Re: c++/10449: [3.4][New parser] Totally bogus error message on missing typename

Re: c++/10449: [3.4][New parser] Totally bogus error message on missing typename

[bangerth]

Synopsis: [3.4][New parser] Totally bogus error message on missing typename

State-Changed-From-To: open->analyzed
State-Changed-By: bangerth
State-Changed-When: Tue Apr 22 23:00:42 2003
State-Changed-Why:
    I think the message is actually not so unclear: you try to
    call
        return Identity<T>::type(t);
    Since you don't have the missing "typename" here, the
    compiler assumes that Identity<T>::type is either a member
    function of Identity<T>, or a pointer-to-function. The
    argument is Foo&, so that's why you get
        error: no match for call to `(Foo) (Foo&)'
    since Identity::type is actually "Foo". So I guess the
    only complaint could be that the compiler shouldn't
    expand the typedef (that it shouldn't see here, after all),
    and rather print something like
        error: no match for call to `Identity<T>::type(Foo&)'
    But then, this wouldn't be much clearer either.
    
    I think the only thing we could do is to check if we such
    a match cannot be found, whether there's a type of that 
    name and then issue a "missing typename?" message. But
    that would amount to re-inventing the implicit typename
    machinery, which we just got rid of.
    
    I leave it to the experts to decide what to do here...
    
    W.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10449