0_Rootfs: Add GCC 14

[eschnett]

Companion PR to JuliaPackaging/BinaryBuilderBase.jl#403.

[eschnett]

I'm looking for comments on these changes. If they look fine I can generate and deploy the artifacts.

[imciner2]

Have you looked through the patches to see which are already upstream and included in the GCC 14 release? I know for sure that you need to drop the gcc9_vsnprintf-str patch since it is already upstream, and I would suspect also the gcc1220-libgomp-win patch should be dropped.

[eschnett]

There is a gazillion patches which don't apply any more. Since most are applied via patch || true that doesn't matter. Should I remove all that are not necessary?

I assume that's only GCC patches you're talking about, not anything related to binutils or glibc?

[imciner2]

IMO, when adding a new GCC version we should verify all the patches are either being applied or are no longer needed and dropped. Keeping a lot of old patches that don't apply present in the new GCC's directory clutters things up.

[eschnett]

This is GCC 14.2. Is that the version we're going to use?

[giordano]

IMO, when adding a new GCC version we should verify all the patches are either being applied or are no longer needed and dropped.

Agreed, I don't like the patch || true style, it's hard to change it for old versions we should avoid it for new stuff

[eschnett]

It's impossible to avoid. There are too many combinations of (GCC, glibc, binutils, architecture), and who knows which patches are necessary for what combination.

So far, the patches are separated by GCC version only. We could separate them by binutils and glibc version as well, but that might not cover all the cases. (For example, the configure flags necessary for glibc 2.17 depend on the GCC version.)

Furthermore, I don't think anyone has tried to re-build an older GCC version in a long time, so I'd assume that's all bit-rotted away.

[eschnett]

Here we go: https://github.com/JuliaPackaging/Yggdrasil/releases/tag/GCCBootstrap-v14.2.0

[giordano]

Furthermore, I don't think anyone has tried to re-build an older GCC version in a long time, so I'd assume that's all bit-rotted away.

I haven't tried very recently, but the build scripts are surprisingly somewhat robust.

For the benefit of future readers (myself included) what was the problem with JuliaPackaging/BinaryBuilderBase.jl#398 (comment)?

[eschnett]

I added this comment to explain:

        # These flags are necessary for GCC 14. GCC 14 defaults to a
        # modern version of C, too modern for the old glibc libraries we are
        # trying to build. Various configure tests would fail otherwise. (Why
        # declare variables or functions if they default to int anyway?)
        GLIBC_CFLAGS="${CFLAGS} -g -O2 -Wno-implicit-int -Wno-implicit-function-declaration -Wno-builtin-declaration-mismatch -Wno-array-parameter"

[giordano]

I haven't tried very recently, but the build scripts are surprisingly somewhat robust.

JuliaPackaging/BinaryBuilderBase.jl#318 rebuilt gcc 4-12 for *86*-linux-gnu little over one year ago, and since then the only major change was the upgrade to GCC 13. Unless we did something super bad there, the script is supposed to still be working for all those versions.

[imciner2]

I haven't tried very recently, but the build scripts are surprisingly somewhat robust.

I also rebuilt GCC 11-12 to fix headers and GCC 11-13 for the new Mingw version in January and it worked with no hiccups: #8005

[eschnett]

I see this error while building for x86_64-apple-darwin14:

[02:47:00] Assertion failed: targetAtom != NULL (/workspace/srcdir/cctools-port/cctools/ld64/src/ld/ld.hpp: Fixup: 776)
[02:47:00] collect2: fatal error: ld terminated with signal 6 [Aborted], core dumped
[02:47:00] compilation terminated.
[02:47:00] make[1]: *** [Makefile:1005: libgcc_s.dylib] Error 1

What's the likely fix? Update to a newer macOS SDK? Or try a different binutils?

[giordano]

Cctools more than binutils:

Yggdrasil/0_RootFS/gcc_sources.jl

Lines 158 to 164 in 51880c6

	# MacOS doesn't actually use binutils, it uses cctools
	binutils_sources = [
	GitSource("https://github.com/tpoechtrager/apple-libtapi.git",
	"a66284251b46d591ee4a0cb4cf561b92a0c138d8"),
	GitSource("https://github.com/tpoechtrager/cctools-port.git",
	"634a084377ee2e2932c66459b0396edf76da2e9f"),
	]

Maybe it's time to upgrade to a newer version (as we do for binutils), for example https://github.com/tpoechtrager/cctools-port/tree/1010.6-ld64-951.9

[eschnett]

Why is there no aarch64-apple-darwin* architecture for GCC?

[giordano]

For that architecture we build a fork of GCC: https://github.com/iains/gcc-darwin-arm64/. Upstream GCC still doesn't support it, I think.

[eschnett]

I'm giving up on building GCC 14 for x86_64-apple-darwin14. I've added my current state to the PR. I cannot figure out how to build the new version of cctools.

