[Piglit] [PATCH] arb_framebuffer_no_attachments: add params set&get test

Tapani tapani.palli at intel.com
Mon Apr 20 20:53:09 PDT 2015


On 04/20/2015 10:14 PM, Ilia Mirkin wrote:
> On Mon, Apr 20, 2015 at 9:22 AM, Tapani <tapani.palli at intel.com> wrote:
>> On 04/20/2015 04:13 PM, Ilia Mirkin wrote:
>>> On Mon, Apr 20, 2015 at 9:04 AM, Tapani <tapani.palli at intel.com> wrote:
>>>> On 04/20/2015 03:58 PM, Ilia Mirkin wrote:
>>>>> On Mon, Apr 20, 2015 at 4:16 AM, Tapani Pälli <tapani.palli at intel.com>
>>>>> wrote:
>>>>>> All other tests except invalid_enum_check pass on Nvidia
>>>>>> binary driver (version 346.35).
>>>>>>
>>>>>> Signed-off-by: Tapani Pälli <tapani.palli at intel.com>
>>>>>> ---
>>>>>>     tests/all.py                                       |   1 +
>>>>>>     .../CMakeLists.gl.txt                              |   1 +
>>>>>>     tests/spec/arb_framebuffer_no_attachments/params.c | 297
>>>>>> +++++++++++++++++++++
>>>>>>     3 files changed, 299 insertions(+)
>>>>>>     create mode 100755
>>>>>> tests/spec/arb_framebuffer_no_attachments/params.c
>>>>>>
>>>>>> diff --git a/tests/all.py b/tests/all.py
>>>>>> index 18124b7..018b37d 100755
>>>>>> --- a/tests/all.py
>>>>>> +++ b/tests/all.py
>>>>>> @@ -2295,6 +2295,7 @@ with profile.group_manager(
>>>>>>             PiglitGLTest,
>>>>>>             grouptools.join('spec', 'ARB_framebuffer_no_attachments'))
>>>>>> as
>>>>>> g:
>>>>>>         g(['arb_framebuffer_no_attachments-minmax'],
>>>>>> run_concurrent=False)
>>>>>> +    g(['arb_framebuffer_no_attachments-params'], run_concurrent=False)
>>>>> Why are both of these run_concurrent=False?
>>>>
>>>> Well ... just because of copy-paste. I'm not sure about the state of this
>>>> field. Will someone go through all the tests in some point and change
>>>> them
>>>> concurrent or how is this supposed to work?
>>>>
>>> Each test has to be evaluated individually, so ideally you'd get it
>>> right when you first add it. Either it does something funky that
>>> requires no other tests to be running (e.g. futzing with global state,
>>> timing, etc), or it doesn't. For the vast, vast, vast majority of
>>> tests, it doesn't.
>>
>> OK, then it should be safe to make these run concurrent. I can test and
>> change this when committing.
> Sounds good. Another thing that I've been going on like a broken
> record about is whether DSA will be core-only. If the plan is to make
> it core-only, then these tests should require core profile, otherwise
> they won't test the DSA bits in mesa.

Good point, this test should be made 3.0 Core then.

// Tapani



More information about the Piglit mailing list