[Piglit] [PATCH] DSA: fix error value for *TextureSubImage* when target doesn't match

[Martin Peres]

On 22/04/15 20:06, Ilia Mirkin wrote:
> [fixing Laura's email]
>
> On Wed, Apr 22, 2015 at 1:05 PM, Ilia Mirkin <imirkin at alum.mit.edu> wrote:
>> On Wed, Apr 22, 2015 at 12:53 PM, Arthur Huillet <arthur.huillet at free.fr> wrote:
>>> Hi,
>>>
>>> On 2015-04-22 18:37, Ilia Mirkin wrote:
>>>>  From the ARB_dsa spec:
>>>>
>>>>      An INVALID_ENUM error is generated by *TexSubImage* if <target> does
>>>>      not match the command, as shown in table 8.subtarg.
>>>>
>>>>      An INVALID_OPERATION error is generated by *TextureSubImage* if the
>>>>      effective target of <texture> does not match the command, as shown in
>>>>      table 8.subtarg.
>>>>
>>>> Probably the source of the confusion?
>>>
>>> Yes, I believe this was the source of the confusion. At least it confused
>>> *me* for a few minutes :) That's why I'm suggesting the policy to always
>>> quote spec, so that reviewers and users can more easily spot these problems.
>> In a universe where you're getting too many piglit contributions and
>> have you do quality control *somehow*, sure. But in this one, getting
>> people to write tests is like pulling teeth (and I'm guilty of this as
>> well), so adding even more onerous requirements on test cases seems
>> like it might be a losing proposition.
>>
>>>> AFAIK a bunch of spec bugs were
>>>> filed about these little inconsistencies... (But I don't know exactly
>>>> what they were, no Khronos access for me.)
>>>
>>> I haven't seen one. I haven't looked very hard.
>>>
>>> Does the patch look OK to you?
>> Yeah, patch looks perfectly fine. But I'd like to hear from Laura (or
>> Martin perhaps, who reviewed a lot of these) about whether this was
>> done on purpose or not, pending a spec fix.

I am not aware of any bug report filed by Laura on this. Given that she 
wrote this code in October last year, I am sure she would have received 
an answer from Khronos if she had filed a bug. This would suggest it is 
a bug in the test.

Let's wait for her comment to make sure of it.

Anyway, thanks for the patch Arthur! I know it can be frustrating to 
have to check every failing test while wondering whether this is a test 
or a driver bug. We should encourage developers writing piglit tests to 
try them on multiple drivers that already support the extension they are 
testing. This is however impossible to mandate.

Do you try your tests against Intel's or AMD's driver for Windows? If 
so, why not start working on them in the open and push them to piglit 
since it now supports Windows?