http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55123
--- Comment #5 from Marc Glisse <glisse at gcc dot gnu.org> 2012-10-29 19:50:59 UTC --- The standard doesn't say anything about it, but it would also be possible to tweak one of the address overloads in allocator so there is no ambiguity when both types are the same. No idea if that's desirable though, I haven't followed the allocator work.