https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83388
--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> --- (In reply to Jan Hubicka from comment #7) > I believe we do not stream any references to references and once we start to > do we will need to invent machinery to keep them intact like we have for > edges and symbols. So I would go for removal of the refernece. > > I however wonder if this is the best place to remove statements from > function. I would expect either fixup_cfg or one of the IPA transform stages > to be responsible for this instead. So where exactly? Ideally in something that has to walk the whole IL again. > Why flag_sanitize is not optimization and not streamed along with function > bodies? It seems to make sense to mix sanitized and unsanitized units. We've tried that (merging options on lto reads) and it just didn't work and was a complete nightmare. There are many different sanitizers, and they are discovered and lowered and handled at different stages (FEs, gimplifier, ubsan pass, asan pass, tsan pass, sanopt pass, ...).
