[Piglit] [PATCH 2/5] gles2: new glReadPixels unit test

[Matt Turner]

On Thu, Dec 20, 2012 at 5:06 PM, Tom Gall <tom.gall at linaro.org> wrote:
> Hi Eric,
>
> On Thu, Dec 20, 2012 at 6:30 PM, Eric Anholt <eric at anholt.net> wrote:
>> Matt Turner <mattst88 at gmail.com> writes:
>>
>>> On Thu, Dec 20, 2012 at 10:54 AM, Tom Gall <tom.gall at linaro.org> wrote:
>>>> A few variations are commented out due to results that are
>>>> incorrect and need to be followed up with Mesa, such as modifying
>>>> the buffer even tho the format is invalid. There are cases where
>>>> the returned data is incorrectly formated. These can best be
>>>> addressed in time and I thought it best to include the variations
>>>> but in their commented out form.
>>>
>>> I think Eric already told you that it's okay to commit failing tests.
>>
>> Not just okay, it's expected.
>>
>> Other note: I don't think the name "unit" belongs in this.  Nothing in
>> piglit is a unit test according to what I've understood or how unit
>> tests are described on wikipedia.
>
> My intent is it's first in the series of being something "unit" test
> like across the gles2 api. It's a start and at this point I've been
> mostly heads down on the shaders and thinking about piglit + Android.

I agree with Eric. A unit test is something that is in the same
codebase as the code it's testing. A unit test would, for instance,
confirm that internal Mesa functions behave as expected. I don't think
this is a unit test.