Good evening!
I hit a strange bug a while ago (in september) using a rpm (on a
Linux/RedHat 6.0 system) of binaries for Mesa 3.1 Beta. I bypassed it by
using a LOCALLY compiled version of Mesa 3.0. I then put the fault on
the "uncompatible binaries". I hit the very same bug today, in an
attempt to LOCALLY compile Mesa 3.1 in order to take advantage of a G200
accelerator using the GLX library refered to on the main page as
"glx.on.openprojects.net".
Let me tell you about the (unusual) way I use Mesa. I built a JPLib
library using the programming language Sather. Sather is a VERY clean OO
language, whose compiler produces C code. I built my rudimentary GUI
using glut. Sather uses the Boehm garbage collector.
The bug has the following behavior:
The program executes correctly, but eventually requires ALL of my
computer memory. On a PC under Linux/RedHat 6.0, I have to reset the PC:
everything freeses! On a Sun UltraSparc 1 under SunOS 5.5, the failure
is more graceful (signal 10, 11 or "out of memory" messages), but I
still cannot witness correct execution of the program. Notice that my
program runs without trouble using Mesa 3.0, Mesa 2.2.
If you need my program, I will be happy to send you my code, but
remember that it is Sather code. I could also send you the (huge) C-code
generated by the compiler...
Thanks,
JPD
_______________________________________________
Mesa-bug maillist - [EMAIL PROTECTED]
http://lists.mesa3d.org/mailman/listinfo/mesa-bug
_______________________________________________
Mesa-dev maillist - [EMAIL PROTECTED]
http://lists.mesa3d.org/mailman/listinfo/mesa-dev
