Re: c++/9464: [3.4 regression] ICE on missing typename

Re: c++/9464: [3.4 regression] ICE on missing typename

[bangerth]

Old Synopsis: ICE on missing typename
New Synopsis: [3.4 regression] ICE on missing typename

State-Changed-From-To: open->analyzed
State-Changed-By: bangerth
State-Changed-When: Mon Jan 27 22:00:59 2003
State-Changed-Why:
    Confirmed. Fallout from the implicit typename deprecation.
    Here's a similar snippet, which I think is even legal since
    we don't instantiate for a particular value of T (well, legal except for the missing typename, of course):
    -----------------------------------
    template<typename T, typename A> struct X {
        typedef typename A::template foo<T>::type type;
    };
    
    template<typename T, typename = int> struct Y {
        typedef double type;
        typename X<T,T>::type t;
    };
    
    template<class T>
    struct Z {
        typedef Y<T>::type type;
    };
    -------------------------
    
    g/a> /home/bangerth/bin/gcc-3.4-pre/bin/c++ -c x.ii
    x.ii: In instantiation of `Y<T, int>':
    x.ii:12:   instantiated from here
    x.ii:7: error: `Y<T, <anonymous template type parameter> >::t' has incomplete
       type
    x.ii:7: internal compiler error: in cxx_incomplete_type_diagnostic, at
       cp/typeck2.c:278
    Please submit a full bug report,
    with preprocessed source if appropriate.
    
    
    Since an ICE is worse than an error message, I rate this
    as a regression. It gave warnings about the implicit
    typename stuff with 3.3 and 3.2.2, and worked with 2.95
    
    W.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=9464