https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80362
--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> --- (In reply to Alexandre Oliva from comment #9) > I see the same issue of using arg vs op is present in the reciprocal tests > right below the patched ones. AFAICT it is not possible to trigger the > problem in a way that affects the outcome, but... is there any good reason > to keep them different and have people puzzle over the difference? I think > it was not the first time that difference caught my attention and got me > distracted trying to reason it out. Feel free to patch it.