[giordano]

Alternatively, compile https://github.com/tpoechtrager/apple-libdispatch locally

[giordano]

I saved my progress so far with building gcc 14 for x86_64-darwin, but now I'm stuck at

Yggdrasil/0_RootFS/gcc_common.jl

Line 529 in 54d6862

make all-target-libgcc -j${nproc}

[22:25:57] /bin/sh /workspace/srcdir/gcc-14.2.0/libgcc/../mkinstalldirs .
[22:25:57] /workspace/srcdir/gcc_stage1/./gcc/xgcc -B/workspace/srcdir/gcc_stage1/./gcc/ -B/workspace/destdir/x86_64-apple-darwin14/bin/ -B/workspace/destdir/x86_64-apple-darwin14/lib/ -isystem /workspace/destdir/x86_64-apple-darwin14/include -isystem /workspace/destdir/x86_64-apple-darwin14/sys-include    -O2  -g -O2 -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE  -W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem ./include  -mmacosx-version-min=10.5 -fno-common -g -DIN_LIBGCC2 -fbuilding-libgcc -fno-stack-protector   -dynamiclib -nodefaultlibs -install_name /workspace/destdir/x86_64-apple-darwin14/lib/libgcc_s.1.1.dylib -single_module -o ./libgcc_s.dylib -Wl,-exported_symbols_list,libgcc.map -compatibility_version 1 -current_version 1.1 -nodefaultrpaths -g -O2 -B./ _muldi3_s.o _negdi2_s.o _lshrdi3_s.o _ashldi3_s.o _ashrdi3_s.o _cmpdi2_s.o _ucmpdi2_s.o _clear_cache_s.o _trampoline_s.o __main_s.o _absvsi2_s.o _absvdi2_s.o _addvsi3_s.o _addvdi3_s.o _subvsi3_s.o _subvdi3_s.o _mulvsi3_s.o _mulvdi3_s.o _negvsi2_s.o _negvdi2_s.o _ctors_s.o _ffssi2_s.o _ffsdi2_s.o _clz_s.o _clzsi2_s.o _clzdi2_s.o _ctzsi2_s.o _ctzdi2_s.o _popcount_tab_s.o _popcountsi2_s.o _popcountdi2_s.o _paritysi2_s.o _paritydi2_s.o _powisf2_s.o _powidf2_s.o _powixf2_s.o _powitf2_s.o _mulsc3_s.o _muldc3_s.o _mulxc3_s.o _multc3_s.o _divsc3_s.o _divdc3_s.o _divxc3_s.o _divtc3_s.o _bswapsi2_s.o _bswapdi2_s.o _clrsbsi2_s.o _clrsbdi2_s.o _mulbitint3_s.o _fixunssfsi_s.o _fixunsdfsi_s.o _fixunsxfsi_s.o _fixsfdi_s.o _fixdfdi_s.o _fixxfdi_s.o _fixunssfdi_s.o _fixunsdfdi_s.o _fixunsxfdi_s.o _floatdisf_s.o _floatdidf_s.o _floatdixf_s.o _floatundisf_s.o _floatundidf_s.o _floatundixf_s.o _fixsfti_s.o _fixdfti_s.o _fixxfti_s.o _fixtfti_s.o _fixunssfti_s.o _fixunsdfti_s.o _fixunsxfti_s.o _fixunstfti_s.o _floattisf_s.o _floattidf_s.o _floattixf_s.o _floattitf_s.o _floatuntisf_s.o _floatuntidf_s.o _floatuntixf_s.o _floatuntitf_s.o _divdi3_s.o _moddi3_s.o _divmoddi4_s.o _udivdi3_s.o _umoddi3_s.o _udivmoddi4_s.o _udiv_w_sdiv_s.o _divmodbitint4_s.o darwin-64_s.o cpuinfo_s.o sfp-exceptions_s.o _divhc3_s.o _mulhc3_s.o addtf3_s.o divtf3_s.o eqtf2_s.o getf2_s.o letf2_s.o multf3_s.o negtf2_s.o subtf3_s.o unordtf2_s.o fixtfsi_s.o fixunstfsi_s.o floatsitf_s.o floatunsitf_s.o fixtfdi_s.o fixunstfdi_s.o floatditf_s.o floatunditf_s.o fixtfti_s.o fixunstfti_s.o floattitf_s.o floatuntitf_s.o fixsfbitint_s.o floatbitintsf_s.o fixdfbitint_s.o floatbitintdf_s.o extendhfsf2_s.o extendhfdf2_s.o extendhftf2_s.o extendhfxf2_s.o extendsfdf2_s.o extendsftf2_s.o extenddftf2_s.o extendxftf2_s.o extendbfsf2_s.o trunctfhf2_s.o truncxfhf2_s.o truncdfhf2_s.o truncsfhf2_s.o trunctfsf2_s.o truncdfsf2_s.o trunctfdf2_s.o trunctfxf2_s.o trunctfbf2_s.o truncxfbf2_s.o truncdfbf2_s.o truncsfbf2_s.o trunchfbf2_s.o fixtfbitint_s.o floatbitinttf_s.o fixhfti_s.o fixunshfti_s.o floattihf_s.o floatuntihf_s.o floattibf_s.o floatuntibf_s.o eqhf2_s.o fixxfbitint_s.o floatbitinthf_s.o floatbitintbf_s.o floatbitintxf_s.o enable-execute-stack_s.o hardcfr_s.o strub_s.o heap-trampoline_s.o emutls_s.o libgcc.a -lSystem
[22:25:57] Assertion failed: targetAtom != NULL (/workspace/srcdir/cctools-port/cctools/ld64/src/ld/ld.hpp: Fixup: 902)
[22:25:57] collect2: fatal error: ld terminated with signal 6 [Aborted]
[22:25:57] compilation terminated.
[22:25:57] make[1]: *** [Makefile:1005: libgcc_s.dylib] Error 1
[22:25:57] make[1]: *** Waiting for unfinished jobs....
[22:25:58] make[1]: Leaving directory '/workspace/srcdir/gcc_stage1/x86_64-apple-darwin14/libgcc'
[22:25:58] make: *** [Makefile:14143: all-target-libgcc] Error 2

