AgeCommit message (Collapse)Author
2012-04-25cmake: Define WIN32_LEAN_AND_MEAN for the whole project.José Fonseca
This avoids namespace pollution when including windows.h, and we don't really need use or need the whole lot of headers that get included by windows.h. In particular, this fixes failure to build ext_framebuffer_multisample accuracy test on windows because some obscure Windows headers define "small" as "char".
2012-04-25glean: Add missing stdlib.h include for system() call.José Fonseca
Include is necessary when building with WIN32_LEAN_AND_MEAN.
2012-04-24Add a test of the accuracy of MSAA rendering.Paul Berry
This is the test I've been using to prototype MSAA on the i965 driver. Unlike the existing EXT_framebuffer_multisample tests (which so far mostly check corner cases at the API level), this test verifies that correct pixels are rendered when MSAA is enabled. It's a bit long for a typical piglit test, since in order to verify that the MSAA color, depth, and stencil buffers are all configured correctly, it has to render some fairly complex scenes.
2012-04-24glsl-1.10: Add a test for a bad shader not breaking a linked program.Eric Anholt
This was a test I came up with from a description of a failure in the webgl conformance suite, but which I don't know the actual sequence of GL calls for. It passes on Mesa, so it didn't catch what I was hoping for. I don't see previous coverage of this in piglit, though.
2012-04-24texelFetch: Add an option for this program that does texelFetchOffset().Eric Anholt
2012-04-24all.tests: Clean up textureSize/texelFetch test setup.Eric Anholt
We can avoid some of the group setup thanks to groups handling '/' characters now.
2012-04-24texelFetch: Add support for *sampler2DRect in GLSL 1.40.Eric Anholt
2012-04-24glsl-1.40: Add tests for the 5 new reserved keywords.Eric Anholt
2012-04-24textureSize: Add testing of GLSL 1.40's texture buffers.Eric Anholt
2012-04-24textureSize: Fix filter setup so that integer rect textures are complete.Eric Anholt
We were leaving things as linear filtering, which is invalid for integer textures.
2012-04-23framework: Remove 'abort' and 'trap' results in favor of 'crash'.Kenneth Graunke
While distinguishing between 'fail' and 'crash' in extremely useful, I've never encountered a situation where 'abort' or 'trap' offers more insight than simply calling it a 'crash'. We may as well simplify things a bit. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org> Acked-by: Marek Olšák <maraeo@gmail.com>
2012-04-21Add test to verify glDrawPixels with allowed pixel formats and data typesAnuj Phogat
v2: Added testing for all supported data types and format/type mismatch v3: Defined new functions, loosened up the tolerence for 2_3_3 and 3_3_2 formats. Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com> Reviewed-by: Brian Paul <brianp@vmware.com>
2012-04-20Rewrite copyteximage to test all texture targets and internal formatsAnuj Phogat
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
2012-04-18framework: Remove GleanTest's duplicate 'results' subdirectory.Kenneth Graunke
In other words, if running piglit-run.py tests/all.tests results/foo, put glean's output in results/foo/glean instead of the more redundant results/foo/results/glean. Reviewed-by: Eric Anholt <eric@anholt.net>
2012-04-18framework: Split a TestProfile.prepare_test_list helper out of run().Kenneth Graunke
Before scheduling or running the tests, run() prepared the final list of tests to run: first, flatten the Group() hierarchy; second, filter out tests based on the -t and -x options. It makes sense to have this as a helper function. Doing so will also enable other utilities that (for example) print a list of tests that would be run and their command line programs. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org> Reviewed-by: Eric Anholt <eric@anholt.net>
2012-04-18Add utility function piglit_draw_triangleAnuj Phogat
These utility functions to draw triangle would help in testing antialiasing and depth related issues. v2: Defined piglit_draw_triangle in terms of piglit_draw_triangle_z. Modified the comments Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com> Reviewed-by: Brian Paul <brianp@vmware.com>
2012-04-17Add new basic GLSL 1.30 textureLod() miplevel tests.Kenneth Graunke
