internal (c++) compiler error for unbound friend class template in a namespace

[Martin v. Loewis]

> Source example:
>   namespace nsn {
> 
>     template< class Type >
>     class Test {
>       template< class T2 >
>       friend class Test ;
>     };
> 
>   } // namespace 

Thanks for your bug report.The mainline compiler (2.96 20000511
(experimental)) compiles this without problems, so the bug appears to
be fixed.


http://www.gnu.org/software/gcc/bugs.html

If you question this line of reasoning, please discuss it in one of
the public C++ fora first, eg. comp.lang.c++.moderated, or
comp.std.c++.

http://gcc.gnu.org/cgi-bin/fom.cgi?file=12

Regards,
Martin

> 
> Incidently,
> 
> class Test {
>   template< class T2 >
>   friend class Test ;
> };
> 
> seems to generate an internal compiler error, regardless of whether it
> is declared in a namespace or not. I know it's bogus-code, but it still
> shouldn't cause an internal compiler error, right :-)

I've put that into GNATS.

> Reporting bugs to gcc is hard. If you go to the web site it sends you
> off to the info pages for your local installation. 

Are you talking about

http://www.gnu.org/software/gcc/bugs.html

Where exactly does it send you to the info pages? It offers you a
choice of reading the information either online or offline.

> For a brand new redhat 6.2 the info pages says to report to
> egcs-bugs@egcs.cygnus.com.

That's probably because redhat includes the egcs 1.1 manual. Had they
included the gcc 2.95 manual, the bug reporting address would have
been different. This is out of our control, please complain to RedHat
- we can't change the info files after they got on the RedHat CD-ROM
:-)

> This mailing list does not appear to exist anymore. How clever is
> it, that the online resource sends you off to a possible outdated
> e-mail address? :-)

Had you followed the online docs link, to
http://gcc.gnu.org/onlinedocs/gcc_8.html#SEC138 , you would not be
directed to an outdated e-mail address.

The EGCS project had a number of positive and a few negative
effects. Having dangling references to the now-terminated EGCS project
is one of the negative effects.

As for ease of reporting: Did you try the GNATS link on that page?

Regards,
Martin