https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107476
Andrew Pinski <pinskia at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |needs-bisection Known to fail| |13.1.0 --- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> --- So it looks like there was some jump threading that happened here: if (one.0_59 > 1) goto <bb 12>; [50.00%] else goto <bb 18>; [50.00%] <bb 12> [local count: 131211251]: ivtmp.16_58 = (unsigned long) &MEM <struct TMLN[1]> [(void *)&ttx_ldsc + 520B]; <bb 13> [local count: 131211252]: # ivtmp.14_38 = PHI <ivtmp.14_5(17), 1(12)> # ivtmp.16_18 = PHI <ivtmp.16_42(17), ivtmp.16_58(12)> _9 = (struct TMLN *) ivtmp.16_18; temp_10 = tmxr_getc_ln (_9); if (temp_10 != 0) goto <bb 14>; [50.00%] else goto <bb 17>; [50.00%] But for some reason the trunk does not do that any more. Would be interesting to see when the warning disappeared from the trunk.