http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51328
--- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> 2011-12-19 15:58:13 UTC --- (In reply to comment #3) > Thus I guess rejecting it earlier is fine, but we don't want to misled people > by saying that 'typename T::~T' would be ok. Agreed. I would expect it to be straightforward to catch this at the point we currently give the misleading message.
