Bug 83520 – getShaderFormatPrecision should return sensible values

• Component: WebGL
• Status: RESOLVED FIXED

[Zhenyao Mo]

Originally the values are copied from chromium port, which are wrong.

It's already fixed in chromium side.

Now we need to also fix it for mac and qt port.

(Also, I reverted my changes to the khronos conformance test so it's checking again whether the returned values are sensible)

[Zhenyao Mo]

Created attachment 136341 [details]
Patch

[Zhenyao Mo]

Ken, please have a look.

The value returns here are documented in the GLES 2 spec, and they are also used by chromium port.

Also, the test is already updated in khronos (it was there in the first place, then I deleted them because we return negative values, now I reverted my change).

[Kenneth Russell]

Comment on attachment 136341 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=136341&action=review

> LayoutTests/fast/canvas/webgl/shader-precision-format.html:56
> +shouldBeTrue('shaderPrecisionFormat.rangeMax >=0 && shaderPrecisionFormat.rangeMax <= 128');

Are these two values wise? What would happen on a graphics card that supports double-precision floating-point values internally? I think that the conformance test should at least allow for that possibility.

[Zhenyao Mo]

(In reply to comment #3)
> (From update of attachment 136341 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=136341&action=review
> 
> > LayoutTests/fast/canvas/webgl/shader-precision-format.html:56
> > +shouldBeTrue('shaderPrecisionFormat.rangeMax >=0 && shaderPrecisionFormat.rangeMax <= 128');
> 
> Are these two values wise? What would happen on a graphics card that supports double-precision floating-point values internally? I think that the conformance test should at least allow for that possibility.

I don't know how they come up with the upper limit.  I checked the spec, I only see the minimum requirement for each value.  I'll update the conformance test and do another sync.

[Zhenyao Mo]

Created attachment 136355 [details]
Patch

[Zhenyao Mo]

Updated the khronos conformance test due to GLES shading language spec sec 4.5, then synced to LayoutTests.

Please have another look.

[Kenneth Russell]

Comment on attachment 136355 [details]
Patch

Thanks, the test looks a lot better. r=me

[WebKit Review Bot]

Comment on attachment 136355 [details]
Patch

Attachment 136355 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/12376173

New failing tests:
platform/chromium/virtual/gpu/fast/canvas/webgl/shader-precision-format.html
fast/canvas/webgl/shader-precision-format.html

[WebKit Review Bot]

Created attachment 136403 [details]
Archive of layout-test-results from ec2-cr-linux-01

The attached test failures were seen while running run-webkit-tests on the chromium-ews.
Bot: ec2-cr-linux-01  Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'>  Platform: Linux-2.6.35-28-virtual-x86_64-with-Ubuntu-10.10-maverick

[Zhenyao Mo]

Committed r113733: <http://trac.webkit.org/changeset/113733>