Re: [Chicken-users] segfault using cairo/SDL

[Markus Klotzbuecher]

On Thu, Dec 22, 2011 at 10:57:39AM +0100, Markus Klotzbuecher wrote:

> On Thu, Dec 22, 2011 at 09:41:05AM +0100, Christian Kellermann wrote:
> > * Markus Klotzbuecher <address@hidden> [111221 21:12]:
> > > Dear All,
> > > 
> > > I'm trying to use the cairo and sdl eggs for a small visualization
> > > tool. One thing I would like to do is to use the mouse to move around
> > > the current figure by clicking and holding it. However, I am getting
> > > frequent segfaults after a few seconds of this moving around. The
> > > crash happens both when running compiled and interpreted. Below an
> > > slightly modified version of the test-cairo.scm from the cairo egg
> > > that implements this "dragging" (I would have pasted it to
> > > http://paste.call-cc.org/ , but the service seems down)
> > > 
> > > Is anyone else using these eggs and has seen such problems? Does
> > > anyone have a suggestions how to track this down?
> > 
> > Yes, I am currently using it for a small doodle framework for easy
> > visualisations and games. I haven't looked (yet) into mouse support.
> > The SDL egg is rather dusty and there might be such dark corners.
> > 
> > I will file this as a bug and will have a look at it during the
> > holidays.
> 
> Great, much appreciated!
> 
> > If you find the time you could build a chicken with debugging
> > information (add DEBUGBUILD=1 to the make call) and show be the
> > backtrace of that. This might be of more help than the standard
> > backtrace.
> 
> Fine, i'll try to do that during the holidays!

The problem seems to be triggered by the use of text-extents. I
uploaded a newer testcase to the bug report that makes heavier use of
extents and also crashes almost immediately on my machine. I also
added a backtraces from a CHICKEN built with DEBUGBUILD=1.

Best regards
Markus