<div dir="ltr">On 14 October 2013 11:33, Ian Romanick <span dir="ltr"><<a href="mailto:idr@freedesktop.org" target="_blank">idr@freedesktop.org</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On 10/14/2013 11:20 AM, Paul Berry wrote:<br>
> On 1 October 2013 18:22, Ian Romanick <<a href="mailto:idr@freedesktop.org">idr@freedesktop.org</a><br>
</div><div class="im">> <mailto:<a href="mailto:idr@freedesktop.org">idr@freedesktop.org</a>>> wrote:<br>
><br>
> From: Ian Romanick <<a href="mailto:ian.d.romanick@intel.com">ian.d.romanick@intel.com</a><br>
</div>> <mailto:<a href="mailto:ian.d.romanick@intel.com">ian.d.romanick@intel.com</a>>><br>
<div class="im">><br>
> Verify that the right shaders are used in the right combinations several<br>
> ways.<br>
><br>
> * The vertex shader has information baked-in that determines the X<br>
> position<br>
> of the block on the screen.<br>
><br>
> * The fragment shader has information baked-in that determines how the<br>
> block is colored. This is combined with data passed from the vertex<br>
> shader.<br>
><br>
> Since data is passed from the vertex shader to the fragment shader, the<br>
> test can use either rendezvous-by-name (default) or<br>
> rendezvous-by-location<br>
> (with --by-location command line parameter).<br>
><br>
> This test passes in both modes on NVIDIA (304.64 on GTX 260).<br>
><br>
> Signed-off-by: Ian Romanick <<a href="mailto:ian.d.romanick@intel.com">ian.d.romanick@intel.com</a><br>
</div>> <mailto:<a href="mailto:ian.d.romanick@intel.com">ian.d.romanick@intel.com</a>>><br>
<div><div class="h5">> ---<br>
> tests/all.tests | 2 +<br>
> .../arb_separate_shader_objects/400-combinations.c | 377<br>
> +++++++++++++++++++++<br>
> .../arb_separate_shader_objects/CMakeLists.gl.txt | 1 +<br>
> 3 files changed, 380 insertions(+)<br>
> create mode 100644<br>
> tests/spec/arb_separate_shader_objects/400-combinations.c<br>
><br>
> diff --git a/tests/all.tests b/tests/all.tests<br>
> index 33556af..1e79514 100644<br>
> --- a/tests/all.tests<br>
> +++ b/tests/all.tests<br>
> @@ -1279,6 +1279,8 @@ arb_separate_shader_objects['Mix<br>
> BindProgramPipeline and UseProgram'] = concurre<br>
> arb_separate_shader_objects['ProgramUniform coverage'] =<br>
> concurrent_test('arb_separate_shader_object-ProgramUniform-coverage')<br>
> arb_separate_shader_objects['Rendezvous by location'] =<br>
> plain_test('arb_separate_shader_object-rendezvous_by_location -fbo')<br>
> arb_separate_shader_objects['ValidateProgramPipeline'] =<br>
> concurrent_test('arb_separate_shader_object-ValidateProgramPipeline')<br>
> +arb_separate_shader_objects['400 combinations by location'] =<br>
> plain_test('arb_separate_shader_object-400-combinations -fbo<br>
> --by-location')<br>
> +arb_separate_shader_objects['400 combinations by name'] =<br>
> plain_test('arb_separate_shader_object-400-combinations -fbo')<br>
><br>
> # Group ARB_sampler_objects<br>
> arb_sampler_objects = Group()<br>
> diff --git<br>
> a/tests/spec/arb_separate_shader_objects/400-combinations.c<br>
> b/tests/spec/arb_separate_shader_objects/400-combinations.c<br>
> new file mode 100644<br>
> index 0000000..8593b6e<br>
> --- /dev/null<br>
> +++ b/tests/spec/arb_separate_shader_objects/400-combinations.c<br>
> @@ -0,0 +1,377 @@<br>
> +/*<br>
> + * Copyright © 2013 Intel Corporation<br>
> + *<br>
> + * Permission is hereby granted, free of charge, to any person<br>
