[Piglit] glSecondaryColorPointer with size 4

[Brian Paul]

On 07/18/2010 11:32 PM, Vinson Lee wrote:
> Is 4 a valid size parameter for glSecondaryColorPointer?
>
> The piglit tests bgra-sec-color-pointer and glsl-routing call glSecondaryColorPointer with a size parameter of 4. Both these tests also crash on Mac OS X when glDrawArrays is called after glSecondaryColorPointer.
>
> The following is from the OpenGL documentation.
>
> glSecondaryColorPointer(size, type, stride, pointer)
> size - Specifies the number of components per color. Must be 3.
> GL_INVALID_VALUE is generated if size is not 3.
>
>
> EXT_vertex_array_bgra specification
>      5.  So should 4 be allowed as a valid size parameter for
>          glSecondaryColorPointer?
>
>          RESOLVED:  Not for this extension.  We leave it to another
>          extension or core revision could extend the secondary color to
>          4 components.

AFAICT, only 3 should be legal, but NVIDIA's driver allows 4 (as does 
Mesa since 2000).

Here's a couple patches for the piglit tests to use 3 instead of 4. 
Could you see if these help with OS X?

I also made a minor fix to bgra-sec-color-pointer.c so that a visual 
with an alpha channel is requested since the blend mode depends on 
destination alpha.

-Brian
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: bgra-sec-color-pointer.c-patch
URL: <http://lists.freedesktop.org/archives/piglit/attachments/20100719/c364890a/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: glsl-routing.c-patch
URL: <http://lists.freedesktop.org/archives/piglit/attachments/20100719/c364890a/attachment.asc>