https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115715
--- Comment #4 from Xi Ruoyao <xry111 at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #3) > Confirmed. This is a QoI issue though. We should do more rejections on > non-dependent types inside templates. Should we open an enhancement like "reject templates which will be ill-formed in a way independent to template parameters" and mark all these reports duplicates of it? Otherwise it's too easy to construct 1000+ different cases and filling up the bugzilla.