This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: weird segfaults @ g++2.8.1, weird mis-linking egcs 1.1
- To: oliva at dcc dot unicamp dot br
- Subject: Re: weird segfaults @ g++2.8.1, weird mis-linking egcs 1.1
- From: Yotam Medini <yotam at avanticorp dot com>
- Date: Mon, 12 Oct 1998 08:26:36 -0700
- CC: egcs-bugs at cygnus dot com, khan at xraylith dot wisc dot edu
- Reply-to: yotam_medini at avanticorp dot com
> Sender: oliva@dcc.unicamp.br
> Cc: egcs-bugs@cygnus.com, ...
> From: Alexandre Oliva <oliva@dcc.unicamp.br>
> Date: 12 Oct 1998 09:54:02 -200
> Believe me, it is a simplification of your testcase, that I'm
Nice, thanks!
> > .../egcs/Solaris/bin/g++ -g -Wl,-R.../egcs/Solaris/lib
> -o sp.new sp.o tpl.o 2>&1 | \
> > tee link.err | /home/seg/yotam/build/egcs/Solaris/bin/c++filt
> > sp.o: In function `WbReferenceCount::unique(void) const':
> > .../WbPointer.hh:16: undefined reference to
> `ostream::operator<<(ostream &(*)(ostream &))'
> > .../WbPointer.hh:16: undefined reference to
> `ostream::operator<<(ostream &(*)(ostream &))'
>
> I couldn't reproduce this problem; are you sure it doesn't have
> anything to do with a wrong library path?
As I subsequently posted, with egcs-1.1 I mistakenly compiled
my case without -fsqaungle as I did for building libstdc++.
Once I compiled it with -fsqaungle it linked well
and produced the same segmentation fault as with g++-2.8.1.
By the way, I think the _source_ reference given by the
'undefined reference' error could be improved.
> Alexandre Oliva
-- yotam