Details
- Reviewers
Mordante - Group Reviewers
Restricted Project Restricted Project - Commits
- rGf29c54998d23: [libc++] Fix problems with GCC 13 and switch to it in the CI
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
libcxx/test/std/utilities/meta/meta.rel/is_convertible.pass.cpp | ||
---|---|---|
117–126 | This code seems bogus to me and GCC disagrees with clang here. I've filed https://github.com/llvm/llvm-project/issues/64402, since I'm quite certain that the typedef above shouldn't be accepted. If the clang folks agree with me I'll also file a bug report against GCC. |
Fixes
libcxx/test/std/utilities/meta/meta.rel/is_convertible.pass.cpp | ||
---|---|---|
117–126 | Never mind. This is actually valid code and GCC probably has a bug in __is_convertible. |
Thanks for working on this! Please update the official supported GCC version in the libc++ documentation,
libcxx/test/std/utilities/meta/meta.rel/is_convertible.pass.cpp | ||
---|---|---|
118 | All versions of GCC or just 13? |
libcxx/test/std/utilities/meta/meta.rel/is_convertible.pass.cpp | ||
---|---|---|
118 | Just GCC 13 |
In general happy, but I would like to have a short look after my comment has been addressed.
libcxx/test/std/utilities/meta/meta.rel/is_convertible.pass.cpp | ||
---|---|---|
118 | Maybe add that information in the comment and in the header too. something like |
Address comments
libcxx/test/std/utilities/meta/meta.rel/is_convertible.pass.cpp | ||
---|---|---|
118 | It's already fixed in the GCC trunk. I've just made this an LLVM 19 TODO, since we will probably switch to GCC 14 then. |
LGTM!
libcxxabi/test/catch_member_function_pointer_02.pass.cpp | ||
---|---|---|
18 | After you land this can you make a follow-up draft review to remove the gcc-11 and gcc-12 XFAILS here and elsewhere. |
libcxxabi/test/catch_member_function_pointer_02.pass.cpp | ||
---|---|---|
18 | That was my plan. |
This code seems bogus to me and GCC disagrees with clang here. I've filed https://github.com/llvm/llvm-project/issues/64402, since I'm quite certain that the typedef above shouldn't be accepted. If the clang folks agree with me I'll also file a bug report against GCC.