> obtaining a<br>
> + * copy of this software and associated documentation files (the<br>
> "Software"),<br>
> + * to deal in the Software without restriction, including without<br>
> limitation<br>
> + * the rights to use, copy, modify, merge, publish, distribute,<br>
> sublicense,<br>
> + * and/or sell copies of the Software, and to permit persons to<br>
> whom the<br>
> + * Software is furnished to do so, subject to the following conditions:<br>
> + *<br>
> + * The above copyright notice and this permission notice (including<br>
> the next<br>
> + * paragraph) shall be included in all copies or substantial<br>
> portions of the<br>
> + * Software.<br>
> + *<br>
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,<br>
> EXPRESS OR<br>
> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF<br>
> MERCHANTABILITY,<br>
> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO<br>
> EVENT SHALL<br>
> + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,<br>
> DAMAGES OR OTHER<br>
> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,<br>
> ARISING<br>
> + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR<br>
> OTHER DEALINGS<br>
> + * IN THE SOFTWARE.<br>
> + */<br>
> +<br>
> +/**<br>
> + * \file 400-combinations.c<br>
> + * Combine 20 vertex shaders and 20 fragment shaders in various ways.<br>
> + *<br>
> + * Verify that the right shaders are used in the right combinations<br>
> several<br>
> + * ways.<br>
> + *<br>
> + * * The vertex shader has information baked-in that determines the<br>
> X position<br>
> + * of the block on the screen.<br>
> + *<br>
> + * * The fragment shader has information baked-in that determines<br>
> how the<br>
> + * block is colored. This is combined with data passed from the<br>
> vertex<br>
> + * shader.<br>
> + *<br>
> + * Since data is passed from the vertex shader to the fragment<br>
> shader, the<br>
> + * test can use either rendezvous-by-name (default) or<br>
> rendezvous-by-location<br>
> + * (with --by-location command line parameter).<br>
> + */<br>
> +#include "piglit-util-gl-common.h"<br>
> +<br>
> +PIGLIT_GL_TEST_CONFIG_BEGIN<br>
> +<br>
> + config.supports_gl_compat_version = 10;<br>
> + config.supports_gl_core_version = 31;<br>
> + config.window_visual = PIGLIT_GL_VISUAL_RGB |<br>
> PIGLIT_GL_VISUAL_DOUBLE;<br>
><br>
><br>
> I'd feel more comfortable if we added:<br>
><br>
> config.window_width = (tile_size + border_size) * ARRAY_SIZE(vs_programs);<br>
> config.window_height = (tile_size + border_size) * ARRAY_SIZE(fs_programs);<br>
><br>
> That way (a) the fact that this test requires a sufficiently large<br>
> window is explicit, and (b) if we try to run this test with a different<br>
> default piglit window size (as we often do when running piglit with a<br>
> hardware simulator) it won't fail.<br>
<br>
</div></div>Good call! I like that. I feel like we should collect some of these<br>
tricks into an "Effective piglit" wiki or something....<br>
<div><div class="h5"><br>
> +<br>
> +PIGLIT_GL_TEST_CONFIG_END<br>
> +<br>
> +/**<br>
> + * Size of each square that will be drawn.<br>
> + */<br>
> +static const unsigned tile_size = 5;<br>
> +<br>
> +/**<br>
> + * Size of the gap between the squares.<br>
> + */<br>
> +static const unsigned border_size = 2;<br>
> +<br>
> +static GLuint vs_programs[20];<br>
> +static GLuint fs_programs[20];<br>
> +<br>
> +static GLuint pipe;<br>
> +<br>
