| Age | Commit message (Collapse) | Author |
|---|
|
Prevent tempdir from being GC-ed before addToStoreFromDump completes
|
|
|
|
nix shell: document how to invoke multiple commands from the command …
|
|
Docs: Add nix develop --command entry
|
|
Implements the approach suggested by feedback on PR #6994, where
tempdir paths are created in the store (now with an exclusive lock).
As part of this work, the currently-broken and unused
`createTempDirInStore` function is updated to create an exclusive lock
on the temp directory in the store.
The GC now makes a non-blocking attempt to lock any store directories
that "look like" the temp directories created by this function, and if
it can't acquire one, ignores the directory.
|
|
|
|
Stdenv sets this to a bash that doesn't have readline/completion
support, so running 'nix (develop|shell)' inside a 'nix develop' gives
you a crippled shell. So let's just ignore the derivation's $SHELL.
This could break interactive use of build phases that use $SHELL, but
they appear to be fairly rare.
|
|
Disable SA_RESTART for some signals on macOS
|
|
Add example of nix develop being used to execuate a series of script
commands. This is common when doing things like CI/CD, and should be
represented in the official documentation.
Also useful for people looking for the 'nix develop' equivalent of
'nix-shell --run'.
Related:
- https://github.com/NixOS/nix/issues/6908
- https://github.com/NixOS/nix/issues/6908#issuecomment-1229266853
|
|
|
|
|
|
Disables the SA_RESTART behavior on macOS which causes:
> Restarting of pending calls is requested by setting the SA_RESTART bit
> in sa_flags. The affected system calls include read(2), write(2),
> sendto(2), recvfrom(2), sendmsg(2) and recvmsg(2) on a communications
> channel or a slow device (such as a terminal, but not a regular file)
> and during a wait(2) or ioctl(2).
From: https://man.openbsd.org/sigaction#SA_RESTART
This being set on macOS caused a bug where read() calls to the daemon
socket were blocking after a SIGINT was received. As a result,
checkInterrupt was never reached even though the signal was received
by the signal handler thread.
On Linux, SA_RESTART is disabled by default. This probably effects
other BSDs but I don’t have the ability to test it there right now.
|
|
This fixes issue 6823 by placing the tempdir used in LocalStore::addToStoreFromDump
outside the Nix store, where automatic GC is no longer a concern.
|
|
readDerivation is pretty slow, and while it may not be significant for
some use cases, on things like ghc-nix where we have thousands of
derivations is really slows things down.
So, this just doesn’t do the impure derivation check if the impure
derivation experimental feature is disabled. Perhaps we could cache
the result of isPure() and keep the check, but this is a quick fix to
for the slowdown introduced with impure derivations features in 2.8.0.
|
|
|
|
documentation
|
|
This reverts commit c530cda345377370c52a616d608de88b9d67cd40, reversing
changes made to 4adcdff5c1d5f9f135c4ec61d690890443c19e6a.
|
|
Fix nested flake input overrides
|
|
do not render relative links in help pages
|
|
|
|
|
|
this simplifies the setup a lot, and avoids weird looking `./file.md`
links showing up.
it also does not show regular URLs any more. currently the command
reference only has few of them, and not showing them in the offline
documentation is hopefully not a big deal.
instead of building more special-case solutions, clumsily preprocessing
the input, or issuing verbal rules on dealing with URLs, should better
be solved sustainably by not rendering relative links in `lowdown`:
https://github.com/kristapsdz/lowdown/issues/105
|
|
fix(libstore): allow access to trustd on macOS
|
|
This was caused by -L calling setLogFormat() again, which caused the
creation of a new progress bar without destroying the old one. So we
had two progress bars clobbering each other.
We should change 'logger' to be a smart pointer, but I'll do that in a
future PR.
Fixes #6931.
|
|
98e361ad4c1a26d4ffe4762a6f33bb9e39321a39 introduced a regression where
previously stored attributes were replaced by placeholders. As a
result, a command like 'nix build nixpkgs#hello' had to be executed at
least twice to get caching.
This code does not seem necessary for suggestions to work.
|
|
|
|
Make nix copy parallel again
|
|
|
|
|
|
This issue made it impossible for clients using a serve protocol of
version <= 2.3 to use the `cmdBuildDerivation` command of servers using
a protocol of version >= 2.6. The faulty version check makes the server
send back build outputs that the client is not expecting.
|
|
nix develop: Ignore some more bash special variables
|
|
JSON: print paths as strings without copying them to the store
|
|
Fixes #6940.
|
|
nix-shell: specify which outputs from bashInteractive to build
|
|
|
|
This hang for some reason didn't trigger in the Nix build, but did
running 'make installcheck' interactively. What happened:
* Store::addMultipleToStore() calls a SinkToSource object to copy a
path, which in turn calls LegacySSHStore::narFromPath(), which
acquires a connection.
* The SinkToSource object is not destroyed after the last bytes has
been read, so the coroutine's stack is still alive and its
destructors are not run. So the connection is not released.
* Then when the next path is copied, because max-connections = 1,
LegacySSHStore::narFromPath() hangs forever waiting for a connection
to be released.
The fix is to make sure that the source object is destroyed when we're
done with it.
|
|
|
|
|
|
nix repl: Stop the progress bar
|
|
after `nix eval --json` and `nix-instantiate --eval --json`.
|
|
Makes `printValueAsJSON` not copy paths to the store for `nix eval
--json`, `nix-instantiate --eval --json` and `nix-env --json`.
Fixes https://github.com/NixOS/nix/issues/5612
|
|
add a nix.conf option to set a download speed limit
|
|
Fix default profile path for root in nix profile documentation
|
|
The repl was broken since c3769c68465bae971ab6bb48cfcdea85b61ea83a. In
general, the progress bar is incompatible with the repl.
|
|
Avoid producing incomplete JSON
|
|
|
|
Co-authored-by: Anderson Torres <torres.anderson.85@protonmail.com>
|
|
|
|
|
|
RewritingSink can handle being fed input where a reference crosses a
chunk boundary. we don't need to load the whole source into memory, and
in fact *not* loading the whole source lets nix build FODs that do not
fit into memory (eg fetchurl'ing data files larger than system memory).
|
