huhu233 wrote: > I don't understand this. The code is a strict aliasing violation, so why > should clang work around it?
Hi, @arsenm, the violation is due to the compiler emitting different TBAA for `svst1` (long) and other users of `res2` (long long). TBAA is invisible to ACLE users, and the compiler should tell them that this kind of code may cause errors, or work around it (e.g., emit `omnipotent char` to `svst1` if there is bitcast). https://github.com/llvm/llvm-project/pull/107793 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
