https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78731
--- Comment #2 from Jim Wilson <wilson at gcc dot gnu.org> --- The compiler is failing in jump threading as called by dom2. Jeff's 2015-09-30 patch is triggering in dom1, and changing the CFG just enough to prevent the jump threading bug from triggering. So it isn't a proper fix. Disabling it, and bisecting again, I found bug 67892, another patch from Jeff, which looks like a proper fix. However, there is a problem. The bug was reported against gcc 5, fixed in gcc 6, and then closed, but is still broken in gcc 5. I think it is in the wrong state. Presumably, it should still be open, and then fixed in gcc 5 if reasonably safe, or documented as a known bug that we won't fix in gcc 5.
