[Piglit] [PATCH 0/3] Fix some program input / output issues with ARB_piq

[Ian Romanick]

On 05/26/2016 08:01 AM, Martin Peres wrote:
> On 26/05/16 01:36, Ian Romanick wrote:
>> After spending about a week with dEQP, GL_OES_shader_io_blocks, and
>> GL_OES_geometry_shader, I believe there are some errors with the
>> expectations of our existing ARB_piq tests.  I have included spec
>> justifications for the changes, and the changes match tests that I have
>> seen in dEQP.
>>
>> I would really like to know the before / after results of the updated
>> test on modern NVIDIA and AMD closed-source drivers.
> 
> Thanks for doing this! Here are the result on GeForce GT 430/PCIe/SSE2,
> NVIDIA 355.11
> 
> $ ./piglit summary console -d before/ after
> spec/arb_program_interface_query/arb_program_interface_query-resource-query/GL_PROGRAM_INPUT(gs):
> pass fail
> spec/arb_program_interface_query/arb_program_interface_query-resource-query/GL_PROGRAM_INPUT(gs,fs):
> pass fail
> spec/arb_program_interface_query/arb_program_interface_query-resource-query/GL_PROGRAM_INPUT(tcs):
> pass fail
> spec/arb_program_interface_query/arb_program_interface_query-resource-query/GL_PROGRAM_INPUT(tcs,tes):
> pass fail
> spec/arb_program_interface_query/arb_program_interface_query-resource-query/GL_PROGRAM_INPUT(tes):
> pass fail

Interesting... is it just that they don't have gl_PerVertex on things,
or is the arrayness also different?  I want to put proper annotations in
the commit messages.

> Martin