The assertion is https://github.com/tpoechtrager/cctools-port/blob/81f205e8ca6bbf2fdbcb6948132454fd1f97839e/cctools/ld64/src/ld/ld.hpp#L902.

At a quick glance it looks similar to NixOS/nixpkgs#105573, but I have yet to study that bug report and understand how it was resolved. But what's interesting is that NixOS/nixpkgs#105573 (comment) claims they hit the issue on tpoechtrager/cctools-port@43f32a4, which is before tpoechtrager/cctools-port@634a084, the commit we've built so far and we didn't have this problem before.

Any help in debugging this (even in the form of googling or rubber-ducking) would be much appreciated 🙂

[giordano]

As far as I can tell, all errors mentioned in NixOS/nixpkgs#105573 about the failed assertion are related to -mmacosx-version-min=10.5, which is also the version used in the failed compiler invocation above. That flag was introduced by gcc-mirror/gcc@20b8779, before -mmacosx-version-min=10.4 was used unconditionally. Now I'm wondering if cctools just crashes badly on -mmacosx-version-min=10.5 specifically, that'd be hilarious. I need to investigate this hypothesis.

Edit:

sandbox:${WORKSPACE}/srcdir/gcc_stage1/x86_64-apple-darwin14/libgcc # /workspace/srcdir/gcc_stage1/./gcc/xgcc -B/workspace/srcdir/gcc_stage1/./gcc/ -B/workspace/destdir/x86_64-apple-darwin14/bin/ -B/workspace/destdir/x86_64-apple-darwin14/lib/ -isystem /workspace/destdir/x86_64-apple-darwin14/include -isystem /workspace/destdir/x86_64-apple-darwin14/sys-include    -O2  -g -O2 -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE  -W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem ./include  -mmacosx-version-min=10.5 -fno-common -g -DIN_LIBGCC2 -fbuilding-libgcc -fno-stack-protector   -dynamiclib -nodefaultlibs -install_name /workspace/destdir/x86_64-apple-darwin14/lib/libgcc_s.1.1.dylib -single_module -o ./libgcc_s.dylib -Wl,-exported_symbols_list,libgcc.map -compatibility_version 1 -current_version 1.1 -nodefaultrpaths -g -O2 -B./ _muldi3_s.o _negdi2_s.o _lshrdi3_s.o _ashldi3_s.o _ashrdi3_s.o _cmpdi2_s.o _ucmpdi2_s.o _clear_cache_s.o _trampoline_s.o __main_s.o _absvsi2_s.o _absvdi2_s.o _addvsi3_s.o _addvdi3_s.o _subvsi3_s.o _subvdi3_s.o _mulvsi3_s.o _mulvdi3_s.o _negvsi2_s.o _negvdi2_s.o _ctors_s.o _ffssi2_s.o _ffsdi2_s.o _clz_s.o _clzsi2_s.o _clzdi2_s.o _ctzsi2_s.o _ctzdi2_s.o _popcount_tab_s.o _popcountsi2_s.o _popcountdi2_s.o _paritysi2_s.o _paritydi2_s.o _powisf2_s.o _powidf2_s.o _powixf2_s.o _powitf2_s.o _mulsc3_s.o _muldc3_s.o _mulxc3_s.o _multc3_s.o _divsc3_s.o _divdc3_s.o _divxc3_s.o _divtc3_s.o _bswapsi2_s.o _bswapdi2_s.o _clrsbsi2_s.o _clrsbdi2_s.o _mulbitint3_s.o _fixunssfsi_s.o _fixunsdfsi_s.o _fixunsxfsi_s.o _fixsfdi_s.o _fixdfdi_s.o _fixxfdi_s.o _fixunssfdi_s.o _fixunsdfdi_s.o _fixunsxfdi_s.o _floatdisf_s.o _floatdidf_s.o _floatdixf_s.o _floatundisf_s.o _floatundidf_s.o _floatundixf_s.o _fixsfti_s.o _fixdfti_s.o _fixxfti_s.o _fixtfti_s.o _fixunssfti_s.o _fixunsdfti_s.o _fixunsxfti_s.o _fixunstfti_s.o _floattisf_s.o _floattidf_s.o _floattixf_s.o _floattitf_s.o _floatuntisf_s.o _floatuntidf_s.o _floatuntixf_s.o _floatuntitf_s.o _divdi3_s.o _moddi3_s.o _divmoddi4_s.o _udivdi3_s.o _umoddi3_s.o _udivmoddi4_s.o _udiv_w_sdiv_s.o _divmodbitint4_s.o darwin-64_s.o cpuinfo_s.o sfp-exceptions_s.o _divhc3_s.o _mulhc3_s.o addtf3_s.o divtf3_s.o eqtf2_s.o getf2_s.o letf2_s.o multf3_s.o negtf2_s.o subtf3_s.o unordtf2_s.o fixtfsi_s.o fixunstfsi_s.o floatsitf_s.o floatunsitf_s.o fixtfdi_s.o fixunstfdi_s.o floatditf_s.o floatunditf_s.o fixtfti_s.o fixunstfti_s.o floattitf_s.o floatuntitf_s.o fixsfbitint_s.o floatbitintsf_s.o fixdfbitint_s.o floatbitintdf_s.o extendhfsf2_s.o extendhfdf2_s.o extendhftf2_s.o extendhfxf2_s.o extendsfdf2_s.o extendsftf2_s.o extenddftf2_s.o extendxftf2_s.o extendbfsf2_s.o trunctfhf2_s.o truncxfhf2_s.o truncdfhf2_s.o truncsfhf2_s.o trunctfsf2_s.o truncdfsf2_s.o trunctfdf2_s.o trunctfxf2_s.o trunctfbf2_s.o truncxfbf2_s.o truncdfbf2_s.o truncsfbf2_s.o trunchfbf2_s.o fixtfbitint_s.o floatbitinttf_s.o fixhfti_s.o fixunshfti_s.o floattihf_s.o floatuntihf_s.o floattibf_s.o floatuntibf_s.o eqhf2_s.o fixxfbitint_s.o floatbitinthf_s.o floatbitintbf_s.o floatbitintxf_s.o enable-execute-stack_s.o hardcfr_s.o strub_s.o heap-trampoline_s.o emutls_s.o libgcc.a -lSystem
Assertion failed: targetAtom != NULL (/workspace/srcdir/cctools-port/cctools/ld64/src/ld/ld.hpp: Fixup: 902)
collect2: fatal error: ld terminated with signal 6 [Aborted]
compilation terminated.
sandbox:${WORKSPACE}/srcdir/gcc_stage1/x86_64-apple-darwin14/libgcc # /workspace/srcdir/gcc_stage1/./gcc/xgcc -B/workspace/srcdir/gcc_stage1/./gcc/ -B/workspace/destdir/x86_64-apple-darwin14/bin/ -B/workspace/destdir/x86_64-apple-darwin14/lib/ -isystem /workspace/destdir/x86_64-apple-darwin14/include -isystem /workspace/destdir/x86_64-apple-darwin14/sys-include    -O2  -g -O2 -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE  -W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem ./include  -mmacosx-version-min=10.6 -fno-common -g -DIN_LIBGCC2 -fbuilding-libgcc -fno-stack-protector   -dynamiclib -nodefaultlibs -install_name /workspace/destdir/x86_64-apple-darwin14/lib/libgcc_s.1.1.dylib -single_module -o ./libgcc_s.dylib -Wl,-exported_symbols_list,libgcc.map -compatibility_version 1 -current_version 1.1 -nodefaultrpaths -g -O2 -B./ _muldi3_s.o _negdi2_s.o _lshrdi3_s.o _ashldi3_s.o _ashrdi3_s.o _cmpdi2_s.o _ucmpdi2_s.o _clear_cache_s.o _trampoline_s.o __main_s.o _absvsi2_s.o _absvdi2_s.o _addvsi3_s.o _addvdi3_s.o _subvsi3_s.o _subvdi3_s.o _mulvsi3_s.o _mulvdi3_s.o _negvsi2_s.o _negvdi2_s.o _ctors_s.o _ffssi2_s.o _ffsdi2_s.o _clz_s.o _clzsi2_s.o _clzdi2_s.o _ctzsi2_s.o _ctzdi2_s.o _popcount_tab_s.o _popcountsi2_s.o _popcountdi2_s.o _paritysi2_s.o _paritydi2_s.o _powisf2_s.o _powidf2_s.o _powixf2_s.o _powitf2_s.o _mulsc3_s.o _muldc3_s.o _mulxc3_s.o _multc3_s.o _divsc3_s.o _divdc3_s.o _divxc3_s.o _divtc3_s.o _bswapsi2_s.o _bswapdi2_s.o _clrsbsi2_s.o _clrsbdi2_s.o _mulbitint3_s.o _fixunssfsi_s.o _fixunsdfsi_s.o _fixunsxfsi_s.o _fixsfdi_s.o _fixdfdi_s.o _fixxfdi_s.o _fixunssfdi_s.o _fixunsdfdi_s.o _fixunsxfdi_s.o _floatdisf_s.o _floatdidf_s.o _floatdixf_s.o _floatundisf_s.o _floatundidf_s.o _floatundixf_s.o _fixsfti_s.o _fixdfti_s.o _fixxfti_s.o _fixtfti_s.o _fixunssfti_s.o _fixunsdfti_s.o _fixunsxfti_s.o _fixunstfti_s.o _floattisf_s.o _floattidf_s.o _floattixf_s.o _floattitf_s.o _floatuntisf_s.o _floatuntidf_s.o _floatuntixf_s.o _floatuntitf_s.o _divdi3_s.o _moddi3_s.o _divmoddi4_s.o _udivdi3_s.o _umoddi3_s.o _udivmoddi4_s.o _udiv_w_sdiv_s.o _divmodbitint4_s.o darwin-64_s.o cpuinfo_s.o sfp-exceptions_s.o _divhc3_s.o _mulhc3_s.o addtf3_s.o divtf3_s.o eqtf2_s.o getf2_s.o letf2_s.o multf3_s.o negtf2_s.o subtf3_s.o unordtf2_s.o fixtfsi_s.o fixunstfsi_s.o floatsitf_s.o floatunsitf_s.o fixtfdi_s.o fixunstfdi_s.o floatditf_s.o floatunditf_s.o fixtfti_s.o fixunstfti_s.o floattitf_s.o floatuntitf_s.o fixsfbitint_s.o floatbitintsf_s.o fixdfbitint_s.o floatbitintdf_s.o extendhfsf2_s.o extendhfdf2_s.o extendhftf2_s.o extendhfxf2_s.o extendsfdf2_s.o extendsftf2_s.o extenddftf2_s.o extendxftf2_s.o extendbfsf2_s.o trunctfhf2_s.o truncxfhf2_s.o truncdfhf2_s.o truncsfhf2_s.o trunctfsf2_s.o truncdfsf2_s.o trunctfdf2_s.o trunctfxf2_s.o trunctfbf2_s.o truncxfbf2_s.o truncdfbf2_s.o truncsfbf2_s.o trunchfbf2_s.o fixtfbitint_s.o floatbitinttf_s.o fixhfti_s.o fixunshfti_s.o floattihf_s.o floatuntihf_s.o floattibf_s.o floatuntibf_s.o eqhf2_s.o fixxfbitint_s.o floatbitinthf_s.o
floatbitintbf_s.o floatbitintxf_s.o enable-execute-stack_s.o hardcfr_s.o strub_s.o heap-trampoline_s.o emutls_s.o libgcc.a -lSystem
sandbox:${WORKSPACE}/srcdir/gcc_stage1/x86_64-apple-darwin14/libgcc #

