https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104506
--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #2) > Now do we want to add a check inside STRIP_USELESS_TYPE_CONVERSION for > error_mark_node, I am still deciding. Oh STRIP_TYPE_NOPS has a check for error_mark_node which was done in 2000 by r0-29080 . So maybe it just should too.