http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14710
--- Comment #8 from Jason Merrill <jason at gcc dot gnu.org> 2012-03-14 18:00:33 UTC --- I wouldn't warn about a cast that changes the type or value category of an expression at all, or a cast that ends up being the same type as the result of template instantiation. But a warning for (A*)&a seems reasonable.
