[Piglit] [PATCH] tests/spec/glsl-1.50/execution/fragcoord-layout-qualifiers-conflicting-case-7.shader_test conflict with spec

[Anuj Phogat]

On Wed, Mar 4, 2015 at 6:25 PM, Matt Turner <mattst88 at gmail.com> wrote:
> On Tue, Mar 3, 2015 at 6:43 PM, Guo, Johney <Weijun.Guo at amd.com> wrote:
>>
>> According to spec, this test should expected to error  instead of success.
I had a discussion with Ian when I added this test to piglit and khronos
conformance. We concluded that spec restriction shouldn't apply to a case
when you redeclare without any layout qualifiers because it matches the
default declaration of gl_FragCoord. NVIDIA proprietary linux driver also
allows the test to link. Making this case fail to link might break some
application making this harmless mistake.


>> The behavior is different with Khronos conformance test GL44-CTS.gtf32.GL3Tests.fragment_coord_conventions.fragment_coord_conventions_32_multishader_error_1   ( subcase redeclaration_with_no_layout_qualifiers_and_no_redeclaration_but_use).
Conformance test behavior matches the piglit test behavior.

>>
>> *     "If gl_FragCoord is redeclared in any fragment shader in a program,
>>  *      it must be redeclared in all the fragment shaders in that program
>>  *      that have a static use gl_FragCoord. All redeclarations of
>>  *      gl_FragCoord in all fragment shaders in a single program must have
>>  *      the same set of qualifiers."
>>
>> Please check it.
>> ---
>>
>> diff --git a/tests/spec/glsl-1.50/execution/fragcoord-layout-qualifiers-conflicting-case-7.shader_test b/tests/spec/glsl-1.50/execution/fragcoord-layout-qualifiers-conflicting-case-7.shader_test
>> index 13605b4..201c75d 100644
>> --- a/tests/spec/glsl-1.50/execution/fragcoord-layout-qualifiers-conflicting-case-7.shader_test
>> +++ b/tests/spec/glsl-1.50/execution/fragcoord-layout-qualifiers-conflicting-case-7.shader_test
>> @@ -48,4 +48,4 @@ void alpha()
>>
>>
>>  [test]
>> -link success
>> +link error
>
> If you make this change, you should update the comment above the
> [require] section as well.
>
> I think I'd have to agree with your interpretation. It seems like "If
> gl_FragCoord is redeclared in any fragment shader in a program, it
> must be redeclared in all the fragment shaders in that program that
> have a static use [of] gl_FragCoord." is pretty clear that this case
> should fail to link.
>
> Anuj, Jordan, Chris?