LinuxQuestions.org - additional clipping planes in OpenGL (xorg 7.1.1 + radeon9700)

- Programming (https://www.linuxquestions.org/questions/programming-9/)

- - additional clipping planes in OpenGL (xorg 7.1.1 + radeon9700) (https://www.linuxquestions.org/questions/programming-9/additional-clipping-planes-in-opengl-xorg-7-1-1-radeon9700-508354/)

additional clipping planes in OpenGL (xorg 7.1.1 + radeon9700)

Hello everyone,

using OpenGL (and C in Linux-Debian) i want to add additional clipping planes to a scene by glClipPlane(GL_CLIP_PLANE0, eqn) and glEnable(GL_CLIP_PLANE0).
The first thing i noticed: nothing is clipped/changed.

Because of this, i tried the example code from the corresponding chapter in the opengl-redbook. I added two lines of info-output to the clippingTest.

This test renders a wireframe sphere and clips three fourth of it away - you should see only the fourth part of the full sphere. I compiled and ran this in windows, there it works as expected (screenshot in windows, screenshot in linux).

Can anyone verify this behaviour ? Now, i'm wondering if my setup is buggy or xorg is buggy or i am buggy.

The info-output in windows (mingw) is

Code:

$ gcc clippingTest.c -o clippingTest  -lglut32 -lopengl32 -lglu32 -lm && ./clippingTest.exe



Maximum count of additional clip-planes: 6

Depth bits: 16

the output in linux is

Code:

$ gcc -lGLU -lGL -lglut -lm clippingTest.c -o clippingTest && ./clippingTest



libGL warning: 3D driver claims to not support visual 0x4b

Maximum count of additional clip-planes: 6

Depth bits: 24

I suspect the above error about 'visual 0x4b' has something to do with it. But i didn't find anything useful regarding clipping in the web.

Can someone tell me, what 'visual 0x4b' is ? (and all the other visuals listed in my xorg.log)

I'm using Debian (unstable) with kernel 2.6.17.6, xorg 7.1.1 and an ati radeon9700 card. The hardware acceleration seems to work (opensource/xorg driver). glxinfo

Little background: The last assignment in my actual computer graphics course was about simulating the perspective transformed canonical viewing volume (cube with extrema [-1,-1,-1] and [1,1,1]) and its clipping on six sides (which is usually done by opengl on the frustum automagically).

Can anyone give me an hint or can tell me someone who can ?

thanks for reading,
cdex.

I'm running Debian on ppc (iBook) with ATI FireGl Te2 graphics card, and I've noticed the same problem. It only appears when DRI is on, though. With 'Option "NoAccel" "true"' in xorg.conf the clipping planes work fine. I think it might be a bug, maybe in the Mesa libraries.

There is indeed a bug report similar to this filed against Mesa, and it is claimed to have been fixed in 6.5.2, which hasn't made it into Debian yet.

ETA: No, actually it's supposed to have been fixed already in 6.5.1, so maybe this is a different issue.

Hello riot_nrrrd,
thanks for replying - thought nobody will ever answer :)

So, we have to wait ... yesterday, i updated my (unstable) packages and got 'only' mesa 6.5.1-0.5

Currently, I think about creating a (debian) bug report ... its supposed to be the 'libgl1-mesa-glx' package (?)

Additionally, a few days ago I found out that glPointSize() has no effect, but this is probably a (radeon-) driver issue
(because of bug 702)