https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97844
--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Roman Lebedev from comment #2) > (In reply to Andrew Pinski from comment #1) > > unsigned integer overflow DOES NOT EXIST. It is called wrapping only. > > > This is NOT undefined at all. > I agree, and no one said it's undefined. Erm ... SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior /usr/bin/../lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/basic_string.h:403:51 in Am I hallucinating? > > clang is broken. > > It's not, actually. > `-fsanitize=undefined` does *NOT* warn on that, > `-fsanitize=integer` does, and it's explicitly marketed as sanitizing > some behavior that is explicitly *NOT* UB, but very often unintentional. So the UBsan output needs to be fixed. > So the bugreport is correct in a way, > libstdc++ for -std=c++20 contains `-fsanitize=integer` issues. Not a bug.
