For now, we don't NATIVE_SYSTEM_HEADER_DIR because it breaks the
build. However, it points to Glibc in the Nix store (not /usr/include)
so it's kind of okay.
Now gcc is just another build input, making it possible in the future
to have a stdenv that doesn't depend on a C compiler. This is very
useful on NixOS, since it would allow trivial builders like
writeTextFile to work without pulling in the C compiler.
Somewhere the no-sys-dirs.patch got disabled, so gcc was looking in
/usr/local/include and /usr/lib. Since I can't fix the patch easily,
I've borrowed the --sysroot trick from clang-wrapper. This causes
builtin paths to be prefixed with /var/empty
(e.g. /var/empty/usr/lib), which don't exist.
The gcc-wrapper doesn't wrap 'cpp'. This breaks some software (such as
Buildroot) because the 'cpp' they get come from the non-wrapped gcc
package which doesn't know about any standard include paths.
gcc-cross-wrapper is untested.
GCC provides a number of libraries that are used by programs built by
GCC, in particular libgcc_s.so and libstdc++.so. This caused programs
that used these libraries to have a runtime dependency on all of GCC
(~77 MiB). Now they only depend on the "lib" output of GCC (~1.6
MiB).
With this and previous multiple-output improvements, closure sizes are
reduced a lot:
hello: 41 MiB -> 22 MiB
patchelf: 118 MiB -> 23 MiB
pan: 364 MiB -> 90 MiB
This reduces the size of closures a lot by getting rid of Glibc's
header files, static libraries and utilities. For instance, the
closure of GNU Hello went from 41 MiB to 22 MiB.
TODO: We may want to merge back in parts of Glibc that we previously
split off (such as the info files).
This is for consistency with terminology in stdenv (and the terms
"hostDrv" and "buildDrv" are not very intuitive, even if they're
consistent with GNU terminology).
This feature sounds crazy, but it is used in some configure scripts (e.g. xbmc).
This patch causes an almost complete rebuild of Nixpkgs.
Patch submitted by Jan Malakhovski <oxij@oxij.org>.
what the new nix thinks the fuloong is.
Anyone having the old nix should use a nixpkgs previous to this change to build
the new nix. And then, with the new nix, he can use any newer nixpkgs revision.
svn path=/nixpkgs/trunk/; revision=31751
path worth a "-rpath" entry.
This put a problem in the bootstrap, where we built 'zlib' and wanted to link programs with it, and while they got linked well, at runtime the boostrap-tools zlib came first on rpath becaue
the ld call was done with "-dynamic-linker ..." before "-lz".
I saw this trouble on the fuloong, where the zlib in boostrap-tools is not
runtime compatible with that binutils are linked to.
svn path=/nixpkgs/branches/stdenv-updates/; revision=25107
derivation instead of a string, but this did not cover all use cases we had.
Instead of updating the use cases, I made the wrapper accept also a string.
We saw this problem trying to build gcc43_multi.
svn path=/nixpkgs/branches/stdenv-updates/; revision=25029
it properly put the rpath for directly passed .so files, and additionally it
works much faster than the old ld-wrapper.
svn path=/nixpkgs/branches/stdenv-updates/; revision=21978
on the next stdenv-updates.
This would fix the build for many cmake packages, although that requires updating the stdenv
in those for its gcc to use this 2nd wrapper.
I updated paraview and avidemux. I can't recall now more packages with problems, but I was
quite sure there were.
If anyone sees a cmake-built package with the result binaries lacking some rpaths, then try
this wrapper.
svn path=/nixpkgs/trunk/; revision=20669
because it makes linking very slow. Use bash's =~ operator instead
(and only once for each argument). We depend on bash already anyway
because of arrays so it's not a problem.
svn path=/nixpkgs/branches/stdenv-updates/; revision=19699