> +static GLuint vao = 0;<br>
> +static GLuint bo = 0;<br>
> +<br>
> +struct combination {<br>
> + unsigned char row;<br>
> + unsigned char col;<br>
> +};<br>
> +<br>
> +static struct combination combinations[ARRAY_SIZE(vs_programs)<br>
> + * ARRAY_SIZE(fs_programs)];<br>
> +<br>
> +static const char *vs_code =<br>
> + "#version %d\n"<br>
> + "#extension GL_ARB_separate_shader_objects: require\n"<br>
> + "#extension GL_ARB_explicit_attrib_location: require\n"<br>
> + "\n"<br>
> + "layout(location = 0) in vec4 piglit_vertex;\n"<br>
> + "layout(location = 1) in vec3 vertex_color;\n"<br>
> + "\n"<br>
> + "%s out vec3 %s;\n"<br>
> + "\n"<br>
> + "const vec4 offset = vec4(%d, 0, 0, 0);\n"<br>
> + "\n"<br>
> + "uniform mat4 transform;\n"<br>
> + "\n"<br>
> + "void main()\n"<br>
> + "{\n"<br>
> + " gl_Position = transform * (piglit_vertex + offset);\n"<br>
> + " %s = vertex_color;\n"<br>
> + "}\n"<br>
> + ;<br>
> +<br>
> +static const char *fs_code =<br>
> + "#version %d\n"<br>
> + "#extension GL_ARB_separate_shader_objects: require\n"<br>
> + "#extension GL_ARB_explicit_attrib_location: enable\n"<br>
> + "\n"<br>
> + "#if __VERSION__ >= 130\n"<br>
> + "layout(location = 0) out vec4 out_color;\n"<br>
> + "#else\n"<br>
> + "#define out_color gl_FragColor\n"<br>
> + "#endif\n"<br>
> + "\n"<br>
> + "%s in vec3 %s;\n"<br>
> + "\n"<br>
> + "const vec3 color_offset = vec3(%d, %d, %d);\n"<br>
> + "\n"<br>
> + "void main()\n"<br>
> + "{\n"<br>
> + " out_color = vec4(%s + color_offset, 1.);\n"<br>
> + "}\n"<br>
> + ;<br>
> +<br>
> +enum piglit_result<br>
> +piglit_display(void)<br>
> +{<br>
> + unsigned i;<br>
> + unsigned j;<br>
> +<br>
> + static const float expected[] = {<br>
> + 0.0f, 1.0f, 0.0f, 1.0f<br>
> + };<br>
> +<br>
> + /* This is stored in row-major order. Note the GL_TRUE<br>
> parameter to<br>
> + * the glProgramUniformMatrix4fv call below.<br>
> + */<br>
> + const float transform[16] = {<br>
> + 2.f / piglit_width, 0.0f, 0.0f, -1.0f,<br>
> + 0.0f, 2.f / piglit_height, 0.0f, -1.0f,<br>
> + 0.0f, 0.0f, 0.0f, 0.0f,<br>
> + 0.0f, 0.0f, 0.0f, 1.0f,<br>
> + };<br>
> +<br>
> + bool pass = true;<br>
> +<br>
> + glClearColor(.5f, .5f, .5f, 1.f);<br>
> + glClear(GL_COLOR_BUFFER_BIT);<br>
> +<br>
> + for (i = 0; i < ARRAY_SIZE(vs_programs); i++) {<br>
> + const GLint loc =<br>
> + glGetUniformLocation(vs_programs[i],<br>
> "transform");<br>
> +<br>
> + glProgramUniformMatrix4fv(vs_programs[i], loc, 1,<br>
> GL_TRUE,<br>
> + transform);<br>
> + }<br>
> +<br>
> + glBindProgramPipeline(pipe);<br>
> +<br>
> + for (i = 0; i < ARRAY_SIZE(combinations); i++) {<br>
> + const unsigned row = combinations[i].row;<br>
> + const unsigned col = combinations[i].col;<br>
> +<br>
> + glUseProgramStages(pipe, GL_VERTEX_SHADER_BIT,<br>
> + vs_programs[col]);<br>
> + glUseProgramStages(pipe, GL_FRAGMENT_SHADER_BIT,<br>
> + fs_programs[row]);<br>
> + glDrawArrays(GL_TRIANGLE_FAN, row * 4, 4);<br>
> + }<br>
> +<br>
> + glBindProgramPipeline(0);<br>
> +<br>
> + for (i = 0; i < ARRAY_SIZE(vs_programs); i++) {<br>
> + for (j = 0; j < ARRAY_SIZE(fs_programs); j++) {<br>
> + const unsigned x = (i * tile_size)<br>
> + + ((i + 1) * border_size);<br>
> + const unsigned y = (j * tile_size)<br>
> + + ((j + 1) * border_size);<br>
> +<br>
> + pass = piglit_probe_rect_rgba(x, y,<br>
> + tile_size,<br>
> tile_size,<br>
> + expected)<br>
> + && pass;<br>
> + }<br>
> + }<br>
> +<br>
