Merge pull request #1030 from pikajude/rsd-cc

Implement resolve-system-dependencies in C++

7 files changed, 207 insertions, 147 deletions

diff --git a/Makefile b/Makefile
index 11b3309f8..6b2d97cdd 100644
--- a/Makefile
+++ b/Makefile
@@ -13,6 +13,7 @@ makefiles = \
   src/nix-collect-garbage/local.mk \
   src/nix-prefetch-url/local.mk \
   src/buildenv/local.mk \
+  src/resolve-system-dependencies/local.mk \
   perl/local.mk \
   scripts/local.mk \
   corepkgs/local.mk \
diff --git a/configure.ac b/configure.ac
index a9e6b4313..1ae0e782d 100644
--- a/configure.ac
+++ b/configure.ac
@@ -246,23 +246,6 @@ AC_MSG_RESULT(yes)
 AC_SUBST(perlFlags)
 
 
-# Check for otool, an optional dependency on Darwin.
-AC_PATH_PROG(otool, otool)
-AC_MSG_CHECKING([that otool works])
-case $host_os in
-  darwin*)
-    if test -z "$otool" || ! $otool --version 2>/dev/null; then
-      AC_MSG_RESULT(no)
-      AC_MSG_ERROR([Can't get version from otool; do you need to install developer tools?])
-    fi
-    AC_MSG_RESULT(yes)
-    ;;
-  *)
-    AC_MSG_RESULT(not needed)
-    ;;
-esac
-
-
 # Whether to build the Perl bindings
 AC_MSG_CHECKING([whether to build the Perl bindings])
 AC_ARG_ENABLE(perl-bindings, AC_HELP_STRING([--enable-perl-bindings],
diff --git a/scripts/local.mk b/scripts/local.mk
index fef256451..9852d8298 100644
--- a/scripts/local.mk
+++ b/scripts/local.mk
@@ -11,19 +11,12 @@ nix_noinst_scripts := \
   $(d)/nix-profile.sh \
   $(d)/nix-reduce-build
 
-ifeq ($(OS), Darwin)
-  nix_noinst_scripts += $(d)/resolve-system-dependencies.pl
-endif
-
 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-program-in, $(d)/build-remote.pl, $(libexecdir)/nix))
-ifeq ($(OS), Darwin)
-  $(eval $(call install-program-in, $(d)/resolve-system-dependencies.pl, $(libexecdir)/nix))
-endif
 $(eval $(call install-symlink, nix-build, $(bindir)/nix-shell))
 
 clean-files += $(nix_bin_scripts) $(nix_noinst_scripts)
diff --git a/scripts/resolve-system-dependencies.pl.in b/scripts/resolve-system-dependencies.pl.in
deleted file mode 100755
index a20f0dc02..000000000
--- a/scripts/resolve-system-dependencies.pl.in
+++ /dev/null
@@ -1,122 +0,0 @@
-#! @perl@ -w @perlFlags@
-
-use utf8;
-use strict;
-use warnings;
-use Cwd qw(realpath);
-use Errno;
-use File::Basename qw(dirname);
-use File::Path qw(make_path);
-use File::Spec::Functions qw(catfile);
-use List::Util qw(reduce);
-use IPC::Open3;
-use Nix::Config;
-use Nix::Store qw(derivationFromPath);
-use POSIX qw(uname);
-use Storable qw(lock_retrieve lock_store);
-
-my ($sysname, undef, $version, undef, $machine) = uname;
-$sysname =~ /Darwin/ or die "This tool is only meant to be used on Darwin systems.";
-
-my $cache = "$Nix::Config::stateDir/dependency-maps/$machine-$sysname-$version.map";
-
-make_path dirname($cache);
-
-our $DEPS;
-eval {
-  $DEPS = lock_retrieve($cache);
-};
-
-if($!{ENOENT}) {
-  lock_store {}, $cache;
-  $DEPS = {};
-} elsif($@) {
-  die "Unable to obtain a lock on dependency-map file $cache: $@";
-}
-
-sub mkset(@) {
-  my %set;
-  @set{@_} = ();
-  \%set
-}
-
-sub union($$) {
-  my ($set1, $set2) = @_;
-  my %new = (%$set1, %$set2);
-  \%new
-}
-
-sub cache_filepath($) {
-  my $fp = shift;
-  $fp =~ s/-/--/g;
-  $fp =~ s/\//-/g;
-  $fp =~ s/^-//g;
-  catfile $cache, $fp
-}
-
-sub resolve_tree {
-  sub resolve_tree_inner {
-    my ($lib, $TREE) = @_;
-    return if (defined $TREE->{$lib});
-    $TREE->{$lib} = mkset(@{cache_get($lib)});
-    foreach my $dep (keys %{$TREE->{$lib}}) {
-      resolve_tree_inner($dep, $TREE);
-    }
-    values %$TREE
-  }
-
-  reduce { union($a, $b) } {}, resolve_tree_inner(@_)
-}
-
-sub cache_get {
-  my $key = shift;
-  if (defined $DEPS->{$key}) {
-    $DEPS->{$key}
-  } else {
-    cache_insert($key);
-    cache_get($key)
-  }
-}
-
-sub cache_insert($) {
-  my $key = shift;
-  print STDERR "Finding dependencies for $key...\n";
-  my @deps = find_deps($key);
-  $DEPS->{$key} = \@deps;
-}
-
-sub find_deps($) {
-  my $lib = shift;
-  my($chld_in, $chld_out, $chld_err);
-  my $pid = open3($chld_in, $chld_out, $chld_err, "@otool@", "-L", "-arch", "x86_64", $lib);
-  waitpid($pid, 0);
-  my $line = readline $chld_out;
-  if($? == 0 and $line !~ /not an object file/) {
-    my @libs;
-    while(<$chld_out>) {
-      my $dep = (split /\s+/)[1];
-      push @libs, $dep unless $dep eq $lib or $dep =~ /\@rpath/;
-    }
-    @libs
-  } elsif (-l $lib) {
-    (realpath($lib))
-  } else {
-    ()
-  }
-}
-
-if (defined $ARGV[0]) {
-  my $deps = derivationFromPath($ARGV[0])->{"env"}->{"__impureHostDeps"};
-  if (defined $deps) {
-    my @files = split(/\s+/, $deps);
-    my $depcache = {};
-    my $depset = reduce { union($a, $b) } (map { resolve_tree($_, $depcache) } @files);
-    print "extra-chroot-dirs\n";
-    print join("\n", keys %$depset);
-    print "\n";
-  }
-  lock_store($DEPS, $cache);
-} else {
-  print STDERR "Usage: $0 path/to/derivation.drv\n";
-  exit 1
-}
diff --git a/src/libstore/globals.cc b/src/libstore/globals.cc
index 7bf48be37..ecf81e8eb 100644
--- a/src/libstore/globals.cc
+++ b/src/libstore/globals.cc
@@ -76,7 +76,7 @@ void Settings::processEnvironment()
 
