[Piglit] [PATCH] arb_program_interface_query: update and fix arrays of arrays tests

Tapani Pälli tapani.palli at intel.com
Tue Jun 30 03:24:15 PDT 2015

On 06/30/2015 12:48 PM, Timothy Arceri wrote:
> On Mon, 2015-06-29 at 13:49 +0300, Tapani Pälli wrote:
>> On 06/27/2015 02:14 AM, Timothy Arceri wrote:
>>> Add a 3D array test and fix the expected result for an array
>>> without
>>> a subscript attached to the name.
>>> The arb_program_interface_query spec says:
>>> "If <name> exactly matches the name string of one of the active
>>> resources
>>> for <programInterface>, the index of the matched resource is
>>> returned.
>>> Additionally, if <name> would exactly match the name string of an
>>> active
>>> resource if "[0]" were appended to <name>, the index of the matched
>>> resource is returned."
>>> So the index should not be INVALID_INDEX when there is no
>>> subscript.
>> Did you check if these changes pass on nvidia/amd?
>> Shouldn't one give 'vs_input2[0]' as parameter? Then it would
>> definitely
>> qualify because 'vs_input2[0][0]' is ok but having only 'vs_input2'
>> means name check against 'vs_input2' and 'vs_input2[0]' which would
>> not
>> qualify (?)
>> This was my understanding when writing the code. So how I interpret
>> the
>> spec is that if you have multiple levels, you'll need to specify
>> until
>> the last level and optionally [0] for that last one.
> Yes you are right its all defined in Issues 7 & 8 at the bottom of the
> spec. I'll send a V2 that covers more AoA combinations, I'm working on
> some fixes for ARB_program_interface_query in Mesa as part of my AoA
> work I hope to have a new series out at the end of the week.

Great to hear that! I've noticed several failures within OpenGL ES 3.1 
conformance tests for ARB_program_interface_query and I think some of 
those were AoA subtests.

// Tapani

More information about the Piglit mailing list