Re: gEDA-user: gschem segmentation fault

[Steve Brown <sbrown@xxxxxxxx>]

Dimitri Princen wrote:

> Well, so far it never happened when I tried to open a schematic, but
> I'd like to try to set G_SLICE=always-malloc, but where can I set it?
> Is it a feature of valgrind? (I've never used it, but then again, I'm
> quite new to linux, I am starting to feel at home though!)
>
> Regards, Dimitri
>
> On 4/7/06, Steve Brown <sbrown@xxxxxxxx> wrote:
>  
>
> > Dimitri Princen wrote:
> >    
> >
> > > I recently submitted a bug regarding a segmentation fault in gschem
> > > (see http://sourceforge.net/tracker/index.php?func=detail&aid=1455979&group_id=161080&atid=818426)
> > > but I'd like to know if there is a workaround available.
> > >
> > > I've succesfully compiled the cvs version, but the problem still
> > > occures. Furthermore, is does not only happens when I delete multiple
> > > components but also when I delete the components individualy (or
> > > single nets for that matter).
> > >
> > > By the way, didn't anybody else experienced this problem? I'm sure I'm
> > > not the only one deleting components :)
> > >
> > > (If anybody is interested, I can post a backtrace)
> > >
> > > regards, Dimitri
> > >
> > >      
> > I've got the same problem. One of my schematics when just opened and
> > closed will reliably segv. Valgrind gives only minimal complaints.
> >
> > If I set  G_SLICE=always-malloc , to turn off the glib slice allocator,
> > no segv. Maybe the problem isn't entirely with gschem.
> >
> > Steve
> >
> >
> >
> >    
It's an environment variable
Try something like.
   G_SLICE=always-malloc gschem file.sch
and see if it makes any difference.
Steve