https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82518
--- Comment #27 from Aldy Hernandez <aldyh at gcc dot gnu.org> --- (In reply to Christophe Lyon from comment #26) > I've manually built or tried to build several revisions: > * 197671: build OK, test fails to run at -fno-vect-cost-model -O3 -g > * 197669: same (!) > * 197815: GCC fails to build > * 197816: same > * 197900: same > > So.... although I still see the test failing, I don't understand why bisect > said the guilty commit would be between 197671 and 197815, given that the > test fails at 197669.... What's the first revision before 197669 that shows the test succeeding?