These simply test that miplevel selection and nearest/linear min filtering works. They don't test texture formats, coordinates, or sampler dimensionality. These exposed a bug in i965 VS texturing, where we failed to upload the sampler state and thus had bogus filter modes. This caused us to get solid red for every LOD value. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-17Rewrite glsl-fs-texturelod-01 as a shader_runner test.Kenneth Graunke
This is quite a bit simpler. Technically, I don't need to set the min/mag filters here since "texture miptree" does it already, but it's self-documenting to include that in the test itself. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-17shader_runner: Add support for adjusting the probe tolerance.Kenneth Graunke
Sometimes you just can't rely on the GPU to give the default level of precision. It's useful to be able to adjust this. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-17shader_runner: Add ability to create textures with per-miplevel colors.Kenneth Graunke
For basic mipmapped texture testing, it's often useful to create textures where each miplevel has a unique color. The next commit will use this to simplify an existing test case, and subsequent commits will introduce new tests based on this functionality. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-17shader_runner: Add support for setting the Min/Mag filters.Kenneth Graunke
It's useful and easy to do. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-17shader_runner: Change texparameter syntax to include a space.Kenneth Graunke
"texparameter 2DArray depth_mode greater" makes more sense than the prior spaceless variant. The texture target is a glTexParameter parameter. Unfortunately, changing the syntax means updating existing shader tests. Changes were made automatically via the following shell command: $ find tests -name \*.shader_test | xargs sed -i \ 's/^texparameter/texparameter /g' Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-17shader_runner: Add support for arbitrary texparameter targets.Kenneth Graunke
Previously, we hardcoded the sampler name into the texparameter command itself (i.e. texparameter2DArray) and relied on the main sscanf loop to parse that for us. That doesn't work well once you start adding more than a handful of texture tests. Parsing the texture target in handle_texparameter() allows us to cleanly support the full set. This preserves the spaceless command names for compatibility (and ease of review). The next commit will fix that. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-17shader_runner: Rename generic "mode_table" variable for clarity.Kenneth Graunke
OpenGL has a lot of different kind of modes, so "mode_table" is a bit too generic of a name. Rename it to "drawing_mode_table" for clarity. While we're at it, unify the "mode_table" struct with the "string_to_enum" struct in handle_texparameter, as they're identical, and the latter has a more descriptive name. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-04-10Add test to verify glCopyPixels()Anuj Phogat
This test varifies glCopyPixels() with GL_COLOR, GL_DEPTH and GL_STENCIL v2: Defined few local variables for x,y positions and width, height Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com> Reviewed-by: Brian Paul <brianp@vmware.com>
2012-04-09glx: Verify that GLX_DONT_CARE can be used with bitmask attributes (this ↵Ian Romanick
time for sure) This reproduces Mesa bugzilla #47478. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-04-09GL_ARB_texture_buffer_object/formats: Add a test for all the format options.Eric Anholt
2012-04-09util: Move code for making a half float from a float to shared code.Eric Anholt
v2: Don't forget to reindent to standard style. Reviewed-by: Dave Airlie <airlied@redhat.com>
2012-04-09GL_ARB_texture_buffer_object/unused-name: Test another little bit of spec.Eric Anholt
v2: Fix all.tests line, fix running in -fbo mode by bumping our unused texture number.
2012-04-09GL_ARB_texture_buffer_object/get: New test for remaining queries.Eric Anholt
2012-04-09GL_ARB_texture_buffer_object/dlist: New test for display list handling.Eric Anholt
2012-04-09glsl-1.30: New tests for discard exiting the shader.Eric Anholt
Thes will infinite loop and hang the GPU on failure to implement the correct behavior. This is the result of a clarification in revision 9 of the 1.30 spec, which I noticed while working on GLSL 1.40 (which also got this clarification). The change was not backported to 1.20.
