efriedma-quic wrote: It doesn't seem like a good idea to make target-independent logic behave in target-specific ways; that's going to confusing for both people hacking on clang, and for users if it's user-visible. Is there some way we can make this logic consistent across targets?
Maybe we can explicitly translate the "arch" feature into enabling/disabling all the relevant features? https://github.com/llvm/llvm-project/pull/75804 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits