jryans wrote: > If it's useful to keep `noipa` orthogonal at the LLVM level, why is it not > similarly useful at the Clang level?
Ah yes, I agree there is potential utility in keeping `noipa` orthogonal at Clang level as well. I don't have any experience with new Clang attributes, so I assumed (perhaps wrongly) before this discussion that Clang may prefer to closely match the meaning that GCC has given to the source-level `noipa` attribute. If we are free to deviate from GCC's semantics for the source-level attribute, then yes, I suggest preserving orthogonal control at the Clang level. So, the source-level attribute would then only need to set `noipa` at the LLVM IR level. https://github.com/llvm/llvm-project/pull/207502 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
