes/init.sh @@ -41,8 +41,8 @@ cat > $templatesDir/trivial/flake.nix <<EOF description = "A flake for building Hello World"; outputs = { self, nixpkgs }: { - packages.x86_64-linux = rec { - hello = nixpkgs.legacyPackages.x86_64-linux.hello; + packages.$system = rec { + hello = nixpkgs.legacyPackages.$system.hello; default = hello; }; }; diff --git a/tests/flakes/inputs.sh b/tests/flakes/inputs.sh new file mode 100644 index 000000000..80620488a --- /dev/null +++ b/tests/flakes/inputs.sh @@ -0,0 +1,80 @@ +source ./common.sh + +requireGit + + +test_subdir_self_path() { + baseDir=$TEST_ROOT/$RANDOM + flakeDir=$baseDir/b-low + mkdir -p $flakeDir + writeSimpleFlake $baseDir + writeSimpleFlake $flakeDir + + echo all good > $flakeDir/message + cat > $flakeDir/flake.nix <<EOF +{ + outputs = inputs: rec { + packages.$system = rec { + default = + assert builtins.readFile ./message == "all good\n"; + assert builtins.readFile (inputs.self + "/message") == "all good\n"; + import ./simple.nix; + }; + }; +} +EOF + ( + nix build $baseDir?dir=b-low --no-link + ) +} +test_subdir_self_path + + +test_git_subdir_self_path() { + repoDir=$TEST_ROOT/repo-$RANDOM + createGitRepo $repoDir + flakeDir=$repoDir/b-low + mkdir -p $flakeDir + writeSimpleFlake $repoDir + writeSimpleFlake $flakeDir + + echo all good > $flakeDir/message + cat > $flakeDir/flake.nix <<EOF +{ + outputs = inputs: rec { + packages.$system = rec { + default = + assert builtins.readFile ./message == "all good\n"; + assert builtins.readFile (inputs.self + "/message") == "all good\n"; + assert inputs.self.outPath == inputs.self.sourceInfo.outPath + "/b-low"; + import ./simple.nix; + }; + }; +} +EOF + ( + cd $flakeDir + git add . + git commit -m init + # nix build + ) + + clientDir=$TEST_ROOT/client-$RANDOM + mkdir -p $clientDir + cat > $clientDir/flake.nix <<EOF +{ + inputs.inp = { + type = "git"; + url = "file://$repoDir"; + dir = "b-low"; + }; + + outputs = inputs: rec { + packages = inputs.inp.packages; + }; +} +EOF + nix build $clientDir --no-link + +} +test_git_subdir_self_path diff --git a/tests/flakes/mercurial.sh b/tests/flakes/mercurial.sh index 2614006c8..0622c79b7 100644 --- a/tests/flakes/mercurial.sh +++ b/tests/flakes/mercurial.sh @@ -1,9 +1,6 @@ source ./common.sh -if [[ -z $(type -p hg) ]]; then - echo "Mercurial not installed; skipping" - exit 99 -fi +[[ $(type -p hq) ]] || skipTest "Mercurial not installed" flake1Dir=$TEST_ROOT/flake-hg1 mkdir -p $flake1Dir diff --git a/tests/flakes/show.sh b/tests/flakes/show.sh new file mode 100644 index 000000000..dd13264b9 --- /dev/null +++ b/tests/flakes/show.sh @@ -0,0 +1,66 @@ +source ./common.sh + +flakeDir=$TEST_ROOT/flake +mkdir -p "$flakeDir" + +writeSimpleFlake "$flakeDir" +cd "$flakeDir" + + +# By default: Only show the packages content for the current system and no +# legacyPackages at all +nix flake show --json > show-output.json +nix eval --impure --expr ' +let show_output = builtins.fromJSON (builtins.readFile ./show-output.json); +in +assert show_output.packages.someOtherSystem.default == {}; +assert show_output.packages.${builtins.currentSystem}.default.name == "simple"; +assert show_output.legacyPackages.${builtins.currentSystem} == {}; +true +' + +# With `--all-systems`, show the packages for all systems +nix flake show --json --all-systems > show-output.json +nix eval --impure --expr ' +let show_output = builtins.fromJSON (builtins.readFile ./show-output.json); +in +assert show_output.packages.someOtherSystem.default.name == "simple"; +assert show_output.legacyPackages.${builtins.currentSystem} == {}; +true +' + +# With `--legacy`, show the legacy packages +nix flake show --json --legacy > show-output.json +nix eval --impure --expr ' +let show_output = builtins.fromJSON (builtins.readFile ./show-output.json); +in +assert show_output.legacyPackages.${builtins.currentSystem}.hello.name == "simple"; +true +' + +# Test that attributes are only reported when they have actual content +cat >flake.nix <<EOF +{ + description = "Bla bla"; + + outputs = inputs: rec { + apps.$system = { }; + checks.$system = { }; + devShells.$system = { }; + legacyPackages.$system = { }; + packages.$system = { }; + packages.someOtherSystem = { }; + + formatter = { }; + nixosConfigurations = { }; + nixosModules = { }; + }; +} +EOF +nix flake show --json --all-systems > show-output.json +nix eval --impure --expr ' +let show_output = builtins.fromJSON (builtins.readFile ./show-output.json); +in +assert show_output == { }; +true +' diff --git a/tests/flakes/unlocked-override.sh b/tests/flakes/unlocked-override.sh new file mode 100644 index 000000000..8abc8b7d3 --- /dev/null +++ b/tests/flakes/unlocked-override.sh @@ -0,0 +1,30 @@ +source ./common.sh + +requireGit + +flake1Dir=$TEST_ROOT/flake1 +flake2Dir=$TEST_ROOT/flake2 + +createGitRepo $flake1Dir +cat > $flake1Dir/flake.nix <<EOF +{ + outputs = { self }: { x = import ./x.nix; }; +} +EOF +echo 123 > $flake1Dir/x.nix +git -C $flake1Dir add flake.nix x.nix +git -C $flake1Dir commit -m Initial + +createGitRepo $flake2Dir +cat > $flake2Dir/flake.nix <<EOF +{ + outputs = { self, flake1 }: { x = flake1.x; }; +} +EOF +git -C $flake2Dir add flake.nix + +[[ $(nix eval --json $flake2Dir#x --override-input flake1 $TEST_ROOT/flake1) = 123 ]] + +echo 456 > $flake1Dir/x.nix + +[[ $(nix eval --json $flake2Dir#x --override-input flake1 $TEST_ROOT/flake1) = 456 ]] diff --git a/tests/fmt.sh b/tests/fmt.sh index 254681ca2..3c1bd9989 100644 --- a/tests/fmt.sh +++ b/tests/fmt.sh @@ -1,7 +1,5 @@ source common.sh -set -o pipefail - clearStore rm -rf $TEST_HOME/.cache $TEST_HOME/.config $TEST_HOME/.local diff --git a/tests/function-trace.sh b/tests/function-trace.sh index 0b7f49d82..bd804bf18 100755 --- a/tests/function-trace.sh +++ b/tests/function-trace.sh @@ -10,17 +10,15 @@ expect_trace() { --trace-function-calls \ --expr "$expr" 2>&1 \ | grep "function-trace" \ - | sed -e 's/ [0-9]*$//' - ); + | sed -e 's/ [0-9]*$//' \ + || true + ) echo -n "Tracing expression '$expr'" - set +e msg=$(diff -swB \ <(echo "$expect") \ <(echo "$actual") - ); - result=$? - set -e + ) && result=0 || result=$? if [ $result -eq 0 ]; then echo " ok." else @@ -32,40 +30,38 @@ expect_trace() { # failure inside a tryEval expect_trace 'builtins.tryEval (throw "example")' " -function-trace entered (string):1:1 at -function-trace entered (string):1:19 at -function-trace exited (string):1:19 at -function-trace exited (string):1:1 at +function-trace entered «string»:1:1 at +function-trace entered «string»:1:19 at +function-trace exited «string»:1:19 at +function-trace exited «string»:1:1 at " # Missing argument to a formal function expect_trace '({ x }: x) { }' " -function-trace entered (string):1:1 at -function-trace exited (string):1:1 at +function-trace entered «string»:1:1 at +function-trace exited «string»:1:1 at " # Too many arguments to a formal function expect_trace '({ x }: x) { x = "x"; y = "y"; }' " -function-trace entered (string):1:1 at -function-trace exited (string):1:1 at +function-trace entered «string»:1:1 at +function-trace exited «string»:1:1 at " # Not enough arguments to a lambda expect_trace '(x: y: x + y) 1' " -function-trace entered (string):1:1 at -function-trace exited (string):1:1 at +function-trace entered «string»:1:1 at +function-trace exited «string»:1:1 at " # Too many arguments to a lambda expect_trace '(x: x) 1 2' " -function-trace entered (string):1:1 at -function-trace exited (string):1:1 at +function-trace entered «string»:1:1 at +function-trace exited «string»:1:1 at " # Not a function expect_trace '1 2' " -function-trace entered (string):1:1 at -function-trace exited (string):1:1 at +function-trace entered «string»:1:1 at +function-trace exited «string»:1:1 at " - -set -e diff --git a/tests/gc-runtime.sh b/tests/gc-runtime.sh index 6094959cb..dc1826a55 100644 --- a/tests/gc-runtime.sh +++ b/tests/gc-runtime.sh @@ -4,7 +4,7 @@ case $system in *linux*) ;; *) - exit 99; + skipTest "Not running Linux"; esac set -m # enable job control, needed for kill diff --git a/tests/gc.sh b/tests/gc.sh index ad09a8b39..98d6cb032 100644 --- a/tests/gc.sh +++ b/tests/gc.sh @@ -50,3 +50,20 @@ if test -e $outPath/foobar; then false; fi # Check that the store is empty. rmdir $NIX_STORE_DIR/.links rmdir $NIX_STORE_DIR + +## Test `nix-collect-garbage -d` +# `nix-env` doesn't work with CA derivations, so let's ignore that bit if we're +# using them +if [[ -z "${NIX_TESTS_CA_BY_DEFAULT:-}" ]]; then + clearProfiles + # Run two `nix-env` commands, should create two generations of + # the profile + nix-env -f ./user-envs.nix -i foo-1.0 + nix-env -f ./user-envs.nix -i foo-2.0pre1 + [[ $(nix-env --list-generations | wc -l) -eq 2 ]] + + # Clear the profile history. There should be only one generation + # left + nix-collect-garbage -d + [[ $(nix-env --list-generations | wc -l) -eq 1 ]] +fi diff --git a/tests/hash.sh b/tests/hash.sh index e5f75e2cf..34c1bb38a 100644 --- a/tests/hash.sh +++ b/tests/hash.sh @@ -2,13 +2,19 @@ source common.sh try () { printf "%s" "$2" > $TEST_ROOT/vector - hash=$(nix hash file --base16 $EXTRA --type "$1" $TEST_ROOT/vector) - if test "$hash" != "$3"; then - echo "hash $1, expected $3, got $hash" + hash="$(nix-hash --flat ${FORMAT_FLAG-} --type "$1" "$TEST_ROOT/vector")" + if ! (( "${NO_TEST_CLASSIC-}" )) && test "$hash" != "$3"; then + echo "try nix-hash: hash $1, expected $3, got $hash" + exit 1 + fi + hash="$(nix hash file ${FORMAT_FLAG-} --type "$1" "$TEST_ROOT/vector")" + if ! (( "${NO_TEST_NIX_COMMAND-}" )) && test "$hash" != "$3"; then + echo "try nix hash: hash $1, expected $3, got $hash" exit 1 fi } +FORMAT_FLAG=--base16 try md5 "" "d41d8cd98f00b204e9800998ecf8427e" try md5 "a" "0cc175b9c0f1b6a831c399e269772661" try md5 "abc" "900150983cd24fb0d6963f7d28e17f72" @@ -28,16 +34,24 @@ try sha256 "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq" "248d6a61d try sha512 "" "cf83e1357eefb8bdf1542850d66d8007d620e4050b5715dc83f4a921d36ce9ce47d0d13c5d85f2b0ff8318d2877eec2f63b931bd47417a81a538327af927da3e" try sha512 "abc" "ddaf35a193617abacc417349ae20413112e6fa4e89a97ea20a9eeee64b55d39a2192992a274fc1a836ba3c23a3feebbd454d4423643ce80e2a9ac94fa54ca49f" try sha512 "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq" "204a8fc6dda82f0a0ced7beb8e08a41657c16ef468b228a8279be331a703c33596fd15c13b1b07f9aa1d3bea57789ca031ad85c7a71dd70354ec631238ca3445" +unset FORMAT_FLAG -EXTRA=--base32 +FORMAT_FLAG=--base32 try sha256 "abc" "1b8m03r63zqhnjf7l5wnldhh7c134ap5vpj0850ymkq1iyzicy5s" -EXTRA= +unset FORMAT_FLAG -EXTRA=--sri +FORMAT_FLAG=--sri try sha512 "" "sha512-z4PhNX7vuL3xVChQ1m2AB9Yg5AULVxXcg/SpIdNs6c5H0NE8XYXysP+DGNKHfuwvY7kxvUdBeoGlODJ6+SfaPg==" try sha512 "abc" "sha512-3a81oZNherrMQXNJriBBMRLm+k6JqX6iCp7u5ktV05ohkpkqJ0/BqDa6PCOj/uu9RU1EI2Q86A4qmslPpUyknw==" try sha512 "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq" "sha512-IEqPxt2oLwoM7XvrjgikFlfBbvRosiioJ5vjMacDwzWW/RXBOxsH+aodO+pXeJygMa2Fx6cd1wNU7GMSOMo0RQ==" try sha256 "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq" "sha256-JI1qYdIGOLjlwCaTDD5gOaM85Flk/yFn9uzt1BnbBsE=" +unset FORMAT_FLAG + +# nix-hash [--flat] defaults to the Base16 format +NO_TEST_NIX_COMMAND=1 try sha512 "abc" "ddaf35a193617abacc417349ae20413112e6fa4e89a97ea20a9eeee64b55d39a2192992a274fc1a836ba3c23a3feebbd454d4423643ce80e2a9ac94fa54ca49f" + +# nix hash [file|path] defaults to the SRI format +NO_TEST_CLASSIC=1 try sha512 "abc" "sha512-3a81oZNherrMQXNJriBBMRLm+k6JqX6iCp7u5ktV05ohkpkqJ0/BqDa6PCOj/uu9RU1EI2Q86A4qmslPpUyknw==" try2 () { hash=$(nix-hash --type "$1" $TEST_ROOT/hash-path) @@ -69,12 +83,18 @@ try2 md5 "f78b733a68f5edbdf9413899339eaa4a" # Conversion. try3() { + h64=$(nix-hash --type "$1" --to-base64 "$2") + [ "$h64" = "$4" ] h64=$(nix hash to-base64 --type "$1" "$2") [ "$h64" = "$4" ] + sri=$(nix-hash --type "$1" --to-sri "$2") + [ "$sri" = "$1-$4" ] sri=$(nix hash to-sri --type "$1" "$2") [ "$sri" = "$1-$4" ] h32=$(nix-hash --type "$1" --to-base32 "$2") [ "$h32" = "$3" ] + h32=$(nix hash to-base32 --type "$1" "$2") + [ "$h32" = "$3" ] h16=$(nix-hash --type "$1" --to-base16 "$h32") [ "$h16" = "$2" ] h16=$(nix hash to-base16 --type "$1" "$h64") diff --git a/tests/impure-derivations.sh b/tests/impure-derivations.sh index 35ae3f5d3..7595fdd35 100644 --- a/tests/impure-derivations.sh +++ b/tests/impure-derivations.sh @@ -2,16 +2,15 @@ source common.sh requireDaemonNewerThan "2.8pre20220311" -enableFeatures "ca-derivations ca-references impure-derivations" +enableFeatures "ca-derivations impure-derivations" restartDaemon -set -o pipefail - clearStore # Basic test of impure derivations: building one a second time should not use the previous result. printf 0 > $TEST_ROOT/counter +nix build --dry-run --json --file ./impure-derivations.nix impure.all json=$(nix build -L --no-link --json --file ./impure-derivations.nix impure.all) path1=$(echo $json | jq -r .[].outputs.out) path1_stuff=$(echo $json | jq -r .[].outputs.stuff) diff --git a/tests/init.sh b/tests/init.sh index 3c6d5917d..c420e8c9f 100644..100755 --- a/tests/init.sh +++ b/tests/init.sh @@ -1,8 +1,11 @@ -source common.sh +# Don't start the daemon +source common/vars-and-functions.sh test -n "$TEST_ROOT" if test -d "$TEST_ROOT"; then chmod -R u+w "$TEST_ROOT" + # We would delete any daemon socket, so let's stop the daemon first. + killDaemon rm -rf "$TEST_ROOT" fi mkdir "$TEST_ROOT" diff --git a/tests/install-darwin.sh b/tests/install-darwin.sh index 7e44e54c4..ea2b75323 100755 --- a/tests/install-darwin.sh +++ b/tests/install-darwin.sh @@ -4,7 +4,7 @@ set -eux cleanup() { PLIST="/Library/LaunchDaemons/org.nixos.nix-daemon.plist" - if sudo launchctl list | grep -q nix-daemon; then + if sudo launchctl list | grepQuiet nix-daemon; then sudo launchctl unload "$PLIST" fi diff --git a/tests/installer/default.nix b/tests/installer/default.nix new file mode 100644 index 000000000..31d83699d --- /dev/null +++ b/tests/installer/default.nix @@ -0,0 +1,220 @@ +{ binaryTarballs +, nixpkgsFor +}: + +let + + installScripts = { + install-default = { + script = '' + tar -xf ./nix.tar.xz + mv ./nix-* nix + ./nix/install --no-channel-add + ''; + }; + + install-force-no-daemon = { + script = '' + tar -xf ./nix.tar.xz + mv ./nix-* nix + ./nix/install --no-daemon + ''; + }; + + install-force-daemon = { + script = '' + tar -xf ./nix.tar.xz + mv ./nix-* nix + ./nix/install --daemon --no-channel-add + ''; + }; + }; + + disableSELinux = "sudo setenforce 0"; + + images = { + + /* + "ubuntu-14-04" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/ubuntu/boxes/trusty64/versions/20190514.0.0/providers/virtualbox.box"; + hash = "sha256-iUUXyRY8iW7DGirb0zwGgf1fRbLA7wimTJKgP7l/OQ8="; + }; + rootDisk = "box-disk1.vmdk"; + system = "x86_64-linux"; + }; + */ + + "ubuntu-16-04" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/generic/boxes/ubuntu1604/versions/4.1.12/providers/libvirt.box"; + hash = "sha256-lO4oYQR2tCh5auxAYe6bPOgEqOgv3Y3GC1QM1tEEEU8="; + }; + rootDisk = "box.img"; + system = "x86_64-linux"; + }; + + "ubuntu-22-04" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/generic/boxes/ubuntu2204/versions/4.1.12/providers/libvirt.box"; + hash = "sha256-HNll0Qikw/xGIcogni5lz01vUv+R3o8xowP2EtqjuUQ="; + }; + rootDisk = "box.img"; + system = "x86_64-linux"; + }; + + "fedora-36" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/generic/boxes/fedora36/versions/4.1.12/providers/libvirt.box"; + hash = "sha256-rxPgnDnFkTDwvdqn2CV3ZUo3re9AdPtSZ9SvOHNvaks="; + }; + rootDisk = "box.img"; + system = "x86_64-linux"; + postBoot = disableSELinux; + }; + + # Currently fails with 'error while loading shared libraries: + # libsodium.so.23: cannot stat shared object: Invalid argument'. + /* + "rhel-6" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/generic/boxes/rhel6/versions/4.1.12/providers/libvirt.box"; + hash = "sha256-QwzbvRoRRGqUCQptM7X/InRWFSP2sqwRt2HaaO6zBGM="; + }; + rootDisk = "box.img"; + system = "x86_64-linux"; + }; + */ + + "rhel-7" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/generic/boxes/rhel7/versions/4.1.12/providers/libvirt.box"; + hash = "sha256-b4afnqKCO9oWXgYHb9DeQ2berSwOjS27rSd9TxXDc/U="; + }; + rootDisk = "box.img"; + system = "x86_64-linux"; + }; + + "rhel-8" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/generic/boxes/rhel8/versions/4.1.12/providers/libvirt.box"; + hash = "sha256-zFOPjSputy1dPgrQRixBXmlyN88cAKjJ21VvjSWUCUY="; + }; + rootDisk = "box.img"; + system = "x86_64-linux"; + postBoot = disableSELinux; + }; + + "rhel-9" = { + image = import <nix/fetchurl.nix> { + url = "https://app.vagrantup.com/generic/boxes/rhel9/versions/4.1.12/providers/libvirt.box"; + hash = "sha256-vL/FbB3kK1rcSaR627nWmScYGKGk4seSmAdq6N5diMg="; + }; + rootDisk = "box.img"; + system = "x86_64-linux"; + postBoot = disableSELinux; + extraQemuOpts = "-cpu Westmere-v2"; + }; + + }; + + makeTest = imageName: testName: + let image = images.${imageName}; in + with nixpkgsFor.${image.system}.native; + runCommand + "installer-test-${imageName}-${testName}" + { buildInputs = [ qemu_kvm openssh ]; + image = image.image; + postBoot = image.postBoot or ""; + installScript = installScripts.${testName}.script; + binaryTarball = binaryTarballs.${system}; + } + '' + shopt -s nullglob + + echo "Unpacking Vagrant box $image..." + tar xvf $image + + image_type=$(qemu-img info ${image.rootDisk} | sed 's/file format: \(.*\)/\1/; t; d') + + qemu-img create -b ./${image.rootDisk} -F "$image_type" -f qcow2 ./disk.qcow2 + + extra_qemu_opts="${image.extraQemuOpts or ""}" + + # Add the config disk, required by the Ubuntu images. + config_drive=$(echo *configdrive.vmdk || true) + if [[ -n $config_drive ]]; then + extra_qemu_opts+=" -drive id=disk2,file=$config_drive,if=virtio" + fi + + echo "Starting qemu..." + qemu-kvm -m 4096 -nographic \ + -drive id=disk1,file=./disk.qcow2,if=virtio \ + -netdev user,id=net0,restrict=yes,hostfwd=tcp::20022-:22 -device virtio-net-pci,netdev=net0 \ + $extra_qemu_opts & + qemu_pid=$! + trap "kill $qemu_pid" EXIT + + if ! [ -e ./vagrant_insecure_key ]; then + cp ${./vagrant_insecure_key} vagrant_insecure_key + fi + + chmod 0400 ./vagrant_insecure_key + + ssh_opts="-o StrictHostKeyChecking=no -o HostKeyAlgorithms=+ssh-rsa -o PubkeyAcceptedKeyTypes=+ssh-rsa -i ./vagrant_insecure_key" + ssh="ssh -p 20022 -q $ssh_opts vagrant@localhost" + + echo "Waiting for SSH..." + for ((i = 0; i < 120; i++)); do + echo "[ssh] Trying to connect..." + if $ssh -- true; then + echo "[ssh] Connected!" + break + fi + if ! kill -0 $qemu_pid; then + echo "qemu died unexpectedly" + exit 1 + fi + sleep 1 + done + + if [[ -n $postBoot ]]; then + echo "Running post-boot commands..." + $ssh "set -ex; $postBoot" + fi + + echo "Copying installer..." + scp -P 20022 $ssh_opts $binaryTarball/nix-*.tar.xz vagrant@localhost:nix.tar.xz + + echo "Running installer..." + $ssh "set -eux; $installScript" + + echo "Testing Nix installation..." + $ssh <<EOF + set -ex + + # FIXME: get rid of this; ideally ssh should just work. + source ~/.bash_profile || true + source ~/.bash_login || true + source ~/.profile || true + source /etc/bashrc || true + + nix-env --version + nix --extra-experimental-features nix-command store ping + + out=\$(nix-build --no-substitute -E 'derivation { name = "foo"; system = "x86_64-linux"; builder = "/bin/sh"; args = ["-c" "echo foobar > \$out"]; }') + [[ \$(cat \$out) = foobar ]] + EOF + + echo "Done!" + touch $out + ''; + +in + +builtins.mapAttrs (imageName: image: + { ${image.system} = builtins.mapAttrs (testName: test: + makeTest imageName testName + ) installScripts; + } +) images diff --git a/tests/installer/vagrant_insecure_key b/tests/installer/vagrant_insecure_key new file mode 100644 index 000000000..7d6a08390 --- /dev/null +++ b/tests/installer/vagrant_insecure_key @@ -0,0 +1,27 @@ +-----BEGIN RSA PRIVATE KEY----- +MIIEogIBAAKCAQEA6NF8iallvQVp22WDkTkyrtvp9eWW6A8YVr+kz4TjGYe7gHzI +w+niNltGEFHzD8+v1I2YJ6oXevct1YeS0o9HZyN1Q9qgCgzUFtdOKLv6IedplqoP +kcmF0aYet2PkEDo3MlTBckFXPITAMzF8dJSIFo9D8HfdOV0IAdx4O7PtixWKn5y2 +hMNG0zQPyUecp4pzC6kivAIhyfHilFR61RGL+GPXQ2MWZWFYbAGjyiYJnAmCP3NO +Td0jMZEnDkbUvxhMmBYSdETk1rRgm+R4LOzFUGaHqHDLKLX+FIPKcF96hrucXzcW +yLbIbEgE98OHlnVYCzRdK8jlqm8tehUc9c9WhQIBIwKCAQEA4iqWPJXtzZA68mKd +ELs4jJsdyky+ewdZeNds5tjcnHU5zUYE25K+ffJED9qUWICcLZDc81TGWjHyAqD1 +Bw7XpgUwFgeUJwUlzQurAv+/ySnxiwuaGJfhFM1CaQHzfXphgVml+fZUvnJUTvzf +TK2Lg6EdbUE9TarUlBf/xPfuEhMSlIE5keb/Zz3/LUlRg8yDqz5w+QWVJ4utnKnK +iqwZN0mwpwU7YSyJhlT4YV1F3n4YjLswM5wJs2oqm0jssQu/BT0tyEXNDYBLEF4A +sClaWuSJ2kjq7KhrrYXzagqhnSei9ODYFShJu8UWVec3Ihb5ZXlzO6vdNQ1J9Xsf +4m+2ywKBgQD6qFxx/Rv9CNN96l/4rb14HKirC2o/orApiHmHDsURs5rUKDx0f9iP +cXN7S1uePXuJRK/5hsubaOCx3Owd2u9gD6Oq0CsMkE4CUSiJcYrMANtx54cGH7Rk +EjFZxK8xAv1ldELEyxrFqkbE4BKd8QOt414qjvTGyAK+OLD3M2QdCQKBgQDtx8pN +CAxR7yhHbIWT1AH66+XWN8bXq7l3RO/ukeaci98JfkbkxURZhtxV/HHuvUhnPLdX +3TwygPBYZFNo4pzVEhzWoTtnEtrFueKxyc3+LjZpuo+mBlQ6ORtfgkr9gBVphXZG +YEzkCD3lVdl8L4cw9BVpKrJCs1c5taGjDgdInQKBgHm/fVvv96bJxc9x1tffXAcj +3OVdUN0UgXNCSaf/3A/phbeBQe9xS+3mpc4r6qvx+iy69mNBeNZ0xOitIjpjBo2+ +dBEjSBwLk5q5tJqHmy/jKMJL4n9ROlx93XS+njxgibTvU6Fp9w+NOFD/HvxB3Tcz +6+jJF85D5BNAG3DBMKBjAoGBAOAxZvgsKN+JuENXsST7F89Tck2iTcQIT8g5rwWC +P9Vt74yboe2kDT531w8+egz7nAmRBKNM751U/95P9t88EDacDI/Z2OwnuFQHCPDF +llYOUI+SpLJ6/vURRbHSnnn8a/XG+nzedGH5JGqEJNQsz+xT2axM0/W/CRknmGaJ +kda/AoGANWrLCz708y7VYgAtW2Uf1DPOIYMdvo6fxIB5i9ZfISgcJ/bbCUkFrhoH ++vq/5CIWxCPp0f85R4qxxQ5ihxJ0YDQT9Jpx4TMss4PSavPaBH3RXow5Ohe+bYoQ +NE5OgEXk2wVfZczCZpigBKbKZHNYcelXtTt/nP3rsCuGcM4h53s= +-----END RSA PRIVATE KEY----- diff --git a/tests/lang.sh b/tests/lang.sh index c0b0fc58c..cdb4174eb 100644 --- a/tests/lang.sh +++ b/tests/lang.sh @@ -2,13 +2,14 @@ source common.sh export TEST_VAR=foo # for eval-okay-getenv.nix export NIX_REMOTE=dummy:// +export NIX_STORE_DIR=/nix/store -nix-instantiate --eval -E 'builtins.trace "Hello" 123' 2>&1 | grep -q Hello +nix-instantiate --eval -E 'builtins.trace "Hello" 123' 2>&1 | grepQuiet Hello nix-instantiate --eval -E 'builtins.addErrorContext "Hello" 123' 2>&1 -nix-instantiate --trace-verbose --eval -E 'builtins.traceVerbose "Hello" 123' 2>&1 | grep -q Hello -(! nix-instantiate --eval -E 'builtins.traceVerbose "Hello" 123' 2>&1 | grep -q Hello) -(! nix-instantiate --show-trace --eval -E 'builtins.addErrorContext "Hello" 123' 2>&1 | grep -q Hello) -nix-instantiate --show-trace --eval -E 'builtins.addErrorContext "Hello" (throw "Foo")' 2>&1 | grep -q Hello +nix-instantiate --trace-verbose --eval -E 'builtins.traceVerbose "Hello" 123' 2>&1 | grepQuiet Hello +nix-instantiate --eval -E 'builtins.traceVerbose "Hello" 123' 2>&1 | grepQuietInverse Hello +nix-instantiate --show-trace --eval -E 'builtins.addErrorContext "Hello" 123' 2>&1 | grepQuietInverse Hello +expectStderr 1 nix-instantiate --show-trace --eval -E 'builtins.addErrorContext "Hello" (throw "Foo")' | grepQuiet