https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87813
--- Comment #6 from Aldy Hernandez <aldyh at gcc dot gnu.org> --- (In reply to Martin Sebor from comment #5) > The sprintf pass doesn't do any optimization at -O0 but it still runs to > diagnose the subset of mistakes that are detectable even without > optimization. Yes, but my point is that the way things are laid out right now, we are running full evrp at -Og and *that* seems like an optimization, because it's affecting the IL later on.
