Re: octave2.1-headers: mkoctfile generated functions crash randomly

[Jon H. Davis]

On Tue, 2005-11-22 at 21:37 +0100, Rafael Laboissiere wrote:
> Below is a bug report filed against the octave2.1-headers Debian package
> (http://bugs.debian.org/340320).  See my comments after the included
> message.
> 
> * Teemu Ikonen <address@hidden> [2005-11-22 19:34]:
> 
> > Package: octave2.1-headers
> > Version: 2.1.72-4
> > Severity: normal
> > 
> > Oct-functions generated by mkoctfile segfault randomly, but very often. For
> > example, the following innocent-looking function:
> > 
> > -*-*-*-
> > 
> > #include <octave/oct.h>
> > 
> > DEFUN_DLD(uitest, args, , " ")
> > {
> >       int x = 20;
> >       int y = 20;
> >         
> >       return octave_value(Matrix(y,x));
> >               
> > }
> >               
> > -*-*-*-
> > 
> > produces this when ran repeatedly:
> > 
> > octave:1> uitest();
> > octave:2> uitest();
> > octave:3> uitest();
> > octave:4> uitest();
> > panic: Segmentation fault -- stopping myself...
> > attempting to save variables to octave-core'...
> > save to octave-core' complete
> > Segmentation fault
> > 
> > I can usually make Octave crash by running uitest less than 10 times.
> > 
> > Since I run a mixed system of testing and unstable, I'm rather certain that
> > the crashes are caused by some kind of library version mismatch which the
> > dependencies do not take into account. Octave 2.1.69 from sarge does not
> > crash with the test above.
> > 
> > I would appreciate any ideas on how to get oct-files working again.
> 
> I confirm the problem here in my sid chroot.  Your program crashes in a
> systematic way producing the following:
> 
> $ echo 'for i=1:100;printf("%i.",i);fflush(stdout);uitest();end' | octave -qf
> 1.2.3.4.5.6.7.8.9.panic: Segmentation fault -- stopping myself...
> attempting to save variables to `octave-core'...
> save to `octave-core' complete
> Segmentation fault
> 
> What puzzles me is that it works fine in my etch (testing) system.  It looks
> like a library problem.
> 
> How should I proceed in order to find where the bug comes from?
> 
        Isn't sid in the throes of c++ library changes? Until everything gets
changed over there might be a chance of a library mismatch creeping in.

        Jon



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------