https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78039
--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> --- (In reply to Martin Sebor from comment #5) > Sorry, that was the wrong link. The correct link to the email thread where > the approval was requested is here: > https://gcc.gnu.org/ml/gcc-patches/2016-10/msg01094.html > I don't see Jason's approval in the archive but I still have it in my Inbox. That doesn't mean it was a good idea to backport the whole patch. Sure, there is a regression, so you should backport the part of the patch that resolves that regression (ICE, or rejects-valid, ...), but not other parts of the patch that reject what has been previously accepted. People are really not willing to adjust their code in between snapshots from release branches because it got more and more strict in what it accepts. Such changes belong to trunk only.
