On Tue, 29 Apr 2003, Dave Bentham wrote:
> Hi,
>
> I've had chance to do some debugging myself today and perhaps got to the
> bottom of it.
>
> The libsadt_objects.so library in my first gdb script is a red herring.
> Its in fact the previously loaded lib: the Python plugin.
Ah, yes, we've had problems with that before.
> It turns out I had a real old one lying around in /usr/lib/dia, and even
> though I haven't built dia recently with python support it seemed to be
> trying that old plugin: dia did put the error dialog up, but for ages now
> I just'clicked' it away without investigating... :-(
Maybe we should turn off plugins that don't work -- but then of course
people will be confused when they update and don't see the plugin. Even if
we have a big fat warning to turn it back on when fixed, people will
forget. But I guess that's better than these strange crashes.
> After removing the offending files, dia is fine. Just to try it, I
> rebuilt dia with python support and again its okay. (Hopefully I'll
> remember its there in two years building dia 0.768...)
>
> What I don't know is why it SIGSEGVs under xdm sessions and not
> gdm/Gnome... One of the mysteries of software.
Rather mysterious, yes. The plug-in loader was somewhat hardened for 0.91,
but apparently not enough.
> Anyway, thanks for the previous helps, and I can now get back to my
> UMLing.
Happy UMLing! (That does sound funny when you say it out loud:)
-Lars
--
Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| HĂ„rdgrim of Numenor
"I do not agree with a word that you say, but I |----------------------------
will defend to the death your right to say it." | Where are we going, and
--Evelyn Beatrice Hall paraphrasing Voltaire | what's with the handbasket?