gdr@integrable-solutions.net (Gabriel Dos Reis) wrote in message news:<m3u1khxj10.fsf@soliton.integrable-solutions.net>...
> algrant@myrealbox.com (Al Grant) writes:
>
> | Consider the following:
> |
> |  #1:  typedef int T; struct X { typedef T T; };
> |  #2:  typedef int T; struct X { typedef T U; typedef T T; };
> |  #3:  typedef int T; struct X { typedef T U; typedef int T; };
> |  #4:  typedef int T; struct X { typedef T U; typedef char T; };
>
> All those constructs are ill-formed (but no diagnostic is required).
> 3.3.6/2:

Thanks, that seems clear enough.  There is even an example of #1.
Since GNU invoked 3.4.1/7 that was the place I looked.

---
[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://www.jamesd.demon.co.uk/csc/faq.html                       ]

Consider the following:

 #1:  typedef int T; struct X { typedef T T; };
 #2:  typedef int T; struct X { typedef T U; typedef T T; };
 #3:  typedef int T; struct X { typedef T U; typedef int T; };
 #4:  typedef int T; struct X { typedef T U; typedef char T; };

The issue is whether 3.4.1/7 forbids the use of T before, or in,
the definition of S::T.  VC++ and Comeau accept them all.
g++ 3.1 faults them all.  HP aCC accepts #1 only.

The applicable text of 3.4.1/7 is

  A name used in the definition of a class X outside of a member
  function body or nested class definition [footnote: this refers
  to unqualified names following the class name] shall be declared
  in one of the following ways
   - before its use in class X
   - ... various other nested class and namespace situations

The name T is used in the definition of X and it is declared
before its use in class X.  So that should be ok.  GNU must be
interpreting the text to mean that if T is declared in X at all,
it must be declared before any unqualified use in X, but I can't
see that's justified.  All 3.4.1/7 seems to be saying is that
lookup in class definitions (apart from member function bodies
and nested classes) works in the usual C/C++ way, i.e.
  typedef int T; struct X { typedef T U; typedef char T; };
works like
  typedef int T; void f() { typedef T U; typedef char T; }

So who's right - g++ or Comeau?

---
[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://www.jamesd.demon.co.uk/csc/faq.html                       ]

algrant@myrealbox.com (Al Grant) writes:

| Consider the following:
|
|  #1:  typedef int T; struct X { typedef T T; };
|  #2:  typedef int T; struct X { typedef T U; typedef T T; };
|  #3:  typedef int T; struct X { typedef T U; typedef int T; };
|  #4:  typedef int T; struct X { typedef T U; typedef char T; };

All those constructs are ill-formed (but no diagnostic is required).
3.3.6/2:

  A name N used in a class S shall refer to the same declaration in
  its context and when re-evaluated in the completed scope of S. No
  diagnostic is required for a violation of this rule.

[...]

| So who's right - g++ or Comeau?

Both.  Just avoid such confusing constructs.

I far more prefer a diagnostic.  YMMV.

--
Gabriel Dos Reis                    gdr@integrable-solutions.net

---
[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://www.jamesd.demon.co.uk/csc/faq.html                       ]