Closed tstellar closed 2 weeks ago
@llvm/issue-subscribers-c-20
Patch for this here: https://reviews.llvm.org/D120456
@llvm/issue-subscribers-clang-driver
@erichkeane @AaronBallman After the deferred concepts patch got merged, are there any other open concepts bugs that you'd like to prioritize before bumping up __cpp_concepts
?
I'd like to see us get p857 done, and at least a majority of the work on p2113 done before we consider concepts "done" enough to up the macro.
I believe there are patches being worked on for both, so I'd be open to a patch to change the macro as soon as those go on.
Sounds good to me :)
The status page indicates this paper was implemented in clang10. Should that be changed?
The status page indicates this paper was implemented in clang10. Should that be changed?
Given that this work is taking some time, it might make sense for us to mark the paper as being partially implemented with an explanation as to what's missing.
Any update on this? libstdc++ is gating some C++23 features (namely std::expected
) behind this.
According to https://clang.llvm.org/cxx_status.html we mention two DRs, but they aren't SUPER related here? We have some problems with deleted special member functions (of which, concepts makes WORSE), but I don't think this should hold up on concepts.
I believe it should be OK to update this feature test macro for Clang19. The std::expected
being gated behind this is unfortunate, and I think that is more important than my lingering doubt on this feature.
So: if someone submits a patch to update this in Clang, i'd accept it.
Because it's early enough in the 19.x cycle and we think our concepts support is reasonably usable, I support setting the feature test macro and giving ourselves some time to see whether that shakes out further issues we need to address before releasing with the macro enabled.
Fixed by #87998
id: P2493R0 paper: https://wg21.link/p2493r0