https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86389
--- Comment #13 from Richard Biener <rguenth at gcc dot gnu.org> --- (In reply to Jan Hubicka from comment #12) > Since the vector_subscript testcase seems indeed to be frontend issue > triggered by -fipa-pta (and soon by modref). I wonder if we should have > fortran bug tracking this or change component here? That's PR102043