Hello set +x @@ -50,10 +51,10 @@ for i in lang/eval-okay-*.nix; do if test -e lang/$i.flags; then flags=$(cat lang/$i.flags) fi - if ! expect 0 env NIX_PATH=lang/dir3:lang/dir4 nix-instantiate $flags --eval --strict lang/$i.nix > lang/$i.out; then + if ! expect 0 env NIX_PATH=lang/dir3:lang/dir4 HOME=/fake-home nix-instantiate $flags --eval --strict lang/$i.nix > lang/$i.out; then echo "FAIL: $i should evaluate" fail=1 - elif ! diff lang/$i.out lang/$i.exp; then + elif ! diff <(< lang/$i.out sed -e "s|$(pwd)|/pwd|g") lang/$i.exp; then echo "FAIL: evaluation result of $i not as expected" fail=1 fi diff --git a/tests/lang/eval-fail-foldlStrict-strict-op-application.nix b/tests/lang/eval-fail-foldlStrict-strict-op-application.nix new file mode 100644 index 000000000..1620cc76e --- /dev/null +++ b/tests/lang/eval-fail-foldlStrict-strict-op-application.nix @@ -0,0 +1,5 @@ +# Tests that the result of applying op is forced even if the value is never used +builtins.foldl' + (_: f: f null) + null + [ (_: throw "Not the final value, but is still forced!") (_: 23) ] diff --git a/tests/lang/eval-okay-closure.exp b/tests/lang/eval-okay-closure.exp new file mode 100644 index 000000000..e7dbf9781 --- /dev/null +++ b/tests/lang/eval-okay-closure.exp @@ -0,0 +1 @@ +[ { foo = true; key = -13; } { foo = true; key = -12; } { foo = true; key = -11; } { foo = true; key = -9; } { foo = true; key = -8; } { foo = true; key = -7; } { foo = true; key = -5; } { foo = true; key = -4; } { foo = true; key = -3; } { key = -1; } { foo = true; key = 0; } { foo = true; key = 1; } { foo = true; key = 2; } { foo = true; key = 4; } { foo = true; key = 5; } { foo = true; key = 6; } { key = 8; } { foo = true; key = 9; } { foo = true; key = 10; } { foo = true; key = 13; } { foo = true; key = 14; } { foo = true; key = 15; } { key = 17; } { foo = true; key = 18; } { foo = true; key = 19; } { foo = true; key = 22; } { foo = true; key = 23; } { key = 26; } { foo = true; key = 27; } { foo = true; key = 28; } { foo = true; key = 31; } { foo = true; key = 32; } { key = 35; } { foo = true; key = 36; } { foo = true; key = 40; } { foo = true; key = 41; } { key = 44; } { foo = true; key = 45; } { foo = true; key = 49; } { key = 53; } { foo = true; key = 54; } { foo = true; key = 58; } { key = 62; } { foo = true; key = 67; } { key = 71; } { key = 80; } ] diff --git a/tests/lang/eval-okay-context-introspection.exp b/tests/lang/eval-okay-context-introspection.exp index 27ba77dda..03b400cc8 100644 --- a/tests/lang/eval-okay-context-introspection.exp +++ b/tests/lang/eval-okay-context-introspection.exp @@ -1 +1 @@ -true +[ true true true true true true ] diff --git a/tests/lang/eval-okay-context-introspection.nix b/tests/lang/eval-okay-context-introspection.nix index 43178bd2e..50a78d946 100644 --- a/tests/lang/eval-okay-context-introspection.nix +++ b/tests/lang/eval-okay-context-introspection.nix @@ -18,7 +18,24 @@ let }; }; - legit-context = builtins.getContext "${path}${drv.outPath}${drv.foo.outPath}${drv.drvPath}"; + combo-path = "${path}${drv.outPath}${drv.foo.outPath}${drv.drvPath}"; + legit-context = builtins.getContext combo-path; - constructed-context = builtins.getContext (builtins.appendContext "" desired-context); -in legit-context == constructed-context + reconstructed-path = builtins.appendContext + (builtins.unsafeDiscardStringContext combo-path) + desired-context; + + # Eta rule for strings with context. + etaRule = str: + str == builtins.appendContext + (builtins.unsafeDiscardStringContext str) + (builtins.getContext str); + +in [ + (legit-context == desired-context) + (reconstructed-path == combo-path) + (etaRule "foo") + (etaRule drv.drvPath) + (etaRule drv.foo.outPath) + (etaRule (builtins.unsafeDiscardOutputDependency drv.drvPath)) +] diff --git a/tests/lang/eval-okay-eq.exp b/tests/lang/eval-okay-eq.exp new file mode 100644 index 000000000..27ba77dda --- /dev/null +++ b/tests/lang/eval-okay-eq.exp @@ -0,0 +1 @@ +true diff --git a/tests/lang/eval-okay-eq.exp.disabled b/tests/lang/eval-okay-eq.exp.disabled deleted file mode 100644 index 2015847b6..000000000 --- a/tests/lang/eval-okay-eq.exp.disabled +++ /dev/null @@ -1 +0,0 @@ -Bool(True) diff --git a/tests/lang/eval-okay-foldlStrict-lazy-elements.exp b/tests/lang/eval-okay-foldlStrict-lazy-elements.exp new file mode 100644 index 000000000..d81cc0710 --- /dev/null +++ b/tests/lang/eval-okay-foldlStrict-lazy-elements.exp @@ -0,0 +1 @@ +42 diff --git a/tests/lang/eval-okay-foldlStrict-lazy-elements.nix b/tests/lang/eval-okay-foldlStrict-lazy-elements.nix new file mode 100644 index 000000000..c666e07f3 --- /dev/null +++ b/tests/lang/eval-okay-foldlStrict-lazy-elements.nix @@ -0,0 +1,9 @@ +# Tests that the rhs argument of op is not forced unconditionally +let + lst = builtins.foldl' + (acc: x: acc ++ [ x ]) + [ ] + [ 42 (throw "this shouldn't be evaluated") ]; +in + +builtins.head lst diff --git a/tests/lang/eval-okay-foldlStrict-lazy-initial-accumulator.exp b/tests/lang/eval-okay-foldlStrict-lazy-initial-accumulator.exp new file mode 100644 index 000000000..d81cc0710 --- /dev/null +++ b/tests/lang/eval-okay-foldlStrict-lazy-initial-accumulator.exp @@ -0,0 +1 @@ +42 diff --git a/tests/lang/eval-okay-foldlStrict-lazy-initial-accumulator.nix b/tests/lang/eval-okay-foldlStrict-lazy-initial-accumulator.nix new file mode 100644 index 000000000..abcd5366a --- /dev/null +++ b/tests/lang/eval-okay-foldlStrict-lazy-initial-accumulator.nix @@ -0,0 +1,6 @@ +# Checks that the nul value for the accumulator is not forced unconditionally. +# Some languages provide a foldl' that is strict in this argument, but Nix does not. +builtins.foldl' + (_: x: x) + (throw "This is never forced") + [ "but the results of applying op are" 42 ] diff --git a/tests/lang/eval-okay-functionargs.exp b/tests/lang/eval-okay-functionargs.exp new file mode 100644 index 000000000..c1c9f8ffa --- /dev/null +++ b/tests/lang/eval-okay-functionargs.exp @@ -0,0 +1 @@ +[ "stdenv" "fetchurl" "aterm-stdenv" "aterm-stdenv2" "libX11" "libXv" "mplayer-stdenv2.libXv-libX11" "mplayer-stdenv2.libXv-libX11_2" "nix-stdenv-aterm-stdenv" "nix-stdenv2-aterm2-stdenv2" ] diff --git a/tests/lang/eval-okay-ind-string.nix b/tests/lang/eval-okay-ind-string.nix index 1669dc064..95d59b508 100644 --- a/tests/lang/eval-okay-ind-string.nix +++ b/tests/lang/eval-okay-ind-string.nix @@ -110,7 +110,7 @@ let And finally to interpret \n etc. as in a string: ''\n, ''\r, ''\t. ''; - # Regression test: antiquotation in '${x}' should work, but didn't. + # Regression test: string interpolation in '${x}' should work, but didn't. s15 = let x = "bla"; in '' foo '${x}' diff --git a/tests/lang/eval-okay-intersectAttrs.exp b/tests/lang/eval-okay-intersectAttrs.exp new file mode 100644 index 000000000..50445bc0e --- /dev/null +++ b/tests/lang/eval-okay-intersectAttrs.exp @@ -0,0 +1 @@ +[ { } { a = 1; } { a = 1; } { a = "a"; } { m = 1; } { m = "m"; } { n = 1; } { n = "n"; } { n = 1; p = 2; } { n = "n"; p = "p"; } { n = 1; p = 2; } { n = "n"; p = "p"; } { a = "a"; b = "b"; c = "c"; d = "d"; e = "e"; f = "f"; g = "g"; h = "h"; i = "i"; j = "j"; k = "k"; l = "l"; m = "m"; n = "n"; o = "o"; p = "p"; q = "q"; r = "r"; s = "s"; t = "t"; u = "u"; v = "v"; w = "w"; x = "x"; y = "y"; z = "z"; } true ] diff --git a/tests/lang/eval-okay-intersectAttrs.nix b/tests/lang/eval-okay-intersectAttrs.nix new file mode 100644 index 000000000..39d49938c --- /dev/null +++ b/tests/lang/eval-okay-intersectAttrs.nix @@ -0,0 +1,50 @@ +let + alphabet = + { a = "a"; + b = "b"; + c = "c"; + d = "d"; + e = "e"; + f = "f"; + g = "g"; + h = "h"; + i = "i"; + j = "j"; + k = "k"; + l = "l"; + m = "m"; + n = "n"; + o = "o"; + p = "p"; + q = "q"; + r = "r"; + s = "s"; + t = "t"; + u = "u"; + v = "v"; + w = "w"; + x = "x"; + y = "y"; + z = "z"; + }; + foo = { + inherit (alphabet) f o b a r z q u x; + aa = throw "aa"; + }; + alphabetFail = builtins.mapAttrs throw alphabet; +in +[ (builtins.intersectAttrs { a = abort "l1"; } { b = abort "r1"; }) + (builtins.intersectAttrs { a = abort "l2"; } { a = 1; }) + (builtins.intersectAttrs alphabetFail { a = 1; }) + (builtins.intersectAttrs { a = abort "laa"; } alphabet) + (builtins.intersectAttrs alphabetFail { m = 1; }) + (builtins.intersectAttrs { m = abort "lam"; } alphabet) + (builtins.intersectAttrs alphabetFail { n = 1; }) + (builtins.intersectAttrs { n = abort "lan"; } alphabet) + (builtins.intersectAttrs alphabetFail { n = 1; p = 2; }) + (builtins.intersectAttrs { n = abort "lan2"; p = abort "lap"; } alphabet) + (builtins.intersectAttrs alphabetFail { n = 1; p = 2; }) + (builtins.intersectAttrs { n = abort "lan2"; p = abort "lap"; } alphabet) + (builtins.intersectAttrs alphabetFail alphabet) + (builtins.intersectAttrs alphabet foo == builtins.intersectAttrs foo alphabet) +] diff --git a/tests/lang/eval-okay-path-antiquotation.exp b/tests/lang/eval-okay-path-antiquotation.exp new file mode 100644 index 000000000..5b8ea0243 --- /dev/null +++ b/tests/lang/eval-okay-path-antiquotation.exp @@ -0,0 +1 @@ +{ absolute = /foo; expr = /pwd/lang/foo/bar; home = /fake-home/foo; notfirst = /pwd/lang/bar/foo; simple = /pwd/lang/foo; slashes = /foo/bar; surrounded = /pwd/lang/a-foo-b; } diff --git a/tests/lang/eval-okay-path.exp b/tests/lang/eval-okay-path.exp new file mode 100644 index 000000000..3ce7f8283 --- /dev/null +++ b/tests/lang/eval-okay-path.exp @@ -0,0 +1 @@ +"/nix/store/ya937r4ydw0l6kayq8jkyqaips9c75jm-output" diff --git a/tests/lang/eval-okay-readDir.exp b/tests/lang/eval-okay-readDir.exp index bf8d2c14e..6413f6d4f 100644 --- a/tests/lang/eval-okay-readDir.exp +++ b/tests/lang/eval-okay-readDir.exp @@ -1 +1 @@ -{ bar = "regular"; foo = "directory"; } +{ bar = "regular"; foo = "directory"; ldir = "symlink"; linked = "symlink"; } diff --git a/tests/lang/eval-okay-readFileType.exp b/tests/lang/eval-okay-readFileType.exp new file mode 100644 index 000000000..6413f6d4f --- /dev/null +++ b/tests/lang/eval-okay-readFileType.exp @@ -0,0 +1 @@ +{ bar = "regular"; foo = "directory"; ldir = "symlink"; linked = "symlink"; } diff --git a/tests/lang/eval-okay-readFileType.nix b/tests/lang/eval-okay-readFileType.nix new file mode 100644 index 000000000..174fb6c3a --- /dev/null +++ b/tests/lang/eval-okay-readFileType.nix @@ -0,0 +1,6 @@ +{ + bar = builtins.readFileType ./readDir/bar; + foo = builtins.readFileType ./readDir/foo; + linked = builtins.readFileType ./readDir/linked; + ldir = builtins.readFileType ./readDir/ldir; +} diff --git a/tests/lang/eval-okay-versions.nix b/tests/lang/eval-okay-versions.nix index e63c36586..e9111f5f4 100644 --- a/tests/lang/eval-okay-versions.nix +++ b/tests/lang/eval-okay-versions.nix @@ -4,6 +4,7 @@ let name2 = "hello"; name3 = "915resolution-0.5.2"; name4 = "xf86-video-i810-1.7.4"; + name5 = "name-that-ends-with-dash--1.0"; eq = 0; lt = builtins.sub 0 1; @@ -23,6 +24,8 @@ let ((builtins.parseDrvName name3).version == "0.5.2") ((builtins.parseDrvName name4).name == "xf86-video-i810") ((builtins.parseDrvName name4).version == "1.7.4") + ((builtins.parseDrvName name5).name == "name-that-ends-with-dash") + ((builtins.parseDrvName name5).version == "-1.0") (versionTest "1.0" "2.3" lt) (versionTest "2.1" "2.3" lt) (versionTest "2.3" "2.3" eq) diff --git a/tests/lang/readDir/ldir b/tests/lang/readDir/ldir new file mode 120000 index 000000000..191028156 --- /dev/null +++ b/tests/lang/readDir/ldir @@ -0,0 +1 @@ +foo
