public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* Failure of 2371.C
@ 1998-02-15  0:16 Martin von Loewis
  1998-02-15 11:52 ` Manfred Hollstein
  0 siblings, 1 reply; 2+ messages in thread
From: Martin von Loewis @ 1998-02-15  0:16 UTC (permalink / raw)
  To: egcs

Is anybody else seeing that compilation of
gcc/testsuite/g++.old-deja/g++.jason/2371.C fails? Or is it my local
modifications that broke it? The symptom is that the compilation won't
terminate.

I've reduced that to a minimal example, which reads

template<class T>
class ListS {
public:
    class Vix {
    public:
	friend int operator!=(const Vix& x1, const Vix& x2)
	    { return 0; }
    };
};
class test {
};
template class ListS<test>;

The problem is that pushdecl () is called twice for
__ne__H1Z4test_RCQ2t5ListS1ZX013VixT0_i while in the global binding,
which will produce a cyclic list for the names in that binding. One
copy is produced when iterating over the friends of Vix, and the other
when iterating over the friends of ListS.

Anybody else seeing this?

Martin


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Failure of 2371.C
  1998-02-15  0:16 Failure of 2371.C Martin von Loewis
@ 1998-02-15 11:52 ` Manfred Hollstein
  0 siblings, 0 replies; 2+ messages in thread
From: Manfred Hollstein @ 1998-02-15 11:52 UTC (permalink / raw)
  To: martin; +Cc: egcs

On Sun, 15 February 1998, 09:15:42, martin@mira.isdn.cs.tu-berlin.de wrote:

 > Is anybody else seeing that compilation of
 > gcc/testsuite/g++.old-deja/g++.jason/2371.C fails? Or is it my local
 > modifications that broke it? The symptom is that the compilation won't
 > terminate.
 > 
 > I've reduced that to a minimal example, which reads
 > 
 > template<class T>
 > class ListS {
 > public:
 >     class Vix {
 >     public:
 > 	friend int operator!=(const Vix& x1, const Vix& x2)
 > 	    { return 0; }
 >     };
 > };
 > class test {
 > };
 > template class ListS<test>;
 > 
 > The problem is that pushdecl () is called twice for
 > __ne__H1Z4test_RCQ2t5ListS1ZX013VixT0_i while in the global binding,
 > which will produce a cyclic list for the names in that binding. One
 > copy is produced when iterating over the friends of Vix, and the other
 > when iterating over the friends of ListS.
 > 
 > Anybody else seeing this?

No problems here on i586-linux-gnulibc1 with egcs from CVS last updated on
1998/02/15 10:20:28 MET.

Manfred

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~1998-02-15 11:52 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1998-02-15  0:16 Failure of 2371.C Martin von Loewis
1998-02-15 11:52 ` Manfred Hollstein

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).