https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61428
--- Comment #5 from Aldy Hernandez <aldyh at gcc dot gnu.org> --- (In reply to Jeffrey A. Law from comment #4) > Aldy, here's another one to try with the new backwards threader. I haven't done any in depth analysis here, but with the new threader we get one threaded path in the early threader, as well as one threaded path in thread1. Neither of these threads was previously done by the backwards threader. Also, there is NO warning emitted :).
