search

oizma / angleproject

Automatically exported from code.google.com/p/angleproject
Other
0 stars 0 forks source link

The bind attrib location WebGL conformance test fails in ANGLE #83

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1. Run Chromium with ANGLE
2. Go to 
https://cvs.khronos.org/svn/repos/registry/trunk/public/webgl/sdk/tests/conforma
nce/gl-bind-attrib-location-test.html

What is the expected output?

The test should pass

What do you see instead?

The test fails

Run chromium with --use-gl=desktop to see it pass using OpenGL

Original issue reported on code.google.com by g...@chromium.org on 4 Dec 2010 at 2:06

GoogleCodeExporter commented 9 years ago 
This is caused by getActiveAttribute not correctly skipping over inactive 
attributes. Should be straightforward to fix...

Original comment by nicolas....@gmail.com on 14 Dec 2010 at 10:39

GoogleCodeExporter commented 9 years ago 
Fixing getActiveAttribute() doesn't make the entire test pass.

The script contains "gl.disableVertexAttribArray(0)" but this call is never 
observed by ANGLE. So this appears to be an Chrome bug. I suspect that it 
attempts to cache the attribute locations, but doesn't take the reassignments 
in account that can happen at linkProgram().

Original comment by nicolas....@gmail.com on 14 Dec 2010 at 1:33

Attachments:

GoogleCodeExporter commented 9 years ago 
Fixed the active attribute querying in r509.

That leaves the gl.disableVertexAttribArray(0) issue. Could someone working on 
Chromium have a look at that?

Original comment by nicolas....@gmail.com on 15 Dec 2010 at 3:55

GoogleCodeExporter commented 9 years ago 
Chromium never disables attrib 0 ever. this is because desktop GL doesn't
function if attrib 0 is disabled. Instead it shadows the disabled flag and
if the user asked for it to be disabled Chromium leaves it enabled and then
fills a buffer with the constant value that would be used if it had been
disabled.

That shouldn't effect whether  the app passes or not

Original comment by g...@chromium.org on 15 Dec 2010 at 5:11

GoogleCodeExporter commented 9 years ago 
This was fixed by r560.

Original comment by nicolas....@gmail.com on 11 Feb 2011 at 1:58