     // should be set with the other config options, but depends on nixLibexecDir
 #ifdef __APPLE__
-    preBuildHook = nixLibexecDir + "/nix/resolve-system-dependencies.pl";
+    preBuildHook = nixLibexecDir + "/nix/resolve-system-dependencies";
 #endif
 }
 
diff --git a/src/resolve-system-dependencies/local.mk b/src/resolve-system-dependencies/local.mk
new file mode 100644
index 000000000..8792a4a25
--- /dev/null
+++ b/src/resolve-system-dependencies/local.mk
@@ -0,0 +1,11 @@
+ifeq ($(OS), Darwin)
+  programs += resolve-system-dependencies
+endif
+
+resolve-system-dependencies_DIR := $(d)
+
+resolve-system-dependencies_INSTALL_DIR := $(libexecdir)/nix
+
+resolve-system-dependencies_LIBS := libstore libmain libutil libformat
+
+resolve-system-dependencies_SOURCES := $(d)/resolve-system-dependencies.cc
diff --git a/src/resolve-system-dependencies/resolve-system-dependencies.cc b/src/resolve-system-dependencies/resolve-system-dependencies.cc
new file mode 100644
index 000000000..a5f0cd7b3
--- /dev/null
+++ b/src/resolve-system-dependencies/resolve-system-dependencies.cc
@@ -0,0 +1,194 @@
+#include "derivations.hh"
+#include "globals.hh"
+#include "shared.hh"
+#include "store-api.hh"
+#include <sys/utsname.h>
+#include <algorithm>
+#include <iostream>
+#include <fstream>
+#include <sys/mman.h>
+#include <fcntl.h>
+#include <mach-o/loader.h>
+#include <mach-o/swap.h>
+
+#define DO_SWAP(x, y) ((x) ? OSSwapInt32(y) : (y))
+
+using namespace nix;
+
+static auto cacheDir = Path{};
+
+Path resolveCacheFile(Path lib) {
+    std::replace(lib.begin(), lib.end(), '/', '%');
+    return cacheDir + "/" + lib;
+}
+
+std::set<string> readCacheFile(const Path & file) {
+    return tokenizeString<set<string>>(readFile(file), "\n");
+}
+
+void writeCacheFile(const Path & file, std::set<string> & deps) {
+    std::ofstream fp;
+    fp.open(file);
+    for (auto & d : deps) {
+        fp << d << "\n";
+    }
+    fp.close();
+}
+
+std::string findDylibName(bool should_swap, ptrdiff_t dylib_command_start) {
+    struct dylib_command *dylc = (struct dylib_command*)dylib_command_start;
+    return std::string((char*)(dylib_command_start + DO_SWAP(should_swap, dylc->dylib.name.offset)));
+}
+
+std::set<std::string> runResolver(const Path & filename) {
+    int fd = open(filename.c_str(), O_RDONLY);
+    struct stat s;
+    fstat(fd, &s);
+    void *obj = mmap(NULL, s.st_size, PROT_READ, MAP_SHARED, fd, 0);
+
+    ptrdiff_t mach64_offset = 0;
+
+    uint32_t magic = ((struct mach_header_64*)obj)->magic;
+    if(magic == FAT_CIGAM || magic == FAT_MAGIC) {
+        bool should_swap = magic == FAT_CIGAM;
+        uint32_t narches = DO_SWAP(should_swap, ((struct fat_header*)obj)->nfat_arch);
+
+        for(uint32_t iter = 0; iter < narches; iter++) {
+            ptrdiff_t header_offset = (ptrdiff_t)obj + sizeof(struct fat_header) * (iter + 1);
+            struct fat_arch* arch = (struct fat_arch*)header_offset;
+            if(DO_SWAP(should_swap, arch->cputype) == CPU_TYPE_X86_64) {
+                mach64_offset = (ptrdiff_t)DO_SWAP(should_swap, arch->offset);
+                break;
+            }
+        }
+        if (mach64_offset == 0) {
+            printMsg(lvlError, format("Could not find any mach64 blobs in file ‘%1%’, continuing...") % filename);
+            return std::set<string>();
+        }
+    } else if (magic == MH_MAGIC_64 || magic == MH_CIGAM_64) {
+        mach64_offset = 0;
+    } else {
+        printMsg(lvlError, format("Object file has unknown magic number ‘%1%’, skipping it...") % magic);
