[Piglit] [PATCH] sso: fix issues in ValidateProgramPipeline

[Ilia Mirkin]

On Mon, Sep 14, 2015 at 4:34 AM, Tapani Pälli <tapani.palli at intel.com> wrote:
> When fixing some ES 3.1 conformance issues I noticed several subtests
> started to fail. Patch fixes 2 issues in the test:
>
> - pipeline has to be bound for validation to return expected results
> - validate_pipe needs to set global value 'pass' to true when validation
>   passes.

Why? Shouldn't pass be initialized to true once at the start, and then
only ever set to false? That way if any subtest fails, the whole test
is marked as failed. This is how most of these piglits work...

>
> This makes subtests to pass with planned Mesa changes.
>
> Signed-off-by: Tapani Pälli <tapani.palli at intel.com>
> ---
>  tests/spec/arb_separate_shader_objects/ValidateProgramPipeline.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/tests/spec/arb_separate_shader_objects/ValidateProgramPipeline.c b/tests/spec/arb_separate_shader_objects/ValidateProgramPipeline.c
> index a2741aa..5c9230ce 100644
> --- a/tests/spec/arb_separate_shader_objects/ValidateProgramPipeline.c
> +++ b/tests/spec/arb_separate_shader_objects/ValidateProgramPipeline.c
> @@ -174,6 +174,7 @@ validate_pipe(GLuint pipe, bool expected, const char *test_name)
>                 pass = false;
>         } else {
>                 piglit_report_subtest_result(PIGLIT_PASS, "%s", test_name);
> +               pass = true;
>         }
>  }
>
> @@ -323,6 +324,7 @@ piglit_init(int argc, char **argv)
>
>         /* Create the pipeline */
>         glGenProgramPipelines(1, &pipe);
> +       glBindProgramPipeline(pipe);
>
>         build_and_validate_pipe(pipe, true,
>                                 "VS/FS program, single glUseProgramStages "
> --
> 2.4.3
>
> _______________________________________________
> Piglit mailing list
> Piglit at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/piglit