Changing -mmacosx-version-min=10.5 to -mmacosx-version-min=10.6 does fix the issue...

I see other people over at conda-forge/gfortran_impl_osx-64-feedstock#82 had exactly the same issue, sigh.

Edit 2: reproducer reduced to ld call only thanks to -Wl,-v:

sandbox:${WORKSPACE}/srcdir/gcc_stage1/x86_64-apple-darwin14/libgcc # /workspace/destdir/bin/x86_64-apple-darwin14-ld -demangle -syslibroot /workspace/destdir/x86_64-apple-darwin14/sys-root/ -dynamic -arch x86_64 -platform_version macos 10.5 0.0 -o ./libgcc_s.dylib -ldylib1.10.5.o -L/workspace/srcdir/gcc_stage1/./gcc -L. -exported_symbols_list libgcc.map _muldi3_s.o _negdi2_s.o _lshrdi3_s.o _ashldi3_s.o _ashrdi3_s.o _cmpdi2_s.o _ucmpdi2_s.o _clear_cache_s.o _trampoline_s.o __main_s.o _absvsi2_s.o _absvdi2_s.o _addvsi3_s.o _addvdi3_s.o _subvsi3_s.o _subvdi3_s.o _mulvsi3_s.o _mulvdi3_s.o _negvsi2_s.o _negvdi2_s.o _ctors_s.o _ffssi2_s.o _ffsdi2_s.o _clz_s.o _clzsi2_s.o _clzdi2_s.o _ctzsi2_s.o _ctzdi2_s.o _popcount_tab_s.o _popcountsi2_s.o _popcountdi2_s.o _paritysi2_s.o _paritydi2_s.o _powisf2_s.o _powidf2_s.o _powixf2_s.o _powitf2_s.o _mulsc3_s.o _muldc3_s.o _mulxc3_s.o _multc3_s.o _divsc3_s.o _divdc3_s.o _divxc3_s.o _divtc3_s.o _bswapsi2_s.o _bswapdi2_s.o _clrsbsi2_s.o _clrsbdi2_s.o _mulbitint3_s.o _fixunssfsi_s.o _fixunsdfsi_s.o _fixunsxfsi_s.o _fixsfdi_s.o _fixdfdi_s.o _fixxfdi_s.o _fixunssfdi_s.o _fixunsdfdi_s.o _fixunsxfdi_s.o _floatdisf_s.o _floatdidf_s.o _floatdixf_s.o _floatundisf_s.o _floatundidf_s.o _floatundixf_s.o _fixsfti_s.o _fixdfti_s.o _fixxfti_s.o _fixtfti_s.o _fixunssfti_s.o _fixunsdfti_s.o _fixunsxfti_s.o _fixunstfti_s.o _floattisf_s.o _floattidf_s.o _floattixf_s.o _floattitf_s.o _floatuntisf_s.o _floatuntidf_s.o _floatuntixf_s.o _floatuntitf_s.o _divdi3_s.o _moddi3_s.o _divmoddi4_s.o _udivdi3_s.o _umoddi3_s.o _udivmoddi4_s.o _udiv_w_sdiv_s.o _divmodbitint4_s.o darwin-64_s.o cpuinfo_s.o sfp-exceptions_s.o _divhc3_s.o _mulhc3_s.o addtf3_s.o divtf3_s.o eqtf2_s.o getf2_s.o letf2_s.o multf3_s.o negtf2_s.o subtf3_s.o unordtf2_s.o fixtfsi_s.o fixunstfsi_s.o floatsitf_s.o floatunsitf_s.o fixtfdi_s.o fixunstfdi_s.o floatditf_s.o floatunditf_s.o fixtfti_s.o fixunstfti_s.o floattitf_s.o floatuntitf_s.o fixsfbitint_s.o floatbitintsf_s.o fixdfbitint_s.o floatbitintdf_s.o extendhfsf2_s.o extendhfdf2_s.o extendhftf2_s.o extendhfxf2_s.o extendsfdf2_s.o extendsftf2_s.o extenddftf2_s.o extendxftf2_s.o extendbfsf2_s.o trunctfhf2_s.o truncxfhf2_s.o truncdfhf2_s.o truncsfhf2_s.o trunctfsf2_s.o truncdfsf2_s.o trunctfdf2_s.o trunctfxf2_s.o trunctfbf2_s.o truncxfbf2_s.o truncdfbf2_s.o truncsfbf2_s.o trunchfbf2_s.o fixtfbitint_s.o floatbitinttf_s.o fixhfti_s.o fixunshfti_s.o floattihf_s.o floatuntihf_s.o floattibf_s.o floatuntibf_s.o eqhf2_s.o fixxfbitint_s.o floatbitinthf_s.o floatbitintbf_s.o floatbitintxf_s.o enable-execute-stack_s.o hardcfr_s.o strub_s.o heap-trampoline_s.o emutls_s.o libgcc.a -lSystem -dylib -dylib_compatibility_version 1 -dylib_current_version 1.1 -dylib_install_name /workspace/destdir/x86_64-apple-darwin14/lib/libgcc_s.1.1.dylib -single_module
Assertion failed: targetAtom != NULL (/workspace/srcdir/cctools-port/cctools/ld64/src/ld/ld.hpp: Fixup: 902)
Aborted

GDB gives me a very unhelpful backtrace:

Thread 1 "x86_64-apple-da" received signal SIGABRT, Aborted.
0x00007ffff7fad455 in setjmp () from /lib/ld-musl-x86_64.so.1
(gdb) bt
#0  0x00007ffff7fad455 in setjmp () from /lib/ld-musl-x86_64.so.1
#1  0x00007ffff7fad5b0 in raise () from /lib/ld-musl-x86_64.so.1
#2  0x0000003000000010 in ?? ()
#3  0x00007fffffffd530 in ?? ()
#4  0x00007fffffffd470 in ?? ()
#5  0x0000000000000000 in ?? ()

[eschnett]

Here is what my new best friend ChatGPT has to say on this issue:
"7. macOS Version
"The -mmacosx-version-min=10.5 flag you're using specifies the minimum macOS version for the build. If you're building on a newer macOS version, there might be issues related to backward compatibility.
Solution: Try adjusting the minimum macOS version to something more recent, such as 10.9 or 10.14, depending on the environment you're targeting."

