[Piglit] [PATCH] readpix-z: asst. test improvements

[Brian Paul]

On 10/18/2012 08:42 AM, Brian Paul wrote:
> On 10/17/2012 05:43 PM, Eric Anholt wrote:
>> Brian Paul<brianp at vmware.com> writes:
>>
>>> 1. Do a simple test that glClear(GL_DEPTH_BUFFER_BIT) works as
>>> expected.
>>> 2. Display the depth buffer contents as a grayscale image when not
>>> running in automatic mode to aid debugging.
>>
>> Since we've got lots of tests that clear the depth buffer and rely
>> on it
>> being cleared, this test doesn't make much sense to me -- if you're
>> trying to sanity check bit quality of just depth glClear on window
>> system framebuffers because fbo-clear-formats isn't covering your case,
>> it seems like you'd want a cross of fbo-clear-formats and
>> glx-visuals-depth.
>
> I'm interested in the specific case of glClear(GL_DEPTH_BUFFER_BIT)
> followed by glReadPixels() with nothing in between. It might not make
> sense to you but it's a case I'm specifically interested in for some
> specific wgl testing ATM.

The comment is probably what mislead you.  The point is to test 
glReadPixels(GL_DEPTH_COMPONENT) in these scenarios (per the test 
name).  I'm relying on glClear(GL_DEPTH_BUFFER_BIT) working properly.

So:
"1. Test glReadPixels(depth) immediatly after glClear(depth)."

-Brian