As discussed in D114786.
Details
- Reviewers
• Quuxplusone - Group Reviewers
Restricted Project - Commits
- rG071762199f19: [libc++] Add tests for the content of <cstddef>
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
libcxx/test/libcxx/language.support/support.types/cstddef.compile.pass.cpp | ||
---|---|---|
19 | The first was presumably a typo, but let's test both here, since otherwise we have no coverage for std::nullptr_t-in-C++03-mode. | |
libcxx/test/std/language.support/support.types/cstddef.compile.pass.cpp | ||
46 |
Address review comments.
libcxx/test/libcxx/language.support/support.types/cstddef.compile.pass.cpp | ||
---|---|---|
19 | Good catch. |
libcxx/test/std/language.support/support.types/cstddef.compile.pass.cpp | ||
---|---|---|
34 | std::nullptr_t is also provided in C++03 mode, right? So this line (and maybe the other too) should go outside the #if, so they get tested in all modes. Otherwise we have no coverage for std::nullptr_t-in-C++03-mode. |
Address review comments.
libcxx/test/std/language.support/support.types/cstddef.compile.pass.cpp | ||
---|---|---|
34 | Hmm, right. I'd still rather move that test to the libc++ specific test, just for clarity. But yeah, good catch again. I guess I don't care enough about C++03! |
All comments have been addressed, this should be fairly uncontroversial to land now. Can address post-commit if needed. Thanks for the review!
The first was presumably a typo, but let's test both here, since otherwise we have no coverage for std::nullptr_t-in-C++03-mode.
Actually, you should probably just copy-paste the other test into this one and test all the things, e.g. ::max_align_t.