\ No newline at end of file diff --git a/tests/lang/readDir/linked b/tests/lang/readDir/linked new file mode 120000 index 000000000..c503f86a0 --- /dev/null +++ b/tests/lang/readDir/linked @@ -0,0 +1 @@ +foo/git-hates-directories
\ No newline at end of file diff --git a/tests/linux-sandbox.sh b/tests/linux-sandbox.sh index 3f304ac2f..5a2cf7abd 100644 --- a/tests/linux-sandbox.sh +++ b/tests/linux-sandbox.sh @@ -4,13 +4,13 @@ needLocalStore "the sandbox only runs on the builder side, so it makes no sense clearStore -if ! canUseSandbox; then exit 99; fi +requireSandboxSupport # Note: we need to bind-mount $SHELL into the chroot. Currently we # only support the case where $SHELL is in the Nix store, because # otherwise things get complicated (e.g. if it's in /bin, do we need # /lib as well?). -if [[ ! $SHELL =~ /nix/store ]]; then exit 99; fi +if [[ ! $SHELL =~ /nix/store ]]; then skipTest "Shell is not from Nix store"; fi chmod -R u+w $TEST_ROOT/store0 || true rm -rf $TEST_ROOT/store0 @@ -35,5 +35,8 @@ nix-build dependencies.nix --no-out-link --check --sandbox-paths /nix/store nix-build check.nix -A nondeterministic --sandbox-paths /nix/store --no-out-link (! nix-build check.nix -A nondeterministic --sandbox-paths /nix/store --no-out-link --check -K 2> $TEST_ROOT/log) -if grep -q 'error: renaming' $TEST_ROOT/log; then false; fi -grep -q 'may not be deterministic' $TEST_ROOT/log +if grepQuiet 'error: renaming' $TEST_ROOT/log; then false; fi +grepQuiet 'may not be deterministic' $TEST_ROOT/log + +# Test that sandboxed builds cannot write to /etc easily +(! nix-build -E 'with import ./config.nix; mkDerivation { name = "etc-write"; buildCommand = "echo > /etc/test"; }' --no-out-link --sandbox-paths /nix/store) diff --git a/tests/local.mk b/tests/local.mk index 5e48ceae1..328f27e90 100644 --- a/tests/local.mk +++ b/tests/local.mk @@ -1,12 +1,19 @@ nix_tests = \ + test-infra.sh \ + init.sh \ flakes/flakes.sh \ flakes/run.sh \ flakes/mercurial.sh \ flakes/circular.sh \ flakes/init.sh \ + flakes/inputs.sh \ flakes/follow-paths.sh \ flakes/bundle.sh \ flakes/check.sh \ + flakes/unlocked-override.sh \ + flakes/absolute-paths.sh \ + flakes/build-paths.sh \ + flakes/flake-in-submodule.sh \ ca/gc.sh \ gc.sh \ remote-store.sh \ @@ -14,9 +21,11 @@ nix_tests = \ fetchMercurial.sh \ gc-auto.sh \ user-envs.sh \ + user-envs-migration.sh \ binary-cache.sh \ multiple-outputs.sh \ ca/build.sh \ + ca/new-build-cmd.sh \ nix-build.sh \ gc-concurrent.sh \ repair.sh \ @@ -90,6 +99,7 @@ nix_tests = \ fmt.sh \ eval-store.sh \ why-depends.sh \ + ca/why-depends.sh \ import-derivation.sh \ ca/import-derivation.sh \ nix_path.sh \ @@ -98,6 +108,8 @@ nix_tests = \ ssh-relay.sh \ plugins.sh \ build.sh \ + build-delete.sh \ + output-normalization.sh \ ca/nix-run.sh \ selfref-gc.sh ca/selfref-gc.sh \ db-migration.sh \ @@ -109,7 +121,10 @@ nix_tests = \ store-ping.sh \ fetchClosure.sh \ completions.sh \ - impure-derivations.sh + flakes/show.sh \ + impure-derivations.sh \ + path-from-hash-part.sh \ + toString-path.sh ifeq ($(HAVE_LIBCPUID), 1) nix_tests += compute-levels.sh @@ -117,11 +132,9 @@ endif install-tests += $(foreach x, $(nix_tests), tests/$(x)) -tests-environment = NIX_REMOTE= $(bash) -e +clean-files += $(d)/common/vars-and-functions.sh $(d)/config.nix $(d)/ca/config.nix -clean-files += $(d)/common.sh $(d)/config.nix $(d)/ca/config.nix - -test-deps += tests/common.sh tests/config.nix tests/ca/config.nix +test-deps += tests/common/vars-and-functions.sh tests/config.nix tests/ca/config.nix ifeq ($(BUILD_SHARED_LIBS), 1) test-deps += tests/plugins/libplugintest.$(SO_EXT) diff --git a/tests/misc.sh b/tests/misc.sh index 2830856ae..60d58310e 100644 --- a/tests/misc.sh +++ b/tests/misc.sh @@ -3,17 +3,17 @@ source common.sh # Tests miscellaneous commands. # Do all commands have help? -#nix-env --help | grep -q install -#nix-store --help | grep -q realise -#nix-instantiate --help | grep -q eval -#nix-hash --help | grep -q base32 +#nix-env --help | grepQuiet install +#nix-store --help | grepQuiet realise +#nix-instantiate --help | grepQuiet eval +#nix-hash --help | grepQuiet base32 # Can we ask for the version number? nix-env --version | grep "$version" # Usage errors. -nix-env --foo 2>&1 | grep "no operation" -nix-env -q --foo 2>&1 | grep "unknown flag" +expect 1 nix-env --foo 2>&1 | grep "no operation" +expect 1 nix-env -q --foo 2>&1 | grep "unknown flag" # Eval Errors. eval_arg_res=$(nix-instantiate --eval -E 'let a = {} // a; in a.foo' 2>&1 || true) diff --git a/tests/multiple-outputs.nix b/tests/multiple-outputs.nix index 624a5dade..413d392e4 100644 --- a/tests/multiple-outputs.nix +++ b/tests/multiple-outputs.nix @@ -31,6 +31,15 @@ rec { helloString = "Hello, world!"; }; + use-a = mkDerivation { + name = "use-a"; + inherit (a) first second; + builder = builtins.toFile "builder.sh" + '' + cat $first/file $second/file >$out + ''; + }; + b = mkDerivation { defaultOutput = assert a.second.helloString == "Hello, world!"; a; firstOutput = assert a.outputName == "first"; a.first.first; @@ -82,9 +91,40 @@ rec { e = mkDerivation { name = "multiple-outputs-e"; - outputs = [ "a" "b" "c" ]; - meta.outputsToInstall = [ "a" "b" ]; - buildCommand = "mkdir $a $b $c"; + outputs = [ "a_a" "b" "c" ]; + meta.outputsToInstall = [ "a_a" "b" ]; + buildCommand = "mkdir $a_a $b $c"; + }; + + independent = mkDerivation { + name = "multiple-outputs-independent"; + outputs = [ "first" "second" ]; + builder = builtins.toFile "builder.sh" + '' + mkdir $first $second + test -z $all + echo "first" > $first/file + echo "second" > $second/file + ''; + }; + + use-independent = mkDerivation { + name = "use-independent"; + inherit (a) first second; + builder = builtins.toFile "builder.sh" + '' + cat $first/file $second/file >$out + ''; + }; + + invalid-output-name-1 = mkDerivation { + name = "invalid-output-name-1"; + outputs = [ "out/"]; + }; + + invalid-output-name-2 = mkDerivation { + name = "invalid-output-name-2"; + outputs = [ "x" "foo$"]; }; } diff --git a/tests/multiple-outputs.sh b/tests/multiple-outputs.sh index 0d45ad35b..330600d08 100644 --- a/tests/multiple-outputs.sh +++ b/tests/multiple-outputs.sh @@ -19,8 +19,8 @@ echo "evaluating c..." # outputs. drvPath=$(nix-instantiate multiple-outputs.nix -A c) #[ "$drvPath" = "$drvPath2" ] -grep -q 'multiple-outputs-a.drv",\["first","second"\]' $drvPath -grep -q 'multiple-outputs-b.drv",\["out"\]' $drvPath +grepQuiet 'multiple-outputs-a.drv",\["first","second"\]' $drvPath +grepQuiet 'multiple-outputs-b.drv",\["out"\]' $drvPath # While we're at it, test the ‘unsafeDiscardOutputDependency’ primop. outPath=$(nix-build multiple-outputs.nix -A d --no-out-link) @@ -83,3 +83,6 @@ nix-store --gc --keep-derivations --keep-outputs nix-store --gc --print-roots rm -rf $NIX_STORE_DIR/.links rmdir $NIX_STORE_DIR + +expect 1 nix build -f multiple-outputs.nix invalid-output-name-1 2>&1 | grep 'contains illegal character' +expect 1 nix build -f multiple-outputs.nix invalid-output-name-2 2>&1 | grep 'contains illegal character' diff --git a/tests/nar-access.sh b/tests/nar-access.sh index dcc2e8a36..d487d58d2 100644 --- a/tests/nar-access.sh +++ b/tests/nar-access.sh @@ -46,8 +46,8 @@ diff -u \ <(echo '{"type":"regular","size":0}' | jq -S) # Test missing files. -nix store ls --json -R $storePath/xyzzy 2>&1 | grep 'does not exist in NAR' -nix store ls $storePath/xyzzy 2>&1 | grep 'does not exist' +expect 1 nix store ls --json -R $storePath/xyzzy 2>&1 | grep 'does not exist in NAR' +expect 1 nix store ls $storePath/xyzzy 2>&1 | grep 'does not exist' # Test failure to dump. if nix-store --dump $storePath >/dev/full ; then diff --git a/tests/nix-channel.sh b/tests/nix-channel.sh index 54b8f5979..dbb3114f1 100644 --- a/tests/nix-channel.sh +++ b/tests/nix-channel.sh @@ -6,12 +6,25 @@ rm -f $TEST_HOME/.nix-channels $TEST_HOME/.nix-profile # Test add/list/remove. nix-channel --add http://foo/bar xyzzy -nix-channel --list | grep -q http://foo/bar +nix-channel --list | grepQuiet http://foo/bar nix-channel --remove xyzzy [ -e $TEST_HOME/.nix-channels ] [ "$(cat $TEST_HOME/.nix-channels)" = '' ] +# Test the XDG Base Directories support + +export NIX_CONFIG="use-xdg-base-directories = true" + +nix-channel --add http://foo/bar xyzzy +nix-channel --list | grepQuiet http://foo/bar +nix-channel --remove xyzzy + +unset NIX_CONFIG + +[ -e $TEST_HOME/.local/state/nix/channels ] +[ "$(cat $TEST_HOME/.local/state/nix/channels)" = '' ] + # Create a channel. rm -rf $TEST_ROOT/foo mkdir -p $TEST_ROOT/foo @@ -28,8 +41,8 @@ nix-channel --update # Do a query. nix-env -qa \* --meta --xml --out-path > $TEST_ROOT/meta.xml -grep -q 'meta.*description.*Random test package' $TEST_ROOT/meta.xml -grep -q 'item.*attrPath="foo".*name="dependencies-top"' $TEST_ROOT/meta.xml +grepQuiet 'meta.*description.*Random test package' $TEST_ROOT/meta.xml +grepQuiet 'item.*attrPath="foo".*name="dependencies-top"' $TEST_ROOT/meta.xml # Do an install. nix-env -i dependencies-top @@ -41,9 +54,9 @@ nix-channel --update # Do a query. nix-env -qa \* --meta --xml --out-path > $TEST_ROOT/meta.xml -grep -q 'meta.*description.*Random test package' $TEST_ROOT/meta.xml -grep -q 'item.*attrPath="bar".*name="dependencies-top"' $TEST_ROOT/meta.xml -grep -q 'item.*attrPath="foo".*name="dependencies-top"' $TEST_ROOT/meta.xml +grepQuiet 'meta.*description.*Random test package' $TEST_ROOT/meta.xml +grepQuiet 'item.*attrPath="bar".*name="dependencies-top"' $TEST_ROOT/meta.xml +grepQuiet 'item.*attrPath="foo".*name="dependencies-top"' $TEST_ROOT/meta.xml # Do an install. nix-env -i dependencies-top diff --git a/tests/nix-profile.sh b/tests/nix-profile.sh index 7ba3235fa..652e8a8f2 100644 --- a/tests/nix-profile.sh +++ b/tests/nix-profile.sh @@ -56,6 +56,14 @@ nix profile history nix profile history | grep "packages.