2012-04-06glx: Verify that GLX_DONT_CARE can be used with bitmask attributesIan Romanick
This reproduces Mesa bugzilla #47478. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-04-05draw-elements-base-vertex-bounds: s/NV/ARB/ and add commentsBrian Paul
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
2012-04-03all.tests: add new ARB_extended_blend_func testsDave Airlie
2012-04-03ARB_blend_func_extended: test all blending modes with explicit attrib indexDave Airlie
This is the same as the other fbo test except it uses explicit attrib locations in the shader to set the correct outputs. Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-04-03ARB_blend_func_extended: test all blending modes (v2)Dave Airlie
This tests all the src1 blending modes along with the SRC_ALPHA_SATURATE mode for dst. (it also tests SRC_ALPHA_SATURATE for src just to be complete). This test should complete the tests necessary for merging ARB_blend_func_extended. cleaned up most of Eric Anholt's review comments Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-04-02GL_EXT_transform_feedback/negative-prims: Remove unused variable.Vinson Lee
Fixes this GCC warning. negative-prims.c: In function ‘piglit_display’: negative-prims.c:114: warning: unused variable ‘tf’ Signed-off-by: Vinson Lee <vlee@freedesktop.org> Reviewed-by: Brian Paul <brianp@vmware.com>
2012-03-31glslparsertest doesn't use commasDave Airlie
This lets these tests run when the two extensions are present Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-04-01tests/all.tests: add ARB_blend_func_extended tests to all.testsDave Airlie
2012-03-31arb blend func extended - check errors at glBeginDave Airlie
The spec says we should error after Begin if the blend factors are applied to an illegal draw buffer. Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-03-31blend-api: fix copyright as pointed out by Eric.Dave Airlie
2012-03-31arb_explicit_location: test for index specified without locationDave Airlie
The ARB_explicit_location says it is illegal to specify index without location in the layout. Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-03-27Add a test for GL_EXT_unpack_subimage on GLES2Neil Roberts
The GL_EXT_unpack_subimage extension just allows GL_UNPACK_ROW_LENGTH, GL_UNPACK_SKIP_PIXELS and GL_UNPACK_SKIP_ROWS on GLES2. This tries setting those three values and if the extension is available it will assert that no error is thrown, otherwise it will assert that GL_INVALID_ENUM is thrown. It will then try to upload some texture data with those settings in such a way that if the values were correctly set it will have one pair of pixels otherwise it will have another. The texture is then rendered to the framebuffer and the pixels are compared so it can verify whether the unpack settings were handled correctly. As far as I can tell, this is the first GLES2 test so the patch also creates an all_es2.tests file to contain it.
2012-03-26gl-3.0/minmax: Fix printf format warnings.Vinson Lee
Fixes these GCC warnings. minmax.c: In function ‘max_test_i’: minmax.c:53:4: warning: format ‘%d’ expects argument of type ‘int’, but argument 5 has type ‘GLint *’ [-Wformat] minmax.c:56:3: warning: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘GLint *’ [-Wformat] Signed-off-by: Vinson Lee <vlee@freedesktop.org> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-03-26html: Use isinstance(x, basestring) rather than "str" and "unicode."Kenneth Graunke
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-03-26Add GLSL tests for loops with function calls that have side-effects.Kenneth Graunke
Mesa's loop unroller incorrectly unrolls these loops 7 times, when the function call within the body clearly makes the loop only execute once. Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
2012-03-26glsl-1.40: Add a test for vertex shaders that don't write gl_Position.Eric Anholt
Transform feedback is used to get a testable output, and it's this possibility of using vertex shader execution without rasterization that led to the change in the spec. Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
2012-03-26GL_EXT_transform_feedback/negative-prims: New test for error behavior.Eric Anholt
Mesa just doesn't emit errors.
2012-03-26textureSize: Add GLSL 1.40 textureSize() to all.tests.Eric Anholt
The samplerBuffer tests are currently SKIPs until the test code is added. Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>