https://gcc.gnu.org/bugzilla/show_bug.cgi?id=39612
--- Comment #37 from Richard Biener <rguenth at gcc dot gnu.org> --- Eww. Fixing the missed optimization causes tons of maybe-uninit warnings. That was probably to be expected since we're introducing undefs on the loop entry edge and conditional uses. We eventually should mark the conditional stores on the exits with no-warning...