$system.default: ∅ -> 1.0" nix profile diff-closures | grep 'env-manifest.nix: ε → ∅' +# Test XDG Base Directories support + +export NIX_CONFIG="use-xdg-base-directories = true" +nix profile remove 1 +nix profile install $flake1Dir +[[ $($TEST_HOME/.local/state/nix/profile/bin/hello) = "Hello World" ]] +unset NIX_CONFIG + # Test upgrading a package. printf NixOS > $flake1Dir/who printf 2.0 > $flake1Dir/version @@ -132,6 +140,36 @@ printf World2 > $flake2Dir/who nix profile install $flake1Dir [[ $($TEST_HOME/.nix-profile/bin/hello) = "Hello World" ]] +expect 1 nix profile install $flake2Dir +diff -u <( + nix --offline profile install $flake2Dir 2>&1 1> /dev/null \ + | grep -vE "^warning: " \ + || true +) <(cat << EOF +error: An existing package already provides the following file: + + $(nix build --no-link --print-out-paths ${flake1Dir}"#default.out")/bin/hello + + This is the conflicting file from the new package: + + $(nix build --no-link --print-out-paths ${flake2Dir}"#default.out")/bin/hello + + To remove the existing package: + + nix profile remove path:${flake1Dir} + + The new package can also be installed next to the existing one by assigning a different priority. + The conflicting packages have a priority of 5. + To prioritise the new package: + + nix profile install path:${flake2Dir} --priority 4 + + To prioritise the existing package: + + nix profile install path:${flake2Dir} --priority 6 +EOF +) +[[ $($TEST_HOME/.nix-profile/bin/hello) = "Hello World" ]] nix profile install $flake2Dir --priority 100 [[ $($TEST_HOME/.nix-profile/bin/hello) = "Hello World" ]] nix profile install $flake2Dir --priority 0 diff --git a/tests/nix-shell.sh b/tests/nix-shell.sh index f291c6f79..044b96d54 100644 --- a/tests/nix-shell.sh +++ b/tests/nix-shell.sh @@ -88,20 +88,43 @@ output=$($TEST_ROOT/spaced\ \\\'\"shell.shebang.rb abc ruby) nix develop -f "$shellDotNix" shellDrv -c bash -c '[[ -n $stdenv ]]' # Ensure `nix develop -c` preserves stdin -echo foo | nix develop -f "$shellDotNix" shellDrv -c cat | grep -q foo +echo foo | nix develop -f "$shellDotNix" shellDrv -c cat | grepQuiet foo # Ensure `nix develop -c` actually executes the command if stdout isn't a terminal -nix develop -f "$shellDotNix" shellDrv -c echo foo |& grep -q foo +nix develop -f "$shellDotNix" shellDrv -c echo foo |& grepQuiet foo # Test 'nix print-dev-env'. -[[ $(nix print-dev-env -f "$shellDotNix" shellDrv --json | jq -r .variables.arr1.value[2]) = '3 4' ]] - -source <(nix print-dev-env -f "$shellDotNix" shellDrv) -[[ -n $stdenv ]] -[[ ${arr1[2]} = "3 4" ]] -[[ ${arr2[1]} = $'\n' ]] -[[ ${arr2[2]} = $'x\ny' ]] -[[ $(fun) = blabla ]] + +nix print-dev-env -f "$shellDotNix" shellDrv > $TEST_ROOT/dev-env.sh +nix print-dev-env -f "$shellDotNix" shellDrv --json > $TEST_ROOT/dev-env.json + +# Ensure `nix print-dev-env --json` contains variable assignments. +[[ $(jq -r .variables.arr1.value[2] $TEST_ROOT/dev-env.json) = '3 4' ]] + +# Run tests involving `source <(nix print-dev-inv)` in subshells to avoid modifying the current +# environment. + +set +u # FIXME: Make print-dev-env `set -u` compliant (issue #7951) + +# Ensure `source <(nix print-dev-env)` modifies the environment. +( + path=$PATH + source $TEST_ROOT/dev-env.sh + [[ -n $stdenv ]] + [[ ${arr1[2]} = "3 4" ]] + [[ ${arr2[1]} = $'\n' ]] + [[ ${arr2[2]} = $'x\ny' ]] + [[ $(fun) = blabla ]] + [[ $PATH = $(jq -r .variables.PATH.value $TEST_ROOT/dev-env.json):$path ]] +) + +# Ensure `source <(nix print-dev-env)` handles the case when PATH is empty. +( + path=$PATH + PATH= + source $TEST_ROOT/dev-env.sh + [[ $PATH = $(PATH=$path jq -r .variables.PATH.value $TEST_ROOT/dev-env.json) ]] +) # Test nix-shell with ellipsis and no `inNixShell` argument (for backwards compat with old nixpkgs) cat >$TEST_ROOT/shell-ellipsis.nix <<EOF diff --git a/tests/nix_path.sh b/tests/nix_path.sh index d3657abf0..2b222b4a1 100644 --- a/tests/nix_path.sh +++ b/tests/nix_path.sh @@ -9,3 +9,6 @@ nix-instantiate --eval -E '<by-relative-path/simple.nix>' --restrict-eval # Should ideally also test this, but there’s no pure way to do it, so just trust me that it works # nix-instantiate --eval -E '<nixpkgs>' -I nixpkgs=channel:nixos-unstable --restrict-eval + +[[ $(nix-instantiate --find-file by-absolute-path/simple.nix) = $PWD/simple.nix ]] +[[ $(nix-instantiate --find-file by-relative-path/simple.nix) = $PWD/simple.nix ]] diff --git a/tests/nixos/authorization.nix b/tests/nixos/authorization.nix new file mode 100644 index 000000000..7e8744dd9 --- /dev/null +++ b/tests/nixos/authorization.nix @@ -0,0 +1,79 @@ +{ + name = "authorization"; + + nodes.machine = { + virtualisation.writableStore = true; + # TODO add a test without allowed-users setting. allowed-users is uncommon among NixOS users. + nix.settings.allowed-users = ["alice" "bob"]; + nix.settings.trusted-users = ["alice"]; + + users.users.alice.isNormalUser = true; + users.users.bob.isNormalUser = true; + users.users.mallory.isNormalUser = true; + + nix.settings.experimental-features = "nix-command"; + }; + + testScript = + let + pathFour = "/nix/store/20xfy868aiic0r0flgzq4n5dq1yvmxkn-four"; + in + '' + machine.wait_for_unit("multi-user.target") + machine.succeed(""" + exec 1>&2 + echo kSELDhobKaF8/VdxIxdP7EQe+Q > one + diff $(nix store add-file one) one + """) + machine.succeed(""" + su --login alice -c ' + set -x + cd ~ + echo ehHtmfuULXYyBV6NBk6QUi8iE0 > two + ls + diff $(echo $(nix store add-file two)) two' 1>&2 + """) + machine.succeed(""" + su --login bob -c ' + set -x + cd ~ + echo 0Jw8RNp7cK0W2AdNbcquofcOVk > three + diff $(nix store add-file three) three + ' 1>&2 + """) + + # We're going to check that a path is not created + machine.succeed(""" + ! [[ -e ${pathFour} ]] + """) + machine.succeed(""" + su --login mallory -c ' + set -x + cd ~ + echo 5mgtDj0ohrWkT50TLR0f4tIIxY > four; + (! nix store add-file four 2>&1) | grep -F "cannot open connection to remote store" + (! nix store add-file four 2>&1) | grep -F "Connection reset by peer" + ! [[ -e ${pathFour} ]] + ' 1>&2 + """) + + # Check that the file _can_ be added, and matches the expected path we were checking + machine.succeed(""" + exec 1>&2 + echo 5mgtDj0ohrWkT50TLR0f4tIIxY > four + four="$(nix store add-file four)" + diff $four four + diff <(echo $four) <(echo ${pathFour}) + """) + + machine.succeed(""" + su --login alice -c 'nix-store --verify --repair' + """) + + machine.succeed(""" + set -x + su --login bob -c '(! nix-store --verify --repair 2>&1)' | tee diag 1>&2 + grep -F "you are not privileged to repair paths" diag + """) + ''; +} diff --git a/tests/nixos/containers/containers.nix b/tests/nixos/containers/containers.nix new file mode 100644 index 000000000..c8ee78a4a --- /dev/null +++ b/tests/nixos/containers/containers.nix @@ -0,0 +1,63 @@ +# Test whether we can run a NixOS container inside a Nix build using systemd-nspawn. +{ lib, nixpkgs, ... }: + +{ + name = "containers"; + + nodes = + { + host = + { config, lib, pkgs, nodes, ... }: + { virtualisation.writableStore = true; + virtualisation.diskSize = 2048; + virtualisation.additionalPaths = + [ pkgs.stdenvNoCC + (import ./systemd-nspawn.nix { inherit nixpkgs; }).toplevel + ]; + virtualisation.memorySize = 4096; + nix.settings.substituters = lib.mkForce [ ]; + nix.extraOptions = + '' + extra-experimental-features = nix-command auto-allocate-uids cgroups + extra-system-features = uid-range + ''; + nix.nixPath = [ "nixpkgs=${nixpkgs}" ]; + }; + }; + + testScript = { nodes }: '' + start_all() + + host.succeed("nix --version >&2") + + # Test that 'id' gives the expected result in various configurations. + + # Existing UIDs, sandbox. + host.succeed("nix build -v --no-auto-allocate-uids --sandbox -L --offline --impure --file ${./id-test.nix} --argstr name id-test-1") + host.succeed("[[ $(cat ./result) = 'uid=1000(nixbld) gid=100(nixbld) groups=100(nixbld)' ]]") + + # Existing UIDs, no sandbox. + host.succeed("nix build -v --no-auto-allocate-uids --no-sandbox -L --offline --impure --file ${./id-test.nix} --argstr name id-test-2") + host.succeed("[[ $(cat ./result) = 'uid=30001(nixbld1) gid=30000(nixbld) groups=30000(nixbld)' ]]") + + # Auto-allocated UIDs, sandbox. + host.succeed("nix build -v --auto-allocate-uids --sandbox -L --offline --impure --file ${./id-test.nix} --argstr name id-test-3") + host.succeed("[[ $(cat ./result) = 'uid=1000(nixbld) gid=100(nixbld) groups=100(nixbld)' ]]") + + # Auto-allocated UIDs, no sandbox. + host.succeed("nix build -v --auto-allocate-uids --no-sandbox -L --offline --impure --file ${./id-test.nix} --argstr name id-test-4") + host.succeed("[[ $(cat ./result) = 'uid=872415232 gid=30000(nixbld) groups=30000(nixbld)' ]]") + + # Auto-allocated UIDs, UID range, sandbox. + host.succeed("nix build -v --auto-allocate-uids --sandbox -L --offline --impure --file ${./id-test.nix} --argstr name id-test-5 --arg uidRange true") + host.succeed("[[ $(cat ./result) = 'uid=0(root) gid=0(root) groups=0(root)' ]]") + + # Auto-allocated UIDs, UID range, no sandbox. + host.fail("nix build -v --auto-allocate-uids --no-sandbox -L --offline --impure --file ${./id-test.nix} --argstr name id-test-6 --arg uidRange true") + + # Run systemd-nspawn in a Nix build. + host.succeed("nix build -v --auto-allocate-uids --sandbox -L --offline --impure --file ${./systemd-nspawn.nix} --argstr nixpkgs ${nixpkgs}") + host.succeed("[[ $(cat ./result/msg) = 'Hello World' ]]") + ''; + +} diff --git a/tests/nixos/containers/id-test.nix b/tests/nixos/containers/id-test.nix new file mode 100644 index 000000000..8eb9d38f9 --- /dev/null +++ b/tests/nixos/containers/id-test.nix @@ -0,0 +1,8 @@ +{ name, uidRange ? false }: + +with import <nixpkgs> {}; + +runCommand name + { requiredSystemFeatures = if uidRange then ["uid-range"] else []; + } + "id; id > $out" diff --git a/tests/nixos/containers/systemd-nspawn.nix b/tests/nixos/containers/systemd-nspawn.nix new file mode 100644 index 000000000..f54f32f2a --- /dev/null +++ b/tests/nixos/containers/systemd-nspawn.nix @@ -0,0 +1,78 @@ +{ nixpkgs }: + +let + + machine = { config, pkgs, ... }: + { + system.stateVersion = "22.05"; + boot.isContainer = true; + systemd.services.console-getty.enable = false; + networking.dhcpcd.enable = false; + + services.httpd = { + enable = true; + adminAddr = "nixos@example.org"; + }; + + systemd.services.test = { + wantedBy = [ "multi-user.target" ]; + after = [ "httpd.service" ]; + script = '' + source /.env + echo "Hello World" > $out/msg + ls -lR /dev > $out/dev + ${pkgs.curl}/bin/curl -sS --fail http://localhost/ > $out/page.html + ''; + unitConfig = { + FailureAction = "exit-force"; + FailureActionExitStatus = 42; + SuccessAction = "exit-force"; + }; + }; + }; + + cfg = (import (nixpkgs + "/nixos/lib/eval-config.nix") { + modules = [ machine ]; + system = "x86_64-linux"; + }); + + config = cfg.config; + +in + +with cfg._module.args.pkgs; + +runCommand "test" + { buildInputs = [ config.system.path ]; + requiredSystemFeatures = [ "uid-range" ]; + toplevel = config.system.build.toplevel; + } + '' + root=$(pwd)/root + mkdir -p $root $root/etc + + export > $root/.env + + # Make /run a tmpfs to shut up a systemd warning. + mkdir /run + mount -t tmpfs none /run + + mount -t cgroup2 none /sys/fs/cgroup + + mkdir -p $out + + chmod +w /etc + touch /etc/os-release + echo a5ea3f98dedc0278b6f3cc8c37eeaeac > /etc/machine-id + + SYSTEMD_NSPAWN_UNIFIED_HIERARCHY=1 \ + ${config.systemd.package}/bin/systemd-nspawn \ + --keep-unit \ + -M ${config.networking.hostName} -D "$root" \ + --register=no \ + --resolv-conf=off \ + --bind-ro=/nix/store \ + --bind=$out \ + --private-network \ + $toplevel/init + '' diff --git a/tests/github-flakes.nix b/tests/nixos/github-flakes.nix index fc481c7e3..e4d347691 100644 --- a/tests/github-flakes.nix +++ b/tests/nixos/github-flakes.nix @@ -1,14 +1,9 @@ -{ nixpkgs, system, overlay }: - -with import (nixpkgs + "/nixos/lib/testing-python.nix") { - inherit system; - extraConfigurations = [ { nixpkgs.overlays = [ overlay ]; } ]; -}; - +{ lib, config, nixpkgs, ... }: let + pkgs = config.nodes.client.nixpkgs.pkgs; - # Generate a fake root CA and a fake api.github.com / channels.nixos.org certificate. - cert = pkgs.runCommand "cert" { buildInputs = [ pkgs.openssl ]; } + # Generate a fake root CA and a fake api.github.com / github.com / channels.nixos.org certificate. + cert = pkgs.runCommand "cert" { nativeBuildInputs = [ pkgs.openssl ]; } '' mkdir -p $out @@ -18,7 +13,7 @@ let openssl req -newkey rsa:2048 -nodes -keyout $out/server.key \ -subj "/C=CN/ST=Denial/L=Springfield/O=Dis/CN=github.com" -out server.csr - openssl x509 -req -extfile <(printf "subjectAltName=DNS:api.github.com,DNS:channels.nixos.org") \ + openssl x509 -req -extfile <(printf "subjectAltName=DNS:api.github.com,DNS:github.com,DNS:channels.nixos.org") \ -days 36500 -in server.csr -CA $out/ca.crt -CAkey ca.key -CAcreateserial -out $out/server.crt ''; @@ -37,6 +32,17 @@ let "owner": "NixOS", "repo": "nixpkgs" } + }, + { + "from": { + "type": "indirect", + "id": "private-flake" + }, + "to": { + "type": "github", + "owner": "fancy-enterprise", + "repo": "private-flake" + } } ], "version": 2 @@ -45,23 +51,41 @@ let destination = "/flake-registry.json"; }; - api = pkgs.runCommand "nixpkgs-flake" {} + private-flake-rev = "9f1dd0df5b54a7dc75b618034482ed42ce34383d"; + + private-flake-api = pkgs.runCommand "private-flake" {} '' - mkdir -p $out/tarball + mkdir -p $out/{commits,tarball} - dir=NixOS-nixpkgs-${nixpkgs.shortRev} - cp -prd ${nixpkgs} $dir - # Set the correct timestamp in the tarball. - find $dir -print0 | xargs -0 touch -t ${builtins.substring 0 12 nixpkgs.lastModifiedDate}.