[Piglit] [PATCH 1/2] Add tests for ARB_transform_feedback3

[Brian Paul]

This is a pretty big patch and I didn't study every line, but it looks
OK overall.  Just minor comments below.

For both: Reviewed-by: Brian Paul <brianp at vmware.com>


On Thu, Jul 5, 2012 at 4:40 AM, Marek Olšák <maraeo at gmail.com> wrote:
> ---
>  tests/all.tests                                 |   11 +
>  tests/spec/ext_transform_feedback/output-type.c |  624 ++++++++++++++++-------
>  2 files changed, 445 insertions(+), 190 deletions(-)
>
> diff --git a/tests/all.tests b/tests/all.tests
> index 07eba93..22cb896 100644
> --- a/tests/all.tests
> +++ b/tests/all.tests
> @@ -1778,6 +1778,17 @@ arb_transform_feedback2 = Group()
>  spec['ARB_transform_feedback2'] = arb_transform_feedback2
>  arb_transform_feedback2['draw-auto'] = PlainExecTest(['arb_transform_feedback2-draw-auto', '-auto'])
>
> +arb_transform_feedback3 = Group()
> +spec['ARB_transform_feedback3'] = arb_transform_feedback3
> +
> +for param in ['gl_NextBuffer-1', 'gl_NextBuffer-2', 'gl_SkipComponents1-1',
> +              'gl_SkipComponents1-2', 'gl_SkipComponents1-3', 'gl_SkipComponents2',
> +              'gl_SkipComponents3', 'gl_SkipComponents4',
> +              'gl_NextBuffer-gl_SkipComponents1-gl_NextBuffer',
> +              'gl_NextBuffer-gl_NextBuffer', 'gl_SkipComponents1234']:
> +        arb_transform_feedback3[param] = concurrent_test(
> +                'ext_transform_feedback-output-type {0}'.format(param))
> +
>  arb_uniform_buffer_object = Group()
>  spec['ARB_uniform_buffer_object'] = arb_uniform_buffer_object
>  import_glsl_parser_tests(spec['ARB_uniform_buffer_object'],
> diff --git a/tests/spec/ext_transform_feedback/output-type.c b/tests/spec/ext_transform_feedback/output-type.c
> index 8b82a1f..016eb01 100644
> --- a/tests/spec/ext_transform_feedback/output-type.c
> +++ b/tests/spec/ext_transform_feedback/output-type.c
> @@ -35,15 +35,18 @@ PIGLIT_GL_TEST_MAIN(
>      32 /*window_height*/,
>      GLUT_DOUBLE | GLUT_RGB | GLUT_ALPHA)
>
> +#define DEFAULT_VALUE 0.123456
> +
>  struct test_desc {
>         const char *name;
>         const char *vs;
>         unsigned num_varyings;
>         const char *varyings[16];
>         bool is_floating_point;
> -       unsigned num_elements;
> -       float expected_float[256];
> -       GLint expected_int[256];
> +       unsigned num_elements[4];
> +       float expected_float[4][256];
> +       GLint expected_int[4][256];

Some comments would be helpful, and/or replace 4 with a MAX_STREAMS #define.
The 256 should probably be a #define also to give some clue to what it means.