+        return std::set<string>();
+    }
+
+    ptrdiff_t mach_header_offset = (ptrdiff_t)obj + mach64_offset;
+    struct mach_header_64 *m_header = (struct mach_header_64 *)mach_header_offset;
+
+    bool should_swap = magic == MH_CIGAM_64;
+    ptrdiff_t cmd_offset = mach_header_offset + sizeof(struct mach_header_64);
+
+    std::set<string> libs;
+    for(uint32_t i = 0; i < DO_SWAP(should_swap, m_header->ncmds); i++) {
+        struct load_command *cmd = (struct load_command*)cmd_offset;
+        switch(DO_SWAP(should_swap, cmd->cmd)) {
+            case LC_LOAD_UPWARD_DYLIB:
+            case LC_LOAD_DYLIB:
+            case LC_REEXPORT_DYLIB:
+                libs.insert(findDylibName(should_swap, cmd_offset));
+                break;
+        }
+        cmd_offset += DO_SWAP(should_swap, cmd->cmdsize);
+    }
+
+    return libs;
+}
+
+bool isSymlink(const Path & path) {
+    struct stat st;
+    if(lstat(path.c_str(), &st))
+        throw SysError(format("getting attributes of path ‘%1%’") % path);
+
+    return S_ISLNK(st.st_mode);
+}
+
+Path resolveSymlink(const Path & path) {
+    char buf[PATH_MAX];
+    ssize_t len = readlink(path.c_str(), buf, sizeof(buf) - 1);
+    if(len != -1) {
+        buf[len] = 0;
+        return Path(buf);
+    } else {
+        throw SysError(format("readlink('%1%')") % path);
+    }
+}
+
+std::set<string> resolveTree(const Path & path, PathSet & deps) {
+    std::set<string> results;
+    if(deps.find(path) != deps.end()) {
+        return std::set<string>();
+    }
+    deps.insert(path);
+    for (auto & lib : runResolver(path)) {
+        results.insert(lib);
+        for (auto & p : resolveTree(lib, deps)) {
+            results.insert(p);
+        }
+    }
+    return results;
+}
+
+std::set<string> getPath(const Path & path) {
+    Path cacheFile = resolveCacheFile(path);
+    if(pathExists(cacheFile)) {
+        return readCacheFile(cacheFile);
+    }
+
+    std::set<string> deps;
+    std::set<string> paths;
+    paths.insert(path);
+
+    Path next_path = Path(path);
+    while(isSymlink(next_path)) {
+        next_path = resolveSymlink(next_path);
+        paths.insert(next_path);
+    }
+
+    for(auto & t : resolveTree(next_path, deps)) {
+        paths.insert(t);
+    }
+
+    writeCacheFile(cacheFile, paths);
+
+    return paths;
+}
+
+int main(int argc, char ** argv) {
+    return handleExceptions(argv[0], [&]() {
+        initNix();
+
+        struct utsname _uname;
+
+        uname(&_uname);
+
+        auto cacheParentDir = (format("%1%/dependency-maps") % settings.nixStateDir).str();
+
+        cacheDir = (format("%1%/%2%-%3%-%4%")
+                % cacheParentDir
+                % _uname.machine
+                % _uname.sysname
+                % _uname.release).str();
+
+        mkdir(cacheParentDir.c_str(), 0755);
+        mkdir(cacheDir.c_str(), 0755);
+
+        auto store = openStore();
+
+        auto drv = store->derivationFromPath(Path(argv[1]));
+        Strings impurePaths = tokenizeString<Strings>(get(drv.env, "__impureHostDeps"));
+
+        std::set<string> all_paths;
+
+        for (auto & path : impurePaths) {
+            for(auto & p : getPath(path)) {
+                all_paths.insert(p);
+            }
+        }
+
+        std::cout << "extra-chroot-dirs" << std::endl;
+        for(auto & path : all_paths) {
+            std::cout << path << std::endl;
+        }
+        std::cout << std::endl;
+    });
+}