philnik777 wrote: > This is missing test coverage for the interesting cases. > > I'm a bit worried about how easy it will be to apply this attribute in > unexpected places. e.g., it will apply to member functions as well as free > functions, will anyone expect that though?
This is explicitly documented, so I'd say yes. This is how the pragma works for better or worse. > It also applies to special functions like `main`, though that's more a > question of the attribute and less about the pragma. I'm not sure whether this is a request for a test? > What's the behavior going to be in C when you get an implicit function > declaration in C89 mode? Does that declaration then pick up this attribute as > well? How about other special member functions we define on behalf of the > user? I'm not 100% sure what you mean by "define on behalf of users". I've added a test for an explicitly defaulted special member. Note though that using `#pragma clang attribute` is basically like explicitly defaulting the special members and adding the attribute there. https://github.com/llvm/llvm-project/pull/145653 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