> + piglit_present_results();<br>
> + return pass ? PIGLIT_PASS : PIGLIT_FAIL;<br>
> +}<br>
> +<br>
> +#define RED(x) ((int) (x / 2))<br>
> +#define GREEN(x) (-(int) x)<br>
> +#define BLUE(x) ((int) (x * 7))<br>
> +<br>
> +void<br>
> +piglit_init(int argc, char **argv)<br>
> +{<br>
> + unsigned glsl_version;<br>
> + unsigned i;<br>
> + unsigned j;<br>
> + unsigned idx;<br>
> + bool es;<br>
> + int glsl_major;<br>
> + int glsl_minor;<br>
> + const char *location;<br>
> + const char *vertex_name;<br>
> + const char *fragment_name;<br>
> +<br>
> + struct vertex {<br>
> + float x;<br>
> + float y;<br>
> + float r;<br>
> + float g;<br>
> + float b;<br>
> + } *vert;<br>
> +<br>
> + piglit_require_extension("GL_ARB_separate_shader_objects");<br>
> + piglit_require_extension("GL_ARB_explicit_attrib_location");<br>
> +<br>
> + if (argc > 1 && strcmp(argv[1], "--by-location") == 0) {<br>
> + location = "layout(location = 3)";<br>
> + vertex_name = "a";<br>
> + fragment_name = "b";<br>
> + } else {<br>
> + location = "";<br>
> + vertex_name = "in_color";<br>
> + fragment_name = "in_color";<br>
> + }<br>
> +<br>
> + /* Some NVIDIA drivers have issues with layout qualifiers, 'in'<br>
> + * keywords, and 'out' keywords in "lower" GLSL versions.<br>
> If the<br>
> + * driver supports GLSL >= 1.40, use 1.40. Otherwise, pick the<br>
> + * highest version that the driver supports.<br>
> + */<br>
> + piglit_get_glsl_version(&es, &glsl_major, &glsl_minor);<br>
> + glsl_version = ((glsl_major * 100) + glsl_minor) >= 140<br>
> + ? 140 : ((glsl_major * 100) + glsl_minor);<br>
> +<br>
> + /* Generate the vertex shader programs. Each vertex shader is<br>
> + * hardcoded to select a specific column on the display.<br>
> + */<br>
> + printf("Generating vertex shaders...\n");<br>
> + for (i = 0; i < ARRAY_SIZE(vs_programs); i++) {<br>
> + const unsigned base_x = (i * tile_size)<br>
> + + ((i + 1) * border_size);<br>
> +<br>
> + char *source = NULL;<br>
> +<br>
> + asprintf(&source, vs_code,<br>
> + glsl_version,<br>
> + location,<br>
> + vertex_name,<br>
> + base_x,<br>
> + vertex_name);<br>
> +<br>
> + vs_programs[i] =<br>
> + glCreateShaderProgramv(GL_VERTEX_SHADER, 1,<br>
> + (const GLchar *const<br>
> *) &source);<br>
> + piglit_link_check_status(vs_programs[i]);<br>
> +<br>
> + if (i == 0)<br>
> + puts(source);<br>
> +<br>
> + free(source);<br>
> + }<br>
> +<br>
> + printf("Generating fragment shaders...\n");<br>
> + for (i = 0; i < ARRAY_SIZE(fs_programs); i++) {<br>
> + char *source = NULL;<br>
> +<br>
> + asprintf(&source, fs_code,<br>
> + glsl_version,<br>
> + location,<br>
> + fragment_name,<br>
> + RED(i), GREEN(i), BLUE(i),<br>
> + fragment_name);<br>
> +<br>
> + fs_programs[i] =<br>
> + glCreateShaderProgramv(GL_FRAGMENT_SHADER, 1,<br>
> + (const GLchar *const<br>
> *) &source);<br>
> + piglit_link_check_status(fs_programs[i]);<br>
> +<br>
> + if (i == 3)<br>
> + puts(source);<br>
> +<br>
> + free(source);<br>
> + }<br>
> +<br>
> + glGenProgramPipelines(1, &pipe);<br>
> +<br>
> + /* Generate vertex data for the tests. The row of each block is<br>
> + * determined by the vertex data. The color data for the<br>
> block comes<br>
> + * from the vertex data and the data baked into the fragment<br>
> shader.<br>
> + */<br>
> + if (piglit_get_gl_version() >= 30<br>
> + ||<br>
> piglit_is_extension_supported("GL_ARB_vertex_array_object")) {<br>
> + glGenVertexArrays(1, &vao);<br>
