heiher wrote: > This change sounds like it might introduce an ABI change, can you confirm > that this is not the case before I accept this into the release branch?
Confirmed, it's an ABI change. Turns out the old behvior never actually matched GCC in the first place. This was just a bug where integer complex types got treated like FP complex types and split across GARs when they shouldn't have been. Since it's fixing something that was already broken, I'd rather get it in sooner than later so we don't keep emitting code that doesn't match GCC's ABI. Thanks! https://github.com/llvm/llvm-project/pull/216034 _______________________________________________ llvm-branch-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