${builtins.substring 12 2 nixpkgs.lastModifiedDate} -- - tar cfz $out/tarball/${nixpkgs.rev} $dir --hard-dereference + # Setup https://docs.github.com/en/rest/commits/commits#get-a-commit + echo '{"sha": "${private-flake-rev}"}' > $out/commits/HEAD + # Setup tarball download via API + dir=private-flake + mkdir $dir + echo '{ outputs = {...}: {}; }' > $dir/flake.nix + tar cfz $out/tarball/${private-flake-rev} $dir --hard-dereference + ''; + + nixpkgs-api = pkgs.runCommand "nixpkgs-flake" {} + '' mkdir -p $out/commits + + # Setup https://docs.github.com/en/rest/commits/commits#get-a-commit echo '{"sha": "${nixpkgs.rev}"}' > $out/commits/HEAD ''; -in + archive = pkgs.runCommand "nixpkgs-flake" {} + '' + mkdir -p $out/archive -makeTest ( + dir=NixOS-nixpkgs-${nixpkgs.shortRev} + cp -prd ${nixpkgs} $dir + # Set the correct timestamp in the tarball. + find $dir -print0 | xargs -0 touch -t ${builtins.substring 0 12 nixpkgs.lastModifiedDate}.${builtins.substring 12 2 nixpkgs.lastModifiedDate} -- + tar cfz $out/archive/${nixpkgs.rev}.tar.gz $dir --hard-dereference + ''; +in { name = "github-flakes"; @@ -93,7 +117,20 @@ makeTest ( sslServerCert = "${cert}/server.crt"; servedDirs = [ { urlPath = "/repos/NixOS/nixpkgs"; - dir = api; + dir = nixpkgs-api; + } + { urlPath = "/repos/fancy-enterprise/private-flake"; + dir = private-flake-api; + } + ]; + }; + services.httpd.virtualHosts."github.com" = + { forceSSL = true; + sslServerKey = "${cert}/server.key"; + sslServerCert = "${cert}/server.crt"; + servedDirs = + [ { urlPath = "/NixOS/nixpkgs"; + dir = archive; } ]; }; @@ -105,11 +142,10 @@ makeTest ( virtualisation.diskSize = 2048; virtualisation.additionalPaths = [ pkgs.hello pkgs.fuse ]; virtualisation.memorySize = 4096; - nix.binaryCaches = lib.mkForce [ ]; + nix.settings.substituters = lib.mkForce [ ]; nix.extraOptions = "experimental-features = nix-command flakes"; - environment.systemPackages = [ pkgs.jq ]; networking.hosts.${(builtins.head nodes.github.config.networking.interfaces.eth1.ipv4.addresses).address} = - [ "channels.nixos.org" "api.github.com" ]; + [ "channels.nixos.org" "api.github.com" "github.com" ]; security.pki.certificateFiles = [ "${cert}/ca.crt" ]; }; }; @@ -121,22 +157,39 @@ makeTest ( start_all() - github.wait_for_unit("httpd.service") + def cat_log(): + github.succeed("cat /var/log/httpd/*.log >&2") - client.succeed("curl -v https://api.github.com/ >&2") - client.succeed("nix registry list | grep nixpkgs") + github.wait_for_unit("httpd.service") - rev = client.succeed("nix flake info nixpkgs --json | jq -r .revision") - assert rev.strip() == "${nixpkgs.rev}", "revision mismatch" + client.succeed("curl -v https://github.com/ >&2") + out = client.succeed("nix registry list") + print(out) + assert "github:NixOS/nixpkgs" in out, "nixpkgs flake not found" + assert "github:fancy-enterprise/private-flake" in out, "private flake not found" + cat_log() + + # If no github access token is provided, nix should use the public archive url... + out = client.succeed("nix flake metadata nixpkgs --json") + print(out) + info = json.loads(out) + assert info["revision"] == "${nixpkgs.rev}", f"revision mismatch: {info['revision']} != ${nixpkgs.rev}" + cat_log() + + # ... otherwise it should use the API + out = client.succeed("nix flake metadata private-flake --json --access-tokens github.com=ghp_000000000000000000000000000000000000 --tarball-ttl 0") + print(out) + info = json.loads(out) + assert info["revision"] == "${private-flake-rev}", f"revision mismatch: {info['revision']} != ${private-flake-rev}" + cat_log() client.succeed("nix registry pin nixpkgs") - - client.succeed("nix flake info nixpkgs --tarball-ttl 0 >&2") + client.succeed("nix flake metadata nixpkgs --tarball-ttl 0 >&2") # Shut down the web server. The flake should be cached on the client. github.succeed("systemctl stop httpd.service") - info = json.loads(client.succeed("nix flake info nixpkgs --json")) + info = json.loads(client.succeed("nix flake metadata nixpkgs --json")) date = time.strftime("%Y%m%d%H%M%S", time.gmtime(info['lastModified'])) assert date == "${nixpkgs.lastModifiedDate}", "time mismatch" @@ -147,4 +200,4 @@ makeTest ( client.succeed("nix build nixpkgs#fuse --tarball-ttl 0") ''; -}) +} diff --git a/tests/nix-copy-closure.nix b/tests/nixos/nix-copy-closure.nix index ba8b2cfc9..66cbfb033 100644 --- a/tests/nix-copy-closure.nix +++ b/tests/nixos/nix-copy-closure.nix @@ -1,13 +1,16 @@ # Test ‘nix-copy-closure’. -{ nixpkgs, system, overlay }: +{ lib, config, nixpkgs, hostPkgs, ... }: -with import (nixpkgs + "/nixos/lib/testing-python.nix") { - inherit system; - extraConfigurations = [ { nixpkgs.overlays = [ overlay ]; } ]; -}; +let + pkgs = config.nodes.client.nixpkgs.pkgs; -makeTest (let pkgA = pkgs.cowsay; pkgB = pkgs.wget; pkgC = pkgs.hello; pkgD = pkgs.tmux; in { + pkgA = pkgs.cowsay; + pkgB = pkgs.wget; + pkgC = pkgs.hello; + pkgD = pkgs.tmux; + +in { name = "nix-copy-closure"; nodes = @@ -15,7 +18,7 @@ makeTest (let pkgA = pkgs.cowsay; pkgB = pkgs.wget; pkgC = pkgs.hello; pkgD = pk { config, lib, pkgs, ... }: { virtualisation.writableStore = true; virtualisation.additionalPaths = [ pkgA pkgD.drvPath ]; - nix.binaryCaches = lib.mkForce [ ]; + nix.settings.substituters = lib.mkForce [ ]; }; server = @@ -74,4 +77,4 @@ makeTest (let pkgA = pkgs.cowsay; pkgB = pkgs.wget; pkgC = pkgs.hello; pkgD = pk # ) # client.succeed("nix-store --check-validity ${pkgC}") ''; -}) +} diff --git a/tests/nss-preload.nix b/tests/nixos/nss-preload.nix index 64b655ba2..cef62e95b 100644 --- a/tests/nss-preload.nix +++ b/tests/nixos/nss-preload.nix @@ -1,11 +1,9 @@ -{ nixpkgs, system, overlay }: - -with import (nixpkgs + "/nixos/lib/testing-python.nix") { - inherit system; - extraConfigurations = [ { nixpkgs.overlays = [ overlay ]; } ]; -}; +{ lib, config, nixpkgs, ... }: let + + pkgs = config.nodes.client.nixpkgs.pkgs; + nix-fetch = pkgs.writeText "fetch.nix" '' derivation { # This derivation is an copy from what is available over at @@ -41,9 +39,7 @@ let ''; in -makeTest ( - -rec { +{ name = "nss-preload"; nodes = { @@ -98,9 +94,9 @@ rec { { address = "192.168.0.10"; prefixLength = 24; } ]; - nix.sandboxPaths = lib.mkForce []; - nix.binaryCaches = lib.mkForce []; - nix.useSandbox = lib.mkForce true; + nix.settings.extra-sandbox-paths = lib.mkForce []; + nix.settings.substituters = lib.mkForce []; + nix.settings.sandbox = lib.mkForce true; }; }; @@ -122,4 +118,4 @@ rec { nix-build ${nix-fetch} >&2 """) ''; -}) +} diff --git a/tests/remote-builds.nix b/tests/nixos/remote-builds.nix index 7b2e6f708..1c96cc787 100644 --- a/tests/remote-builds.nix +++ b/tests/nixos/remote-builds.nix @@ -1,22 +1,21 @@ # Test Nix's remote build feature. -{ nixpkgs, system, overlay }: - -with import (nixpkgs + "/nixos/lib/testing-python.nix") { - inherit system; - extraConfigurations = [ { nixpkgs.overlays = [ overlay ]; } ]; -}; - -makeTest ( +{ config, lib, hostPkgs, ... }: let + pkgs = config.nodes.client.nixpkgs.pkgs; # The configuration of the remote builders. builder = { config, pkgs, ... }: { services.openssh.enable = true; virtualisation.writableStore = true; - nix.useSandbox = true; + nix.settings.sandbox = true; + + # Regression test for use of PID namespaces when /proc has + # filesystems mounted on top of it + # (i.e. /proc/sys/fs/binfmt_misc). + boot.binfmt.emulatedSystems = [ "aarch64-linux" ]; }; # Trivial Nix expression to build remotely. @@ -44,7 +43,7 @@ in client = { config, lib, pkgs, ... }: - { nix.maxJobs = 0; # force remote building + { nix.settings.max-jobs = 0; # force remote building nix.distributedBuilds = true; nix.buildMachines = [ { hostName = "builder1"; @@ -62,7 +61,7 @@ in ]; virtualisation.writableStore = true; virtualisation.additionalPaths = [ config.system.build.extraUtils ]; - nix.binaryCaches = lib.mkForce [ ]; + nix.settings.substituters = lib.mkForce [ ]; programs.ssh.extraConfig = "ConnectTimeout 30"; }; }; @@ -75,7 +74,7 @@ in # Create an SSH key on the client. subprocess.run([ - "${pkgs.openssh}/bin/ssh-keygen", "-t", "ed25519", "-f", "key", "-N", "" + "${hostPkgs.openssh}/bin/ssh-keygen", "-t", "ed25519", "-f", "key", "-N", "" ], capture_output=True, check=True) client.succeed("mkdir -p -m 700 /root/.ssh") client.copy_from_host("key", "/root/.ssh/id_ed25519") @@ -109,4 +108,4 @@ in builder1.block() client.succeed("nix-build ${expr nodes.client.config 4}") ''; -}) +} diff --git a/tests/setuid.nix b/tests/nixos/setuid.nix index a83b1fc3a..2b66320dd 100644 --- a/tests/setuid.nix +++ b/tests/nixos/setuid.nix @@ -1,21 +1,20 @@ # Verify that Linux builds cannot create setuid or setgid binaries. -{ nixpkgs, system, overlay }: +{ lib, config, nixpkgs, ... }: -with import (nixpkgs + "/nixos/lib/testing-python.nix") { - inherit system; - extraConfigurations = [ { nixpkgs.overlays = [ overlay ]; } ]; -}; +let + pkgs = config.nodes.machine.nixpkgs.pkgs; -makeTest { +in +{ name = "setuid"; nodes.machine = { config, lib, pkgs, ... }: { virtualisation.writableStore = true; - nix.binaryCaches = lib.mkForce [ ]; + nix.settings.substituters = lib.mkForce [ ]; nix.nixPath = [ "nixpkgs=${lib.cleanSource pkgs.path}" ]; - virtualisation.additionalPaths = [ pkgs.stdenv pkgs.pkgsi686Linux.stdenv ]; + virtualisation.additionalPaths = [ pkgs.stdenvNoCC pkgs.pkgsi686Linux.stdenvNoCC ]; }; testScript = { nodes }: '' diff --git a/tests/sourcehut-flakes.nix b/tests/nixos/sourcehut-flakes.nix index daa259dd6..a76fed020 100644 --- a/tests/sourcehut-flakes.nix +++ b/tests/nixos/sourcehut-flakes.nix @@ -1,12 +1,8 @@ -{ nixpkgs, system, overlay }: - -with import (nixpkgs + "/nixos/lib/testing-python.nix") -{ - inherit system; - extraConfigurations = [{ nixpkgs.overlays = [ overlay ]; }]; -}; +{ lib, config, hostPkgs, nixpkgs, ... }: let + pkgs = config.nodes.sourcehut.nixpkgs.pkgs; + # Generate a fake root CA and a fake git.sr.ht certificate. cert = pkgs.runCommand "cert" { buildInputs = [ pkgs.openssl ]; } '' @@ -64,8 +60,6 @@ let in -makeTest ( - { name = "sourcehut-flakes"; @@ -108,7 +102,7 @@ makeTest ( virtualisation.diskSize = 2048; virtualisation.additionalPaths = [ pkgs.hello pkgs.fuse ]; virtualisation.memorySize = 4096; - nix.binaryCaches = lib.mkForce [ ]; + nix.settings.substituters = lib.mkForce [ ]; nix.extraOptions = '' experimental-features = nix-command flakes flake-registry = https://git.sr.ht/~NixOS/flake-registry/blob/master/flake-registry.json @@ -164,4 +158,4 @@ makeTest ( client.succeed("nix build nixpkgs#fuse --tarball-ttl 0") ''; - }) +} diff --git a/tests/output-normalization.sh b/tests/output-normalization.sh new file mode 100644 index 000000000..0f6df5e31 --- /dev/null +++ b/tests/output-normalization.sh @@ -0,0 +1,9 @@ +source common.sh + +testNormalization () { + clearStore + outPath=$(nix-build ./simple.nix --no-out-link) + test "$(stat -c %Y $outPath)" -eq 1 +} + +testNormalization diff --git a/tests/path-from-hash-part.sh b/tests/path-from-hash-part.sh new file mode 100644 index 000000000..bdd104434 --- /dev/null +++ b/tests/path-from-hash-part.sh @@ -0,0 +1,10 @@ +source common.sh + +path=$(nix build --no-link --print-out-paths -f simple.nix) + +hash_part=$(basename $path) +hash_part=${hash_part:0:32} + +path2=$(nix store path-from-hash-part $hash_part) + +[[ $path = $path2 ]] diff --git a/tests/plugins.sh b/tests/plugins.sh index 6e278ad9d..baf71a362 100644 --- a/tests/plugins.sh +++ b/tests/plugins.sh @@ -1,10 +1,7 @@ source common.sh -set -o pipefail - if [[ $BUILD_SHARED_LIBS != 1 ]]; then - echo "plugins are not supported" - exit 99 + skipTest "Plugins are not supported" fi res=$(nix --option setting-set true --option plugin-files $PWD/plugins/libplugintest* eval --expr builtins.anotherNull) diff --git a/tests/plugins/local.mk b/tests/plugins/local.mk index 82ad99402..8182a6a83 100644 --- a/tests/plugins/local.mk +++ b/tests/plugins/local.mk @@ -8,4 +8,4 @@ libplugintest_ALLOW_UNDEFINED := 1 libplugintest_EXCLUDE_FROM_LIBRARY_LIST := 1 -libplugintest_CXXFLAGS := -I src/libutil -I src/libexpr +libplugintest_CXXFLAGS := -I src/libutil -I src/libstore -I src/libexpr diff --git a/tests/post-hook.sh b/tests/post-hook.sh index 4eff5f511..0266eb15d 100644 --- a/tests/post-hook.sh +++ b/tests/post-hook.sh @@ -9,8 +9,14 @@ echo 'require-sigs = false' >> $NIX_CONF_DIR/nix.conf restartDaemon +if isDaemonNewer "2.13"; then + pushToStore="$PWD/push-to-store.sh" +else + pushToStore="$PWD/push-to-store-old.sh" +fi + # Build the dependencies and push them to the remote store. -nix-build -o $TEST_ROOT/result dependencies.nix --post-build-hook $PWD/push-to-store.sh +nix-build -o $TEST_ROOT/result dependencies.nix --post-build-hook "$pushToStore" clearStore diff --git a/tests/pure-eval.sh b/tests/pure-eval.sh index b83ab8afe..5334bf28e 100644 --- a/tests/pure-eval.sh +++ b/tests/pure-eval.sh @@ -8,7 +8,7 @@ nix eval --expr 'assert 1 + 2 == 3; true' missingImpureErrorMsg=$(! nix eval --expr 'builtins.readFile ./pure-eval.sh' 2>&1) -echo "$missingImpureErrorMsg" | grep -q -- --impure || \ +echo "$missingImpureErrorMsg" | grepQuiet -- --impure || \ fail "The error message should mention the “--impure” flag to unblock users" [[ $(nix eval --expr 'builtins.pathExists ./pure-eval.sh') == false ]] || \ diff --git a/tests/push-to-store-old.sh b/tests/push-to-store-old.sh new file mode 100755 index 000000000..b1495c9e2 --- /dev/null +++ b/tests/push-to-store-old.sh @@ -0,0 +1,10 @@ +#!/bin/sh + +set -x +set -e + +[ -n "$OUT_PATHS" ] +[ -n "$DRV_PATH" ] + +echo Pushing "$OUT_PATHS" to "$REMOTE_STORE" +printf "%s" "$DRV_PATH" | xargs nix copy --to "$REMOTE_STORE" --no-require-sigs diff --git a/tests/push-to-store.sh b/tests/push-to-store.sh index b1495c9e2..0b090e1b3 100755 --- a/tests/push-to-store.sh +++ b/tests/push-to-store.sh @@ -7,4 +7,4 @@ set -e [ -n "$DRV_PATH" ] echo Pushing "$OUT_PATHS" to "$REMOTE_STORE" -printf "%s" "$DRV_PATH" | xargs nix copy --to "$REMOTE_STORE" --no-require-sigs +printf "%s" "$DRV_PATH"^'*' | xargs nix copy --to "$REMOTE_STORE" --no-require-sigs diff --git a/tests/readfile-context.builder.sh b/tests/readfile-context.builder.sh deleted file mode 100644 index 7084a08cb..000000000 --- a/tests/readfile-context.builder.sh +++ /dev/null @@ -1 +0,0 @@ -echo "$input" > $out diff --git a/tests/readfile-context.nix b/tests/readfile-context.nix index 600036a94..54cd1afd9 100644 --- a/tests/readfile-context.nix +++ b/tests/readfile-context.nix @@ -6,14 +6,23 @@ let dependent = mkDerivation { name = "dependent"; - builder = ./readfile-context.builder.sh; - input = "${input}/hello"; + buildCommand = '' + mkdir -p $out + echo -n "$input1" > "$out/file1" + echo -n "$input2" > "$out/file2" + ''; + input1 = "${input}/hello"; + input2 = "hello"; }; readDependent = mkDerivation { - name = "read-dependent"; - builder = ./readfile-context.builder.sh; - input = builtins.readFile dependent; + # Will evaluate correctly because file2 doesn't have any references, + # even though the `dependent` derivation does. + name = builtins.readFile (dependent + "/file2"); + buildCommand = '' + echo "$input" > "$out" + ''; + input = builtins.readFile (dependent + "/file1"); }; in readDependent diff --git a/tests/recursive.sh b/tests/recursive.sh index 91518d67d..6335d44a5 100644 --- a/tests/recursive.sh +++ b/tests/recursive.sh @@ -4,7 +4,7 @@ sed -i 's/experimental-features .*/& recursive-nix/' "$NIX_CONF_DIR"/nix.conf restartDaemon # FIXME -if [[ $(uname) != Linux ]]; then exit 99; fi +if [[ $(uname) != Linux ]]; then skipTest "Not running Linux"; fi clearStore diff --git a/tests/remote-store.sh b/tests/remote-store.sh index 31210ab47..1ae126794 100644 --- a/tests/remote-store.sh +++ b/tests/remote-store.sh @@ -30,7 +30,3 @@ NIX_REMOTE= nix-store --dump-db > $TEST_ROOT/d2 cmp $TEST_ROOT/d1 $TEST_ROOT/d2 killDaemon - -user=$(whoami) -[ -e $NIX_STATE_DIR/gcroots/per-user/$user ] -[ -e $NIX_STATE_DIR/profiles/per-user/$user ] diff --git a/tests/repl.sh b/tests/repl.sh index c555560cc..be8adb742 100644 --- a/tests/repl.sh +++ b/tests/repl.sh @@ -33,14 +33,14 @@ testRepl () { nix repl "${nixArgs[@]}" <<< "$replCmds" || fail "nix repl does not work twice with the same inputs" # simple.nix prints a PATH during build - echo "$replOutput" | grep -qs 'PATH=' || fail "nix repl :log doesn't output logs" + echo "$replOutput" | grepQuiet -s 'PATH=' || fail "nix repl :log doesn't output logs" local replOutput="$(nix repl "${nixArgs[@]}" <<< "$replFailingCmds" 2>&1)" echo "$replOutput" - echo "$replOutput" | grep -qs 'This should fail' \ + echo "$replOutput" | grepQuiet -s 'This should fail' \ || fail "nix repl :log doesn't output logs for a failed derivation" local replOutput="$(nix repl --show-trace "${nixArgs[@]}" <<< "$replUndefinedVariable" 2>&1)" echo "$replOutput" - echo "$replOutput" | grep -qs "while evaluating the file" \ + echo "$replOutput" | grepQuiet -s "while evaluating the file" \ || fail "nix repl --show-trace doesn't show the trace" nix repl "${nixArgs[@]}" --option pure-eval true 2>&1 <<< "builtins.currentSystem" \ @@ -58,7 +58,7 @@ testReplResponse () { local commands="$1"; shift local expectedResponse="$1"; shift local response="$(nix repl "$@" <<< "$commands")" - echo "$response" | grep -qs "$expectedResponse" \ + echo "$response" | grepQuiet -s "$expectedResponse" \ || fail "repl command set: $commands @@ -121,5 +121,5 @@ sed -i 's/beforeChange/afterChange/' flake/flake.nix echo ":reload" echo "changingThing" ) | nix repl ./flake --experimental-features 'flakes repl-flake') -echo "$replResult" | grep -qs beforeChange -echo "$replResult" | grep -qs afterChange +echo "$replResult" | grepQuiet -s beforeChange +echo "$replResult" | grepQuiet -s afterChange diff --git a/tests/restricted.sh b/tests/restricted.sh index 242b901dd..776893a56 100644 --- a/tests/restricted.sh +++ b/tests/restricted.sh @@ -3,7 +3,7 @@ source common.sh clearStore nix-instantiate --restrict-eval --eval -E '1 + 2' -(! nix-instantiate --restrict-eval ./restricted.nix) +(! nix-instantiate --eval --restrict-eval ./restricted.nix) (! nix-instantiate --eval --restrict-eval <(echo '1 + 2')) nix-instantiate --restrict-eval ./simple.nix -I src=. nix-instantiate --restrict-eval ./simple.nix -I src1=simple.nix -I src2=config.nix -I src3=./simple.builder.sh @@ -48,4 +48,4 @@ output="$(nix eval --raw --restrict-eval -I "$traverseDir" \ --expr "builtins.readFile \"$traverseDir/$goUp$(pwd)/restricted-innocent\"" \ 2>&1 || :)" echo "$output" | grep "is forbidden" -! echo "$output" | grep -F restricted-secret +echo "$output" | grepInverse -F restricted-secret diff --git a/tests/search.sh b/tests/search.sh index 1a98f5b49..8742f8736 100644 --- a/tests/search.sh +++ b/tests/search.sh @@ -20,9 +20,9 @@ clearCache ## Search expressions # Check that empty search string matches all -nix search -f search.nix '' |grep -q foo -nix search -f search.nix '' |grep -q bar -nix search -f search.nix '' |grep -q hello +nix search -f search.nix '' |grepQuiet foo +nix search -f search.nix '' |grepQuiet bar +nix search -f search.nix '' |grepQuiet hello ## Tests for multiple regex/match highlighting diff --git a/tests/shell.sh b/tests/shell.sh index 6a80e8385..d2f7cf14e 100644 --- a/tests/shell.sh +++ b/tests/shell.sh @@ -10,7 +10,7 @@ nix shell -f shell-hello.nix hello -c hello NixOS | grep 'Hello NixOS' nix shell -f shell-hello.nix hello^dev -c hello2 | grep 'Hello2' nix shell -f shell-hello.nix 'hello^*' -c hello2 | grep 'Hello2' -if ! canUseSandbox; then exit 99; fi +requireSandboxSupport chmod -R u+w $TEST_ROOT/store0 || true rm -rf $TEST_ROOT/store0 diff --git a/tests/signing.sh b/tests/signing.sh index 6aafbeb91..9b673c609 100644 --- a/tests/signing.sh +++ b/tests/signing.sh @@ -81,7 +81,7 @@ info=$(nix path-info --store file://$cacheDir --json $outPath2) [[ $info =~ 'cache1.example.org' ]] [[ $info =~ 'cache2.example.org' ]] -# Copying to a diverted store should fail due to a lack of valid signatures. +# Copying to a diverted store should fail due to a lack of signatures by trusted keys. chmod -R u+w $TEST_ROOT/store0 || true rm -rf $TEST_ROOT/store0 (! nix copy --to $TEST_ROOT/store0 $outPath) diff --git a/tests/store-ping.sh b/tests/store-ping.sh index f9427cf0a..9846c7d3d 100644 --- a/tests/store-ping.sh +++ b/tests/store-ping.sh @@ -1,13 +1,17 @@ source common.sh STORE_INFO=$(nix store ping 2>&1) +STORE_INFO_JSON=$(nix store ping --json) echo "$STORE_INFO" | grep "Store URL: ${NIX_REMOTE}" if [[ -v NIX_DAEMON_PACKAGE ]] && isDaemonNewer "2.7.0pre20220126"; then DAEMON_VERSION=$($NIX_DAEMON_PACKAGE/bin/nix-daemon --version | cut -d' ' -f3) echo "$STORE_INFO" | grep "Version: $DAEMON_VERSION" + [[ "$(echo "$STORE_INFO_JSON" | jq -r ".version")" == "$DAEMON_VERSION" ]] fi expect 127 NIX_REMOTE=unix:$PWD/store nix store ping || \ fail "nix store ping on a non-existent store should fail" + +[[ "$(echo "$STORE_INFO_JSON" | jq -r ".url")" == "${NIX_REMOTE:-local}" ]] diff --git a/tests/tarball.sh b/tests/tarball.sh index d5cab879c..5f39658c9 100644 --- a/tests/tarball.sh +++ b/tests/tarball.sh @@ -19,7 +19,7 @@ test_tarball() { tarball=$TEST_ROOT/tarball.tar$ext (cd $TEST_ROOT && tar cf - tarball) | $compressor > $tarball - nix-env -f file://$tarball -qa --out-path | grep -q dependencies + nix-env -f file://$tarball -qa --out-path | grepQuiet dependencies nix-build -o $TEST_ROOT/result file://$tarball @@ -34,7 +34,7 @@ test_tarball() { nix-build -o $TEST_ROOT/result -E "import (fetchTree { type = \"tarball\"; url = file://$tarball; narHash = \"$hash\"; })" # Do not re-fetch paths already present nix-build -o $TEST_ROOT/result -E "import (fetchTree { type = \"tarball\"; url = file:///does-not-exist/must-remain-unused/$tarball; narHash = \"$hash\"; })" - nix-build -o $TEST_ROOT/result -E "import (fetchTree { type = \"tarball\"; url = file://$tarball; narHash = \"sha256-xdKv2pq/IiwLSnBBJXW8hNowI4MrdZfW+SYqDQs7Tzc=\"; })" 2>&1 | grep 'NAR hash mismatch in input' + expectStderr 102 nix-build -o $TEST_ROOT/result -E "import (fetchTree { type = \"tarball\"; url = file://$tarball; narHash = \"sha256-xdKv2pq/IiwLSnBBJXW8hNowI4MrdZfW+SYqDQs7Tzc=\"; })" | grep 'NAR hash mismatch in input' nix-instantiate --strict --eval -E "!