https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80635
--- Comment #74 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #73) > So coming back to this. I think we can transform `VCE<bool>(a) != 0` always > to just `a != 0` even without worring about the range of `a` here as if it > was outside of 0/1 then it would be undefined and that is what my patch > attached in PR 105749 does. That fixes the issue even at -O1. Patch posted for the warning at -O1: https://gcc.gnu.org/pipermail/gcc-patches/2025-October/696763.html
