[Bug c++/12216] New: Bad "implicit typename" warning

[Bug c++/12216] New: Bad "implicit typename" warning

[igodard at pacbell dot net]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12216

           Summary: Bad "implicit typename" warning
           Product: gcc
           Version: 3.3.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: igodard at pacbell dot net
                CC: gcc-bugs at gcc dot gnu dot org

3.3.1 warns about implicit typename on identifier declared in an (inherited) typedef. A typedef can't be anything but a type, so the warning is incorrect. If you attempt to remove the warning by uncommenting the "/* typename */" then you get a compile error.

Ivan

[Bug c++/12216] Bad "implicit typename" warning

[igodard at pacbell dot net]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12216



------- Additional Comments From igodard at pacbell dot net  2003-09-08 21:02 -------
Created an attachment (id=4720)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=4720&action=view)
Compiler output (-v -save-temps)

[Bug c++/12216] Bad "implicit typename" warning

[igodard at pacbell dot net]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12216



------- Additional Comments From igodard at pacbell dot net  2003-09-08 21:03 -------
Created an attachment (id=4721)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=4721&action=view)
Source code (-save-temps)

[Bug c++/12216] Bad "implicit typename" warning

[bangerth at dealii dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12216


bangerth at dealii dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID


------- Additional Comments From bangerth at dealii dot org  2003-09-08 21:41 -------
No. Given your code
-------------------------
template<typename I>
class C : public std::iterator<
              typename std::iterator_traits<I>::Category,
              typename std::iterator_traits<I>::value_type,
              typename std::iterator_traits<I>::difference_type,
              typename std::iterator_traits<I>::pointer,
              typename std::iterator_traits<I>::reference
                                                        > {
public:
                                C(I i) : iter(i) {}

        value_type operator*() const { return *iter; }
private:
        I iter;
        };

int main() {
        }

---------------------------------
the typedef value_type is not visible with two-stage lookup. Your code
is thus invalid, but gcc up to and including 3.3.x accepts it. You
really need to write the thing with the typename, but including the
classname, i.e.
  typename C<I>::value_type
or
  typename std::iterator<...>::value_type

The point is that you need to make the name of the type dependent, so
that it is looked up on instantiation, rather than declaration time. So
the thing before :: must be template dependent, but then you also need
the typename.

W.

[Bug c++/12216] Bad "implicit typename" warning

[igodard at pacbell dot net]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12216



------- Additional Comments From igodard at pacbell dot net  2003-09-09 07:02 -------
Subject: Re:  Bad "implicit typename" warning

Understood. Thanks.

Ivan

----- Original Message ----- 
From: "bangerth at dealii dot org" <gcc-bugzilla@gcc.gnu.org>
To: <igodard@pacbell.net>
Sent: Monday, September 08, 2003 2:41 PM
Subject: [Bug c++/12216] Bad "implicit typename" warning


> PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12216
>
>
> bangerth at dealii dot org changed:
>
>            What    |Removed                     |Added
> --------------------------------------------------------------------------
--
>              Status|UNCONFIRMED                 |RESOLVED
>          Resolution|                            |INVALID
>
>
> ------- Additional Comments From bangerth at dealii dot org  2003-09-08
21:41 -------
> No. Given your code
> -------------------------
> template<typename I>
> class C : public std::iterator<
>               typename std::iterator_traits<I>::Category,
>               typename std::iterator_traits<I>::value_type,
>               typename std::iterator_traits<I>::difference_type,
>               typename std::iterator_traits<I>::pointer,
>               typename std::iterator_traits<I>::reference
>                                                         > {
> public:
>                                 C(I i) : iter(i) {}
>
>         value_type operator*() const { return *iter; }
> private:
>         I iter;
>         };
>
> int main() {
>         }
>
> ---------------------------------
> the typedef value_type is not visible with two-stage lookup. Your code
> is thus invalid, but gcc up to and including 3.3.x accepts it. You
> really need to write the thing with the typename, but including the
> classname, i.e.
>   typename C<I>::value_type
> or
>   typename std::iterator<...>::value_type
>
> The point is that you need to make the name of the type dependent, so
> that it is looked up on instantiation, rather than declaration time. So
> the thing before :: must be template dependent, but then you also need
> the typename.
>
> W.
>
>
>
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.