[eschnett]

Why do we care about 10.5? Julia requires 10.14, and we use the 10.12 SDK?

[giordano]

Why do we care about 10.5? Julia requires 10.14, and we use the 10.12 SDK?

We don't, it's GCC build script which is forcing it: gcc-mirror/gcc@20b8779#diff-5c91aee3ddbc0ea342f51c839978c4fb0bc1a1bb1bf2d160cbcb062431141c79R251, as I said above.

[giordano]

With https://github.com/JuliaPackaging/Yggdrasil/blob/4fb6b23d2032478e41902dba334cb94de2d5c691/0_RootFS/GCCBootstrap%4014/bundled/patches/gcc/gcc1420-Increase-fallback-min-macOS-version-for-libgcc-to-10.8.patch I can build gcc 14 for x86_64-darwin successfully.

[giordano]

I'm not really sure we need to install all of llvm here (which also takes a while to compile), my understanding is that we only need libtapi (for cctools), which can be installed with just install-libtapi, but we also need the header files for cctools below, and that's more complicated, we should eventually have

sandbox:${WORKSPACE}/srcdir/apple-libtapi/build # ls ${includedir}/tapi/
APIVersion.h           Defines.h              LinkerInterfaceFile.h  PackedVersion32.h      Symbol.h               Version.h              Version.inc            tapi.h

All the *.h files should be in src/tapi/include/tapi, but Version.inc is generated inside the build directory.

[giordano]

Uhm, now I'm having troubles compiling glibc for x86_64-linux at

Yggdrasil/0_RootFS/gcc_common.jl

Line 528 in c92735c

make -j${nproc}

[18:06:22] In file included from readlib.c:37:
[18:06:22] ../sysdeps/unix/sysv/linux/x86_64/ldconfig.h:27:30: error: ‘FLAG_ELF_MUSL’ undeclared here (not in a function); did you mean ‘FLAG_ELF’?
[18:06:22]    27 |   { "libc.musl-x86_64.so.1", FLAG_ELF_MUSL }
[18:06:22]       |                              ^~~~~~~~~~~~~
[18:06:22] readlib.c:61:3: note: in expansion of macro ‘SYSDEP_KNOWN_LIBRARY_NAMES’
[18:06:22]    61 |   SYSDEP_KNOWN_LIBRARY_NAMES
[18:06:22]       |   ^~~~~~~~~~~~~~~~~~~~~~~~~~
[...]
[18:06:22] make[2]: *** [../o-iterator.mk:9: /workspace/srcdir/glibc_build/elf/readlib.o] Error 1
[18:06:22] make[2]: *** Waiting for unfinished jobs....

Failed command should be

sandbox:${WORKSPACE}/srcdir/glibc-2.17/elf # x86_64-linux-gnu-gcc readlib.c -c -std=gnu99 -fgnu89-inline  -D_GNU_SOURCE -O2 -Wall -Winline -Wno-array-parameter -Wno-builtin-declaration-mismatch -Wno-implicit-function-declaration -Wno-implicit-int -Wno-int-conversion -Wwrite-strings -fmerge-all-constants -frounding-math -g -Wstrict-prototypes         -I../include -I/workspace/srcdir/glibc_build/elf -I/workspace/srcdir/glibc_build -I../sysdeps/unix/sysv/linux/x86_64/64/nptl -I../sysdeps/unix/sysv/linux/x86_64/64 -I../nptl/sysdeps/unix/sysv/linux/x86_64 -I../nptl/sysdeps/unix/sysv/linux/x86 -I../sysdeps/unix/sysv/linux/x86 -I../sysdeps/unix/sysv/linux/x86_64 -I../sysdeps/unix/sysv/linux/wordsize-64 -I../nptl/sysdeps/unix/sysv/linux -I../nptl/sysdeps/pthread -I../sysdeps/pthread -I../ports/sysdeps/unix/sysv/linux -I../sysdeps/unix/sysv/linux -I../sysdeps/gnu -I../sysdeps/unix/inet -I../nptl/sysdeps/unix/sysv -I../ports/sysdeps/unix/sysv -I../sysdeps/unix/sysv -I../sysdeps/unix/x86_64 -I../nptl/sysdeps/unix -I../ports/sysdeps/unix -I../sysdeps/unix -I../sysdeps/posix -I../nptl/sysdeps/x86_64/64 -I../sysdeps/x86_64/64 -I../sysdeps/x86_64/fpu/multiarch -I../sysdeps/x86_64/fpu -I../sysdeps/x86/fpu -I../sysdeps/x86_64/multiarch -I../nptl/sysdeps/x86_64 -I../sysdeps/x86_64 -I../sysdeps/x86 -I../sysdeps/ieee754/ldbl-96 -I../sysdeps/ieee754/dbl-64/wordsize-64 -I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754/flt-32 -I../sysdeps/wordsize-64 -I../sysdeps/ieee754 -I../sysdeps/generic -I../nptl -I../ports  -I.. -I../libio -I. -nostdinc -isystem /workspace/x86_64-linux-gnu/destdir/bin/../lib/gcc/x86_64-linux-gnu/14.2.0/include -isystem /workspace/x86_64-linux-gnu/destdir/bin/../lib/gcc/x86_64-linux-gnu/14.2.0/include-fixed -isystem /workspace/destdir/x86_64-linux-gnu/sys-root/usr/include -D_LIBC_REENTRANT -include ../include/libc-symbols.h       -o /workspace/srcdir/glibc_build/elf/readlib.o -MD -MP -MF /workspace/srcdir/glibc_build/elf/readlib.o.dt -MT /workspace/srcdir/glibc_build/elf/readlib.o
In file included from readlib.c:37:
../sysdeps/unix/sysv/linux/x86_64/ldconfig.h:27:30: error: ‘FLAG_ELF_MUSL’ undeclared here (not in a function); did you mean ‘FLAG_ELF’?
   27 |   { "libc.musl-x86_64.so.1", FLAG_ELF_MUSL }
      |                              ^~~~~~~~~~~~~
readlib.c:61:3: note: in expansion of macro ‘SYSDEP_KNOWN_LIBRARY_NAMES’
   61 |   SYSDEP_KNOWN_LIBRARY_NAMES
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~

