------- Comment #1 from dgregor at gcc dot gnu dot org 2008-02-15 03:35 ------- This test case actually exposes two (!) related issues, one of which is a rejects-valid:
template<typename... T, template<T...> class X> void foo(X<0>); and the other, which is an ice-on-valid-code: template<typename... T, template<T...p> class X> void foo(X<0>); The issue for both the original test case and this last test case is that coerce_template_parameter_pack doesn't expect to get a pack expansion back from tsubst_pack_expansion, but of course this happens in nested templates like this template template parameter. Note to myself: the parameter pack SIZEOF_EXPR handling doesn't deal with pack expansions coming back from tsubst_pack_expansion, either. I can probably craft a test case to illustrate this issue. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35022
