https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126075
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> --- I would have expected, that without -fnon-call-exceptions, they are transformed to the same IL at some point. It might be that the short-circuiting is correctly preserved with -ftrapping-math because the compare might set IEEE flags?
