Re: glibc 2.2.4pre2

[Jakub Jelinek <jakub at redhat dot com>]

On Wed, Aug 08, 2001 at 09:45:09AM +0200, Thorsten Kukuk wrote:
> On Wed, Aug 08, Frédéric L. W. Meunier wrote:
> 
> > Andreas Jaeger wrote:
> > 
> > >> http://sources.redhat.com/ml/bug-glibc/2001-07/msg00017.html
> > >> The second patch still isn't in gcc-2_95-branch 
> >                 
> > > You have to persuade Bernd Schmidt to apply the patch, GCC is
> > > beyond control of the glibc developers.
> > 
> > Bernd, Jakub: Could you both work together and release the
> > patch ? If 2.2.4 won't support GCC 3.0.1, then Red Hat's 2.96
> > will be the only supported compiler, what's a bad idea.
> 
> Why is Red Hat's 2.96 the only supported compiler ? gcc 2.95.3 from
> SuSE Linux 7.2 also works fine without problems. I think the same
> is true for Debian.

Does atexit work properly there? If yes, the better.
I mean, do you have .hidden __dso_handle symbol in crtbeginS.o?
If not, it will not work properly (and gcc 2.95.x did not have it).
Also, do you have the atexit/i386 patch in (it is in 2.95.4 CVS)?

The __dso_handle patch I posted for gcc 2.95.x was untested, that's why I
asked people who want to compile glibc with gcc 2.95.x to test it out.
Recently some folks have mailed about make check failure even with
__dso_handle patch in some atexit tests. Someone who is able to reproduce it
should debug it, I'm not able to reproduce it with the compilers I'm using
(2.96-RH and 3.0.1).

	Jakub