https://github.com/ojhunt requested changes to this pull request.
I think this PR is wrong. I had assumed that this was a win32 PR because long double is fp64 on win64 so it didn't occur to me to think of this as a win64-ism. On win32, where long double is fp80, clang respects the win32 ABI's 8 byte alignment of long doubles. There is no re-arrangement/ignorance of packing, the type is 8 byte aligned. The problem is that your test case is hitting is that it is using an ABI breaking flag. I asked whether the behavior was correct here: https://github.com/llvm/llvm-project/pull/208256#discussion_r3696552254 but it was resolved without answer. I can't see any evidence of any per-type alignment overrides in MSVC. @StephanTLavavej do you have any thoughts on this? https://github.com/llvm/llvm-project/pull/208256 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
