nikic wrote: Hm, okay, looks like this change is not right, per the failing test. I didn't realize that we're not just doing the alignment check here, but that the ubsan diagnostic handler also uses the address to show the contents of the memory. So we do end up using the pointer provenance.
https://github.com/llvm/llvm-project/pull/183538 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
