https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69857
--- Comment #8 from David Binderman <dcb314 at hotmail dot com> --- (In reply to ktkachov from comment #7) > Yes, that's an approach that can be taken. Once such a case is found, you > could also try using a reducer program like creduce to create a small > testcase appropriate for the testsuite. > > Removing the return there would probably be an optimisation, which is why I > recommended it for GCC 7. GCC 6 is now in a regression fixing stage, so for > it to go in now we would need to find a testcase where we were performing an > optimisation before r197530 and after r197530 (with the bogus return) we > stopped doing. Considering that r197530 was supposed to introduce that > optimisation, it's unlikely that such a case can be found. GCC 6 has been produced. Time to remove this return ?
