Summary: | sys-devel/gcc: support for musl | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Luca Barbato <lu_zero> |
Component: | [OLD] Core system | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
Status: | RESOLVED FIXED | ||
Severity: | enhancement | CC: | ale, bugzie, erikdenstore+gbugs, gentoo, nikoli, toast+misc, tsmksubc |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | http://gcc.gnu.org/PR55807 | ||
See Also: | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55807 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 430702, 571700 |
Description
Luca Barbato
2012-12-25 20:24:12 UTC
patches should be posted to gcc-patches@gcc.gnu.org, not their bugzilla (In reply to Luca Barbato from comment #0) > Musl needs few patches to work with gcc. I already reported them upstream. Luca, did you ever email these patches to gcc-patches@gcc.gnu.org? I see your last comment on the upstream bug is that you're trying to get the author to submit the patches himself. Did he/she submit them? There is a hacky workaround that I've been using. And that is to make a sym link from /lib{,64}/ld-linux-x86-64.so.2 -> /usr/lib/libc.so but we should get it right. how it look these days, is upstream bug still valid? (marked as NEW) First patches are going upstream: http://gcc.gnu.org/ml/gcc-patches/2015-04/msg01068.html 4.9.3 has been stabilized, and still seems to require patches (In reply to Chloe Kudryavtsev from comment #5) > 4.9.3 has been stabilized, and still seems to require patches yes i will be forward porting those patches soon on the musl overlay. as far as upstream goes, it will probably be a long wait. Currently, we are applying very few patches to gcc. This can cause issues, such as in #571700 (DWARF debug info mismatches). In the case of bug 571700, the patch required is found here: https://github.com/GregorR/musl-gcc-patches/blob/master/gcc-config-musl.diff We can see that alpine uses this patch here http://git.alpinelinux.org/cgit/aports/tree/main/gcc/202-musl-config-v3.patch . Is there any reason we aren't using a full patch set, as we await upstream to improve the situation? (In reply to Chloe Kudryavtsev from comment #7) > Currently, we are applying very few patches to gcc. This can cause issues, > such as in #571700 (DWARF debug info mismatches). > > In the case of bug 571700, the patch required is found here: > https://github.com/GregorR/musl-gcc-patches/blob/master/gcc-config-musl.diff > > We can see that alpine uses this patch here > http://git.alpinelinux.org/cgit/aports/tree/main/gcc/202-musl-config-v3. > patch . > > Is there any reason we aren't using a full patch set, as we await upstream > to improve the situation? I'm not against using GregorR's patchset, but we must merge it with ours, in particular with gcc-4.9.3-musl-linker-path.patch else it will break my stages. If you want to propose a merged patchset for 4.9.3 we can test it and see if it works. Then we can maintain it moving forward. This can probably be closed now (since musl support is in gcc-6 and later). |