https://bugs.kde.org/show_bug.cgi?id=374568
--- Comment #16 from RJVB <rjvber...@gmail.com> --- So that's why we don't see it! And so in optimised builds there is in fact no bug in the compiled code. If you have the time and opportunity, could you check if you get the crash with other compilers and -O0 (GCC versions and clang)? It strikes me as really odd that a compiler won't warn about a situation for which it can generate correct code when optimising but not (intentionally?) without optimisation. FWIW, Gcc has `-fisolate-erroneous-path-*`, -fsanitize-undefined-trap-on-error (maybe you already used that?) and -Wsequence-point that are all related to catching undefined behaviour. -- You are receiving this mail because: You are watching all bug changes.
