The original bug doesn't reproduce on Clang, allegedly because of
https://bugs.llvm.org/show_bug.cgi?id=44484
We already test STL's exact test case, in span.cons/deduct.pass.cpp, which I'm touching just for the heck of it.
Details
- Reviewers
jloser ldionne Mordante - Group Reviewers
Restricted Project - Commits
- rG0412c007e3c2: [libc++] Implement LWG3369, tweak CTAD for std::span.
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
LGTM - thanks for looking into it.
libcxx/test/std/containers/views/span.cons/deduct.pass.cpp | ||
---|---|---|
9 | We could do a similar thing for all the files I touched in my recent span PR FYI. We're going to drop support for AppleClang 12.0.0 officially in BuildKite hopefully soon since we only need to support AppleClang 13.0.0 technically. At that point, all the UNSUPPORTED: libcpp-no-concepts go away. |
Can you add a test for the fix. For example STL's example in LWG3369? Or does the Clang bug prevent us from testing it?
libcxx/test/std/containers/views/span.cons/deduct.pass.cpp | ||
---|---|---|
56–58 | @Mordante: I believe this is the exact test case displayed in https://cplusplus.github.io/LWG/issue3369 — and yeah, I assume it passes anyway because of that Clang bug. I don't have the capability to test libc++ with non-Clang compilers to see if it fails... although I confess I don't understand why this test passes even with the gcc buildkite. |
LGTM!
libcxx/test/std/containers/views/span.cons/deduct.pass.cpp | ||
---|---|---|
56–58 | That's indeed the same test. |
I guess that means we are forward-compatible with Clang once they fix PR44484. Let's ship this.
We could do a similar thing for all the files I touched in my recent span PR FYI. We're going to drop support for AppleClang 12.0.0 officially in BuildKite hopefully soon since we only need to support AppleClang 13.0.0 technically. At that point, all the UNSUPPORTED: libcpp-no-concepts go away.