implicit typename warnings with gcc 3.1

implicit typename warnings with gcc 3.1

[Richard Guenther]

Hi!

template <int i>
struct foo {
        typedef int a_t;
};

template <int i>
class blubb : public foo<i>
{
        a_t x;
};

gives

bellatrix:~/src/tests$ g++ -c implicit.cpp
implicit.cpp:10: warning: `typename blubb<i>::a_t' is implicitly a
typename
implicit.cpp:10: warning: implicit typename is deprecated, please see the
   documentation for details

How am I supposed to fix such cases? Is this a gcc deficiency? Intel C++
doesnt complain here.

Thanx, Richard.

--
Richard Guenther <richard.guenther@uni-tuebingen.de>
WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/
The GLAME Project: http://www.glame.de/

Re: implicit typename warnings with gcc 3.1

[Gabriel Dos Reis]

Richard Guenther <rguenth@tat.physik.uni-tuebingen.de> writes:

| Hi!
| 
| template <int i>
| struct foo {
|         typedef int a_t;
| };
| 
| template <int i>
| class blubb : public foo<i>
| {
|         a_t x;
| };
| 
| gives
| 
| bellatrix:~/src/tests$ g++ -c implicit.cpp
| implicit.cpp:10: warning: `typename blubb<i>::a_t' is implicitly a
| typename
| implicit.cpp:10: warning: implicit typename is deprecated, please see the
|    documentation for details
| 
| How am I supposed to fix such cases?

This must be a FAQ.  Dependent base names are not examined during the
first phase of name lookup in template codes.  That means the above
code is ill-formed, since there is no "a_t" in scope. The incriminated
line should read something like: 

   typename foo<i>::a_t x;

| Is this a gcc deficiency? 

No. The only deficiency I see is that it doesn't issue a hard error ;-)

| Intel C++
| doesnt complain here.

Then, it isn't compliant in that regard.

-- Gaby

Re: implicit typename warnings with gcc 3.1

[Richard Guenther]

On 1 Aug 2002, Gabriel Dos Reis wrote:

> Richard Guenther <rguenth@tat.physik.uni-tuebingen.de> writes:
>
> | Hi!
> |
> | template <int i>
> | struct foo {
> |         typedef int a_t;
> | };
> |
> | template <int i>
> | class blubb : public foo<i>
> | {
> |         a_t x;
> | };
> |
> | gives
> |
> | bellatrix:~/src/tests$ g++ -c implicit.cpp
> | implicit.cpp:10: warning: `typename blubb<i>::a_t' is implicitly a
> | typename
> | implicit.cpp:10: warning: implicit typename is deprecated, please see the
> |    documentation for details
> |
> | How am I supposed to fix such cases?
>
> This must be a FAQ.  Dependent base names are not examined during the
> first phase of name lookup in template codes.  That means the above
> code is ill-formed, since there is no "a_t" in scope. The incriminated
> line should read something like:
>
>    typename foo<i>::a_t x;

--
Richard Guenther <richard.guenther@uni-tuebingen.de>
WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/
The GLAME Project: http://www.glame.de/

Re: implicit typename warnings with gcc 3.1

[Nix]

On Thu, 1 Aug 2002, Gabriel Dos Reis said:
> Richard Guenther <rguenth@tat.physik.uni-tuebingen.de> writes:
>| Is this a gcc deficiency? 
> 
> No. The only deficiency I see is that it doesn't issue a hard error ;-)

I fear that if it issued a hard error about half the C++ code in a
random Linux distro would cease to compile[1] :(

In time, when most people have fixed their code, it can be made an
error; but I'd not do it before that myself.

[1] merely an educated guesstimate

-- 
`There's something satisfying about killing JWZ over and over again.'
                                        -- 1i, personal communication

Re: implicit typename warnings with gcc 3.1

[Gabriel Dos Reis]

Nix <nix@esperi.demon.co.uk> writes:

| On Thu, 1 Aug 2002, Gabriel Dos Reis said:
| > Richard Guenther <rguenth@tat.physik.uni-tuebingen.de> writes:
| >| Is this a gcc deficiency? 
| > 
| > No. The only deficiency I see is that it doesn't issue a hard error ;-)
| 
| I fear that if it issued a hard error about half the C++ code in a
| random Linux distro would cease to compile[1] :(

I know that.  Please note the smiley.

-- Gaby