> + glBindVertexArray(vao);<br>
> + }<br>
> +<br>
> + glGenBuffers(1, &bo);<br>
> + glBindBuffer(GL_ARRAY_BUFFER, bo);<br>
> + glBufferData(GL_ARRAY_BUFFER,<br>
> + sizeof(vert[0]) * 4 * ARRAY_SIZE(fs_programs),<br>
> + NULL, GL_STATIC_DRAW);<br>
> +<br>
> + vert = glMapBuffer(GL_ARRAY_BUFFER, GL_WRITE_ONLY);<br>
> +<br>
> + for (i = 0; i < ARRAY_SIZE(fs_programs); i++) {<br>
> + const unsigned base_y = (i * tile_size)<br>
> + + ((i + 1) * border_size);<br>
> +<br>
> + vert[(i * 4) + 0].x = 0.f;<br>
> + vert[(i * 4) + 0].y = (float) base_y;<br>
> + vert[(i * 4) + 0].r = (float) -RED(i);<br>
> + vert[(i * 4) + 0].g = (float) 1 - GREEN(i);<br>
> + vert[(i * 4) + 0].b = (float) -BLUE(i);<br>
> +<br>
> + vert[(i * 4) + 1].x = (float) tile_size;<br>
> + vert[(i * 4) + 1].y = (float) base_y;<br>
> + vert[(i * 4) + 1].r = (float) -RED(i);<br>
> + vert[(i * 4) + 1].g = (float) 1 - GREEN(i);<br>
> + vert[(i * 4) + 1].b = (float) -BLUE(i);<br>
> +<br>
> + vert[(i * 4) + 2].x = (float) tile_size;<br>
> + vert[(i * 4) + 2].y = (float) (base_y + tile_size);<br>
> + vert[(i * 4) + 2].r = (float) -RED(i);<br>
> + vert[(i * 4) + 2].g = (float) 1 - GREEN(i);<br>
> + vert[(i * 4) + 2].b = (float) -BLUE(i);<br>
> +<br>
> + vert[(i * 4) + 3].x = 0.f;<br>
> + vert[(i * 4) + 3].y = (float) (base_y + tile_size);<br>
> + vert[(i * 4) + 3].r = (float) -RED(i);<br>
> + vert[(i * 4) + 3].g = (float) 1 - GREEN(i);<br>
> + vert[(i * 4) + 3].b = (float) -BLUE(i);<br>
> + }<br>
> +<br>
> + glUnmapBuffer(GL_ARRAY_BUFFER);<br>
> +<br>
> + glVertexAttribPointer(0, 2, GL_FLOAT, GL_FALSE, sizeof(vert[0]),<br>
> + (void *)(intptr_t) offsetof(struct<br>
> vertex, x));<br>
> + glVertexAttribPointer(1, 3, GL_FLOAT, GL_FALSE, sizeof(vert[0]),<br>
> + (void *)(intptr_t) offsetof(struct<br>
> vertex, r));<br>
> + glEnableVertexAttribArray(0);<br>
> + glEnableVertexAttribArray(1);<br>
> +<br>
> + /* Generate the set of combinations of vertex shader<br>
> programs and<br>
> + * fragment shader programs that will be used together.<br>
> This is all<br>
> + * the possible combinations. The next step is to shuffle<br>
> list so<br>
> + * that there's (hopefully) no pattern to the access<br>
> combination... to<br>
> + * uncover driver bugs.<br>
> + */<br>
> + idx = 0;<br>
> + for (i = 0; i < ARRAY_SIZE(vs_programs); i++) {<br>
> + for (j = 0; j < ARRAY_SIZE(fs_programs); j++) {<br>
> + combinations[idx].row = j;<br>
> + combinations[idx].col = i;<br>
> + idx++;<br>
> + }<br>
> + }<br>
> +<br>
> + for (i = 0; i < (5 * ARRAY_SIZE(combinations)); i++) {<br>
> + /* Pick a random element from the array.<br>
> + */<br>
> + const unsigned src = rand() % ARRAY_SIZE(combinations);<br>
> +<br>
> + /* Pick a random element from the array that is not<br>
> the same<br>
> + * as the previous element. This is done by picking<br>
> a second<br>
> + * number on the range [1, ARRAY_SIZE(combinations)<br>
> - 2] and<br>
> + * adding it (using modular addition) to the first.<br>
> + */<br>
> + const unsigned delta =<br>
> + (rand() % (ARRAY_SIZE(combinations) - 1)) + 1;<br>
> + const unsigned dst = (src + delta) %<br>
> ARRAY_SIZE(combinations);<br>
> +<br>
> + /* Exchange the two selected elements.<br>
> + */<br>
> + const struct combination temp = combinations[dst];<br>
> + combinations[dst] = combinations[src];<br>
> + combinations[src] = temp;<br>
> + }<br>
><br>
><br>
> A less ad-hoc algorithm for shuffling an array is:<br>
