[Bug c++/53492] [4.7/4.8/4.9 Regression] ICE in retrieve_specialization, at cp/pt.c:985

[reichelt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53492

Volker Reichelt <reichelt at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |accepts-invalid
   Last reconfirmed|2012-05-26 00:00:00         |2014-03-01
                 CC|                            |reichelt at gcc dot gnu.org
   Target Milestone|---                         |4.7.4
            Summary|internal compiler error: in |[4.7/4.8/4.9 Regression]
                   |retrieve_specialization, at |ICE in
                   |cp/pt.c:985                 |retrieve_specialization, at
                   |                            |cp/pt.c:985

--- Comment #4 from Volker Reichelt <reichelt at gcc dot gnu.org> ---
Here's an even simpler testcase:

=====================================================
template<typename> struct A
{
  template<typename> struct B;
};

template<> template<typename T> struct A<T>::B
{
  B(int) {}
};

A<int>::B<int> b;
=====================================================

It crashes since GCC 4.0.0, but was correctly rejected before.
So this qualifies as a regression.

A slightly modified and still invalid testcase also crashes since GCC 4.0.0
but was wrongly accepted before:

=====================================================
template<typename> struct A
{
  template<typename> struct B;
};

template<> template<typename T> struct A<T>::B
{
  B() {}
};

A<int>::B<int> b;
=====================================================

Another variation is wrongly accepted since at least GCC 3.2:

=====================================================
template<typename> struct A
{
  template<typename> struct B;
};

template<> template<typename T> struct A<T>::B {};

A<int>::B<int> b;
=====================================================