Re: gcc-3.3 CVS failing during libjava compile with:./.libs/libgcj.so: undefined reference to `__gcc_personality_v0'

[Martin Schlemmer <azarah at gentoo dot org>]

On Sun, 2003-07-13 at 21:19, Jeff Sturm wrote:
> On 13 Jul 2003, Martin Schlemmer wrote:
> > > Index: libgcc-std.ver
> > > ===================================================================
> > > RCS file: /cvs/gcc/gcc/gcc/libgcc-std.ver,v
> > > retrieving revision 1.21
> > > diff -c -p -r1.21 libgcc-std.ver
> > > *** libgcc-std.ver      7 May 2003 22:11:33 -0000       1.21
> > > --- libgcc-std.ver      1 Jul 2003 03:06:40 -0000
> > > *************** GCC_3.3 {
> > > *** 183,188 ****
> > > --- 183,190 ----
> > >     _Unwind_Backtrace
> > >     _Unwind_Resume_or_Rethrow
> > >     _Unwind_SjLj_Resume_or_Rethrow
> > > +   __gcc_personality_v0
> > > +   __gcc_personality_sj0
> > >   }
> > >
> > >   %inherit GCC_3.4 GCC_3.3
> >
> > This does work ... only question now, is will it cause issues on a
> > system compiled with this gcc, and another gcc (3.4/whatever) is
> > installed that do not export these functions ?
> 
> If a similar patch is approved for 3.4, libgcc_s will export these
> functions, and you'll have no trouble on systems built with 3.4.  3.3.1
> and earlier will be a problem however.
> 

Right, I thought so (*think back at the __divdi3 issue*).

> > Also, this only occurs with current glibc CVS, with NPTL (0.48-0.52).
> > I have not been able to test with linuxthreads, but it works fine
> > with a glibc snapshot of 2003-05-27/9 that also use NPTL (0.38).
> 
> Regardless, the failure can easily be demonstrated with a small testcase.
> Normally I'd suggest a PR be entered with high priority as a 3.3
> regression, however since it is new functionality there really is no
> regression at all.
> 

Ok, so what is the solution ... do we push your patch ?


Thanks,

-- 

Martin Schlemmer

Attachment: signature.asc
Description: This is a digitally signed message part