((import (fetchTree { type = \"tarball\"; url = file://$tarball; narHash = \"$hash\"; })) ? submodules)" >&2 nix-instantiate --strict --eval -E "!((import (fetchTree { type = \"tarball\"; url = file://$tarball; narHash = \"$hash\"; })) ? submodules)" 2>&1 | grep 'true' diff --git a/tests/test-infra.sh b/tests/test-infra.sh new file mode 100644 index 000000000..54ae120e7 --- /dev/null +++ b/tests/test-infra.sh @@ -0,0 +1,85 @@ +# Test the functions for testing themselves! +# Also test some assumptions on how bash works that they rely on. +source common.sh + +# `true` should exit with 0 +expect 0 true + +# `false` should exit with 1 +expect 1 false + +# `expect` will fail when we get it wrong +expect 1 expect 0 false + +noisyTrue () { + echo YAY! >&2 + true +} + +noisyFalse () { + echo NAY! >&2 + false +} + +# These should redirect standard error to standard output +expectStderr 0 noisyTrue | grepQuiet YAY +expectStderr 1 noisyFalse | grepQuiet NAY + +# `set -o pipefile` is enabled + +pipefailure () { + # shellcheck disable=SC2216 + true | false | true +} +expect 1 pipefailure +unset pipefailure + +pipefailure () { + # shellcheck disable=SC2216 + false | true | true +} +expect 1 pipefailure +unset pipefailure + +commandSubstitutionPipeFailure () { + # shellcheck disable=SC2216 + res=$(set -eu -o pipefail; false | true | echo 0) +} +expect 1 commandSubstitutionPipeFailure + +# `set -u` is enabled + +# note (...), making function use subshell, as unbound variable errors +# in the outer shell are *rightly* not recoverable. +useUnbound () ( + set -eu + # shellcheck disable=SC2154 + echo "$thisVariableIsNotBound" +) +expect 1 useUnbound + +# ! alone unfortunately negates `set -e`, but it works in functions: +# shellcheck disable=SC2251 +! true +funBang () { + ! true +} +expect 1 funBang +unset funBang + +# `grep -v -q` is not what we want for exit codes, but `grepInverse` is +# Avoid `grep -v -q`. The following line proves the point, and if it fails, +# we'll know that `grep` had a breaking change or `-v -q` may not be portable. +{ echo foo; echo bar; } | grep -v -q foo +{ echo foo; echo bar; } | expect 1 grepInverse foo + +# `grepQuiet` is quiet +res=$(set -eu -o pipefail; echo foo | grepQuiet foo | wc -c) +(( res == 0 )) +unset res + +# `greqQietInverse` is both +{ echo foo; echo bar; } | expect 1 grepQuietInverse foo +res=$(set -eu -o pipefail; echo foo | expect 1 grepQuietInverse foo | wc -c) +(( res == 0 )) +unset res diff --git a/tests/timeout.sh b/tests/timeout.sh index e3fb3ebcc..b179b79a2 100644 --- a/tests/timeout.sh +++ b/tests/timeout.sh @@ -5,17 +5,14 @@ source common.sh # XXX: This shouldn’t be, but #4813 cause this test to fail needLocalStore "see #4813" -set +e -messages=$(nix-build -Q timeout.nix -A infiniteLoop --timeout 2 2>&1) -status=$? -set -e +messages=$(nix-build -Q timeout.nix -A infiniteLoop --timeout 2 2>&1) && status=0 || status=$? if [ $status -ne 101 ]; then echo "error: 'nix-store' exited with '$status'; should have exited 101" exit 1 fi -if ! echo "$messages" | grep -q "timed out"; then +if echo "$messages" | grepQuietInvert "timed out"; then echo "error: build may have failed for reasons other than timeout; output:" echo "$messages" >&2 exit 1 diff --git a/tests/toString-path.sh b/tests/toString-path.sh new file mode 100644 index 000000000..07eb87465 --- /dev/null +++ b/tests/toString-path.sh @@ -0,0 +1,8 @@ +source common.sh + +mkdir -p $TEST_ROOT/foo +echo bla > $TEST_ROOT/foo/bar + +[[ $(nix eval --raw --impure --expr "builtins.readFile (builtins.toString (builtins.fetchTree { type = \"path\"; path = \"$TEST_ROOT/foo\"; } + \"/bar\"))") = bla ]] + +[[ $(nix eval --json --impure --expr "builtins.readDir (builtins.toString (builtins.fetchTree { type = \"path\"; path = \"$TEST_ROOT/foo\"; }))") = '{"bar":"regular"}' ]] diff --git a/tests/user-envs-migration.sh b/tests/user-envs-migration.sh new file mode 100644 index 000000000..187372b16 --- /dev/null +++ b/tests/user-envs-migration.sh @@ -0,0 +1,35 @@ +# Test that the migration of user environments +# (https://github.com/NixOS/nix/pull/5226) does preserve everything + +source common.sh + +if isDaemonNewer "2.4pre20211005"; then + skipTest "Daemon is too new" +fi + + +killDaemon +unset NIX_REMOTE + +clearStore +clearProfiles +rm -rf ~/.nix-profile + +# Fill the environment using the older Nix +PATH_WITH_NEW_NIX="$PATH" +export PATH="$NIX_DAEMON_PACKAGE/bin:$PATH" + +nix-env -f user-envs.nix -i foo-1.0 +nix-env -f user-envs.nix -i bar-0.1 + +# Migrate to the new profile dir, and ensure that everything’s there +export PATH="$PATH_WITH_NEW_NIX" +nix-env -q # Trigger the migration +( [[ -L ~/.nix-profile ]] && \ + [[ $(readlink ~/.nix-profile) == ~/.local/share/nix/profiles/profile ]] ) || \ + fail "The nix profile should point to the new location" + +(nix-env -q | grep foo && nix-env -q | grep bar && \ + [[ -e ~/.nix-profile/bin/foo ]] && \ + [[ $(nix-env --list-generations | wc -l) == 2 ]]) || + fail "The nix profile should have the same content as before the migration" diff --git a/tests/user-envs.sh b/tests/user-envs.sh index d63fe780a..d1260ba04 100644 --- a/tests/user-envs.sh +++ b/tests/user-envs.sh @@ -1,6 +1,6 @@ source common.sh -if [ -z "$storeCleared" ]; then +if [ -z "${storeCleared-}" ]; then clearStore fi @@ -28,13 +28,13 @@ nix-env -f ./user-envs.nix -qa --json --out-path | jq -e '.[] | select(.name == ] | all' # Query descriptions. -nix-env -f ./user-envs.nix -qa '*' --description | grep -q silly +nix-env -f ./user-envs.nix -qa '*' --description | grepQuiet silly rm -rf $HOME/.nix-defexpr ln -s $(pwd)/user-envs.nix $HOME/.nix-defexpr -nix-env -qa '*' --description | grep -q silly +nix-env -qa '*' --description | grepQuiet silly # Query the system. -nix-env -qa '*' --system | grep -q $system +nix-env -qa '*' --system | grepQuiet $system # Install "foo-1.0". nix-env -i foo-1.0 @@ -42,19 +42,19 @@ nix-env -i foo-1.0 # Query installed: should contain foo-1.0 now (which should be # executable). test "$(nix-env -q '*' | wc -l)" -eq 1 -nix-env -q '*' | grep -q foo-1.0 +nix-env -q '*' | grepQuiet foo-1.0 test "$($profiles/test/bin/foo)" = "foo-1.0" # Test nix-env -qc to compare installed against available packages, and vice versa. -nix-env -qc '*' | grep -q '< 2.0' -nix-env -qac '*' | grep -q '> 1.0' +nix-env -qc '*' | grepQuiet '< 2.0' +nix-env -qac '*' | grepQuiet '> 1.0' # Test the -b flag to filter out source-only packages. [ "$(nix-env -qab | wc -l)" -eq 1 ] # Test the -s flag to get package status. -nix-env -qas | grep -q 'IP- foo-1.0' -nix-env -qas | grep -q -- '--- bar-0.1' +nix-env -qas | grepQuiet 'IP- foo-1.0' +nix-env -qas | grepQuiet -- '--- bar-0.1' # Disable foo. nix-env --set-flag active false foo @@ -74,15 +74,15 @@ nix-env -i foo-2.0pre1 # Query installed: should contain foo-2.0pre1 now. test "$(nix-env -q '*' | wc -l)" -eq 1 -nix-env -q '*' | grep -q foo-2.0pre1 +nix-env -q '*' | grepQuiet foo-2.0pre1 test "$($profiles/test/bin/foo)" = "foo-2.0pre1" # Upgrade "foo": should install foo-2.0. -NIX_PATH=nixpkgs=./user-envs.nix:$NIX_PATH nix-env -f '<nixpkgs>' -u foo +NIX_PATH=nixpkgs=./user-envs.nix:${NIX_PATH-} nix-env -f '<nixpkgs>' -u foo # Query installed: should contain foo-2.0 now. test "$(nix-env -q '*' | wc -l)" -eq 1 -nix-env -q '*' | grep -q foo-2.0 +nix-env -q '*' | grepQuiet foo-2.0 test "$($profiles/test/bin/foo)" = "foo-2.0" # Store the path of foo-2.0. @@ -94,20 +94,20 @@ nix-env -i bar-0.1 nix-env -e foo # Query installed: should only contain bar-0.1 now. -if nix-env -q '*' | grep -q foo; then false; fi -nix-env -q '*' | grep -q bar +if nix-env -q '*' | grepQuiet foo; then false; fi +nix-env -q '*' | grepQuiet bar # Rollback: should bring "foo" back. oldGen="$(nix-store -q --resolve $profiles/test)" nix-env --rollback [ "$(nix-store -q --resolve $profiles/test)" != "$oldGen" ] -nix-env -q '*' | grep -q foo-2.0 -nix-env -q '*' | grep -q bar +nix-env -q '*' | grepQuiet foo-2.0 +nix-env -q '*' | grepQuiet bar # Rollback again: should remove "bar". nix-env --rollback -nix-env -q '*' | grep -q foo-2.0 -if nix-env -q '*' | grep -q bar; then false; fi +nix-env -q '*' | grepQuiet foo-2.0 +if nix-env -q '*' | grepQuiet bar; then false; fi # Count generations. nix-env --list-generations @@ -129,7 +129,7 @@ nix-env --switch-generation 7 # Install foo-1.0, now using its store path. nix-env -i "$outPath10" -nix-env -q '*' | grep -q foo-1.0 +nix-env -q '*' | grepQuiet foo-1.0 nix-store -qR $profiles/test | grep "$outPath10" nix-store -q --referrers-closure $profiles/test | grep "$(nix-store -q --resolve $profiles/test)" [ "$(nix-store -q --deriver "$outPath10")" = $drvPath10 ] @@ -137,12 +137,12 @@ nix-store -q --referrers-closure $profiles/test | grep "$(nix-store -q --resolve # Uninstall foo-1.0, using a symlink to its store path. ln -sfn $outPath10/bin/foo $TEST_ROOT/symlink nix-env -e $TEST_ROOT/symlink -if nix-env -q '*' | grep -q foo; then false; fi -(! nix-store -qR $profiles/test | grep "$outPath10") +if nix-env -q '*' | grepQuiet foo; then false; fi +nix-store -qR $profiles/test | grepInverse "$outPath10" # Install foo-1.0, now using a symlink to its store path. nix-env -i $TEST_ROOT/symlink -nix-env -q '*' | grep -q foo +nix-env -q '*' | grepQuiet foo # Delete all old generations. nix-env --delete-generations old @@ -160,7 +160,7 @@ test "$(nix-env -q '*' | wc -l)" -eq 0 # Installing "foo" should only install the newest foo. nix-env -i foo test "$(nix-env -q '*' | grep foo- | wc -l)" -eq 1 -nix-env -q '*' | grep -q foo-2.0 +nix-env -q '*' | grepQuiet foo-2.0 # On the other hand, this should install both (and should fail due to # a collision). @@ -171,8 +171,8 @@ nix-env -e '*' nix-env -e '*' nix-env -i '*' test "$(nix-env -q '*' | wc -l)" -eq 2 -nix-env -q '*' | grep -q foo-2.0 -nix-env -q '*' | grep -q bar-0.1.1 +nix-env -q '*' | grepQuiet foo-2.0 +nix-env -q '*' | grepQuiet bar-0.1.1 # Test priorities: foo-0.1 has a lower priority than foo-1.0, so it # should be possible to install both without a collision. Also test diff --git a/tests/why-depends.sh b/tests/why-depends.sh index c12941e76..b35a0d1cf 100644 --- a/tests/why-depends.sh +++ b/tests/why-depends.sh @@ -6,6 +6,9 @@ cp ./dependencies.nix ./dependencies.builder0.sh ./config.nix $TEST_HOME cd $TEST_HOME +nix why-depends --derivation --file ./dependencies.nix input2_drv input1_drv +nix why-depends --file ./dependencies.nix input2_drv input1_drv + nix-build ./dependencies.nix -A input0_drv -o dep nix-build ./dependencies.nix -o toplevel @@ -13,9 +16,9 @@ FAST_WHY_DEPENDS_OUTPUT=$(nix why-depends ./toplevel ./dep) PRECISE_WHY_DEPENDS_OUTPUT=$(nix why-depends ./toplevel ./dep --precise) # Both outputs should show that `input-2` is in the dependency chain -echo "$FAST_WHY_DEPENDS_OUTPUT" | grep -q input-2 -echo "$PRECISE_WHY_DEPENDS_OUTPUT" | grep -q input-2 +echo "$FAST_WHY_DEPENDS_OUTPUT" | grepQuiet input-2 +echo "$PRECISE_WHY_DEPENDS_OUTPUT" | grepQuiet input-2 -# But only the “precise” one should refere to `reference-to-input-2` -echo "$FAST_WHY_DEPENDS_OUTPUT" | (! grep -q reference-to-input-2) -echo "$PRECISE_WHY_DEPENDS_OUTPUT" | grep -q reference-to-input-2 +# But only the “precise” one should refer to `reference-to-input-2` +echo "$FAST_WHY_DEPENDS_OUTPUT" | grepQuietInverse reference-to-input-2 +echo "$PRECISE_WHY_DEPENDS_OUTPUT" | grepQuiet reference-to-input-2 |