https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102201
--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> --- CD1 was a C++0x draft: "CD1: A DR issue not resolved in TC1 but included in Committee Draft 1. CD1 was advanced for balloting at the September, 2008 WG21 meeting." DR 613 was resolved by N2253, which was a post-C++03 change, and wasn't adopted as a defect report. So arguably Clang is right to reject it in C++03 mode. We could make it a pedwarn, but I think WONTFIX is probably right. There seems to be no advantage to rejecting it (strictly conforming C++98 programs aren't going to be using SFINAE tricks that could detect this).