Need to figure out where musl is coming from here.

Interestingly,

sandbox:${WORKSPACE}/srcdir/glibc-2.17/elf # tail $(realpath ../sysdeps/unix/sysv/linux/x86_64/ldconfig.h)
#include <sysdeps/generic/ldconfig.h>

#define SYSDEP_KNOWN_INTERPRETER_NAMES \
  { "/lib/ld-linux.so.2", FLAG_ELF_LIBC6 }, \
  { "/libx32/ld-linux-x32.so.2", FLAG_ELF_LIBC6 }, \
  { "/lib64/ld-linux-x86-64.so.2", FLAG_ELF_LIBC6 },
#define SYSDEP_KNOWN_LIBRARY_NAMES \
  { "libc.so.6", FLAG_ELF_LIBC6 },      \
  { "libm.so.6", FLAG_ELF_LIBC6 },  \
  { "libc.musl-x86_64.so.1", FLAG_ELF_MUSL }

but https://github.com/bminor/glibc/blob/glibc-2.17/sysdeps/unix/sysv/linux/x86_64/ldconfig.h doesn't have the last line about musl (which is also a bit surprising when compiling glibc).

Edit: ah, that's coming from

Yggdrasil/0_RootFS/GCCBootstrap@4/bundled/patches/glibc_musl_rejection.patch

Lines 154 to 176 in c92735c

	diff --git a/sysdeps/generic/ldconfig.h b/sysdeps/generic/ldconfig.h
	index 59e20dc..13c4215 100644
	--- a/sysdeps/generic/ldconfig.h
	+++ b/sysdeps/generic/ldconfig.h
	@@ -27,6 +27,7 @@
	#define FLAG_ELF 0x0001
	#define FLAG_ELF_LIBC5 0x0002
	#define FLAG_ELF_LIBC6 0x0003
	+#define FLAG_ELF_MUSL 0x0004
	#define FLAG_REQUIRED_MASK 0xff00
	#define FLAG_SPARC_LIB64 0x0100
	#define FLAG_IA64_LIB64 0x0200
	diff --git a/sysdeps/unix/sysv/linux/x86_64/ldconfig.h b/sysdeps/unix/sysv/linux/x86_64/ldconfig.h
	index c7e9ad6..abd724e 100644
	--- a/sysdeps/unix/sysv/linux/x86_64/ldconfig.h
	+++ b/sysdeps/unix/sysv/linux/x86_64/ldconfig.h
	@@ -23,4 +23,5 @@
	{ "/lib64/ld-linux-x86-64.so.2", FLAG_ELF_LIBC6 },
	#define SYSDEP_KNOWN_LIBRARY_NAMES \
	{ "libc.so.6", FLAG_ELF_LIBC6 }, \
	- { "libm.so.6", FLAG_ELF_LIBC6 },
	+ { "libm.so.6", FLAG_ELF_LIBC6 }, \
	+ { "libc.musl-x86_64.so.1", FLAG_ELF_MUSL }

but the patch to sysdeps/generic/ldconfig.h was rejected. Sigh.

How did this ever work?!? As far I could tell, the patch never applied to glibc 2.12 nor 2.17.

[giordano]

I was able to build all GCC for all platforms locally. I'm now rebuilding and publishing them (could have started doing the publishing the first time round, oh well), tomorrow I'll open the PR to BinaryBuilderBase.