><br>
> for (i = ARRAY_SIZE(combinations); i > 1; i--) {<br>
> j = rand() % i;<br>
> if (j != i - 1)<br>
> swap elements j and i - 1<br>
> }<br>
><br>
> Neglecting deficiencies in rand(), this algorithm produces all possible<br>
> permutations with equal probability.<br>
<br>
</div></div>Okay... I like not having to rand() twice.<br>
<br>
I understand (and empathize) your reservations about using rand. I<br>
think we can cope with them in a couple ways.<br>
<br>
1. Implement a piglit_rand. This guarantees that we'll have the same<br>
generator on all platforms, compilers, etc. This doesn't even have to<br>
be a great generator... a dumb LFSR should do it. Most tests that want<br>
random numbers just want a shuffled sequence or some non-sequential data<br>
values.<br>
<br>
2. Implement some standard piglit options for controlling the generator.<br>
These options would be parsed by the piglit framework code. I'm<br>
thinking --random-seed-clock and --random-seed=<some value>.<br>
<br>
3. Make the random number generator log the seed used when it is first<br>
called.<br>
<br>
4. Have a default seed that can be overridden with one of the previously<br>
mentioned command line options.<br>
<br>
Thoughts?<br></blockquote><div><br></div><div>Personally, I'd vote for a combination of options 1 and 4. That would ensure that vanilla piglit runs are repeatable (even across implementations, which is a nice bonus). Options 2 and 3 would be nice, but I don't think they're strictly necessary--I bet if we didn't do them no one would notice/care.<br>
<br></div><div>Incidentally, in case it wasn't clear from my previous email, I think it's just fine to go ahead with this patch before we've resolved what to do about rand().<br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im"><br>
> With the above two changes, this patch is:<br>
><br>
> Reviewed-by: Paul Berry <<a href="mailto:stereotype441@gmail.com">stereotype441@gmail.com</a><br>
</div>> <mailto:<a href="mailto:stereotype441@gmail.com">stereotype441@gmail.com</a>>><br>
<div class="im">><br>
><br>
> +}<br>
> diff --git<br>
> a/tests/spec/arb_separate_shader_objects/CMakeLists.gl.txt<br>
> b/tests/spec/arb_separate_shader_objects/CMakeLists.gl.txt<br>
> index 32a28ba..2e2e1b8 100644<br>
> --- a/tests/spec/arb_separate_shader_objects/CMakeLists.gl.txt<br>
> +++ b/tests/spec/arb_separate_shader_objects/CMakeLists.gl.txt<br>
> @@ -9,6 +9,7 @@ link_libraries (<br>
> ${OPENGL_glu_LIBRARY}<br>
> )<br>
><br>
> +piglit_add_executable (arb_separate_shader_object-400-combinations<br>
> 400-combinations.c)<br>
> piglit_add_executable<br>
> (arb_separate_shader_object-GetProgramPipelineiv GetProgramPipelineiv.c)<br>
> piglit_add_executable (arb_separate_shader_object-IsProgramPipeline<br>
> IsProgramPipeline.c)<br>
> piglit_add_executable<br>
> (arb_separate_shader_object-mix_pipeline_useprogram<br>
> mix_pipeline_useprogram.c)<br>
> --<br>
> 1.8.1.4<br>
><br>
> _______________________________________________<br>
> Piglit mailing list<br>
</div>> <a href="mailto:Piglit@lists.freedesktop.org">Piglit@lists.freedesktop.org</a> <mailto:<a href="mailto:Piglit@lists.freedesktop.org">Piglit@lists.freedesktop.org</a>><br>
> <a href="http://lists.freedesktop.org/mailman/listinfo/piglit" target="_blank">http://lists.freedesktop.org/mailman/listinfo/piglit</a><br>
<br>
</blockquote></div><br></div></div>