gEDA-user: Intel stencil clear bug with multiple contexts

[Peter Clifton <pcjc2@xxxxxxxxx>]

Hi guys,

This is a vague bug report at best - hence mentioned here in case it
means anything to anyone. I've not really got enough to file anything
useful in Bugzilla yet!

I've been working on GL code for a PCB design app. It uses various bits
in the stencil buffer for sub-compositing. To avoid expensive clear
operations, it uses each of the bits on the stencil buffer independently
(in turn), only clearing the buffer when it runs out of pre-cleared
bits.

I'm using two drawing windows (and two GL contexts), set up with
GtkGlExt.

With the latest mesa code (sorry - no idea when it regressed), calling
glClear (GL_STENCIL_BUFFER_BIT); on my second drawing context seems to
corrupt the main drawing context, preventing its colour buffer being
properly cleared when it is next used.

Since I'd expect the two contexts to be separate, and not affect each
other - I can only assume that the driver is somehow muddling them up
with regards to the stencil buffer.

Removing the glClear (GL_STENCIL_BUFFER_BIT); on my second context, both
windows render "properly". (The second context doesn't _really_ need the
sub-compositing, so I can get away with that for now).

I'm using mesa from the Ubuntu Xorg-edgers PPA:
7.7.0~git20090928.d492e7b0-0ubuntu0

Along with the intel driver:
2:2.8.99.902~git20090923.a92bbcc9-0ubuntu0tormod

Forcing software rendering fixes the issue - and I know this same code
has worked properly with older mesa versions (with HW acceleration).

Any ideas?

Unfortunately I've not got a lot of time at the moment to distil a
succinct test-case.


Regards,

-- 
Peter Clifton

Electrical Engineering Division,
Engineering Department,
University of Cambridge,
9, JJ Thomson Avenue,
Cambridge
CB3 0FA

Tel: +44 (0)7729 980173 - (No signal in the lab!)



_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user