glean and Piglit -- OpenGL driver testing

[Saulius Krasuckas]

* On Thu, 17 Sep 2009, Henri Verbeet wrote:
> * 2009/9/17 Saulius Krasuckas <saulius2 at ar.fi.lt>:
> >
> > Could these be of any use for our graphic guys -- Stefan and co.?
> 
> Well, they're mostly useful when you're maintaining an OpenGL driver. 
> Mesa already uses these.

And what about seeing if our tests (vs Win drivers) aren't really broken?

For example one test-check fails with these OpenGL 1.[34].x and 2.[12].x 
drivers for the adapters:

2.1.8870  ATI Radeon HD 4200
2.0.0     Intel 965/963 Graphics Media Accelerator
1.4.1     GeForce4 MX 440/AGP/SSE 
1.4.0     Intel 915GM
1.3.0     Intel Brookdale-G
1.3.4145  MOBILITY RADEON 7500 DDR x86/SSE2

opengl.c:328: Test failed: Sharing of display lists failed for a context which already shared lists before

But it doesn't fail on these:

3.1.0     GeForce 8600 GTS/PCI/SSE2
3.0.0     GeForce 9600M GT/PCI/SSE2
2.1.2     GeForce FX 5200/AGP/SSE2
2.1.2     GeForce 7300 LE/PCI/SSE2/3DNOW!
2.1.1     GeForce 8600M GS/PCI/SSE2
2.1.1     GeForce 8400M GS/PCI/SSE2
2.0 Chrom Chromium 1.9
1.5 Chrom Chromium 1.9
1.1.0     GDI Generic (old w9x ar virtual boxes)

Aren't you guys having hard time deciding whether this statement:

 322     /* Test 3: Share display lists with a context which already shares display lists with another context.
 323      * According to MSDN the second parameter cannot share any display lists but some buggy drivers might allow it */

is OK ?  (No offence)
I thought driver test suite would give a more thorough answer..

> > Then there is PerceptualDiff utility I found some time ago [3]. 
> >  Guessed, could it also usefull for finding visual regressions of 
> > Wine?  Probably not, as it seems to be used for testing video codecs 
> > (but I may be wrong):
> 
> Possibly, but it would have to be in the context of a larger framework 
> like e.g. CxTest or Appinstall.

And what about D3D rendering discrepancies?