summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2010-02-08clipFlat: relax green color testMarek Olsak1-1/+1
Signed-off-by: Brian Paul <brianp@vmware.com>
2010-02-08texture_srgb: add 1 bit toleranceMarek Olsak1-1/+1
Signed-off-by: Brian Paul <brianp@vmware.com>
2010-02-04texUnits: query/use GL_MAX_COMBINED_TEXTURE_IMAGE_UNITSBrian Paul2-12/+8
Some of the tests were wrong before. This only became apparent after Mesa's handling of GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS was changed (with commit b2a30497cc8b107ea74c3d8fbb646e59a4d55a05).
2010-02-03glsl1: added another cross() test targetting MesaBrian Paul1-0/+21
2010-02-01glsl1: check for GL version 3.1 and 3.2Brian Paul1-1/+3
We're also assuming GL 3.1 or 3.2 has all the backward-compatible GL features. We should also check for GL_ARB_compatibility...
2010-02-01glsl1: add regression test for fd.o bug 26317Brian Paul1-0/+33
2010-01-28glsl1: added some comment parsing testsBrian Paul1-0/+64
2010-01-03Fix MinGW build.José Fonseca1-0/+1
2009-12-21clipFlat: check/test GL_ARB_provoking_vertexBrian Paul1-8/+13
2009-12-21added tokens/typedef for GL_ARB_provoking_vertexBrian Paul1-0/+11
2009-12-16Allow to specify installation directory.José Fonseca5-16/+26
Drop the packaging stuff -- not very cross platform.
2009-12-16Tweak MSVC optimization flags.José Fonseca1-3/+1
2009-12-16Ensure SCons persistent data is put in the build dir and not shared across ↵José Fonseca1-11/+15
variants.
2009-12-14glsl1: added exp2(), log(), log2() testsBrian Paul1-1/+40
2009-12-10glsl1: added literal int, float parsing testsBrian Paul1-0/+31
2009-12-10dsconfig: try exact config match in match() before searchingBrian Paul1-0/+10
This fixes a failed assertion in readpixPerf when comparing results. readpixPerf produces different/extra test results if the visual config has Z and/or stencil. The comparison code was trying to compare results from different visual configs and finding a different number of results so an assertion failed. Now we first look for an exact visual config match before searching for a closest match.
2009-12-09occluQry: new error/warning report functionsBrian Paul1-0/+2
Missed committing the header file on the previous commit.
2009-12-08occluQry: new error/warning report functionsBrian Paul1-32/+50
Don't emit errors/warnings to stderr, send them to the log.
2009-12-08shaderAPI: move some assertions, clean-upsBrian Paul1-13/+8
2009-12-08dsconfig: added equal() method to see if two configs are identicalBrian Paul2-0/+62
2009-12-08dsconfig: zero-out all fields in the string parser constructorBrian Paul2-0/+59
Otherwise, if a field wasn't read from the file, the field had a random value.
2009-12-07glwrap.h: Add PFNGLPOINTPARAMETERIPROC.Vinson Lee1-0/+1
Fixes Mac OS build.
2009-12-03readpixperf: fix bogus res.readBuf valuesBrian Paul1-1/+1
This should have been committed with the previous commit.
2009-12-02readpixperf: fix bogus res.readBuf valuesBrian Paul1-2/+5
Need to write the buffer info to the results file. Otherwise, the value was undefined/random when we read results from a file for comparison.
2009-12-02blendfunc: fix bad calls to nameToFactor()Brian Paul1-2/+10
This fixes some false failures reported by the compareOne() function. Also, add some assertions to catch this kind of bug in the future. Finally, added some minor comments.
2009-12-02texcombine: fix src/dst typo in getresults()Brian Paul1-1/+1
2009-12-02pointsprite: fix signed/unsigned comparison warningBrian Paul1-1/+1
2009-12-02pointsprite: Only test GL_POINT_SPRITE_COORD_ORIGIN if OpenGL 2.0 is supportedIan Romanick2-19/+30
GL_POINT_SPRITE_COORD_ORIGIN is not part of GL_ARB_point_sprite. It was added when the functionality was merged into core OpenGL 2.0. Therefore, if OpenGL 2.0 is not supported, don't test any state related to GL_POINT_SPRITE_COORD_ORIGIN. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Signed-off-by: Brian Paul <brianp@vmware.com>
2009-12-02pointsprite: Test GL_UPPER_LEFT firstIan Romanick2-10/+18
Test the coordinate origin GL_UPPER_LEFT first. Also, log the correct origin string when a test fails. Previously the log message was inverted, and this was very misleading when debugging test failures. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Signed-off-by: Brian Paul <brianp@vmware.com>
2009-12-02pointsprite: Fix grammar / spelling errors in log messages and commentsIan Romanick1-5/+8
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Signed-off-by: Brian Paul <brianp@vmware.com>
2009-12-02pointsprite: Use non-ARB names everywhereIan Romanick1-7/+7
Previously the code used a mixture of, for example, GL_POINT_SPRITE_ARB and GL_POINT_SPRITE. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Signed-off-by: Brian Paul <brianp@vmware.com>
2009-11-08occluqry: Fix memory leak when malloc fails.Vinson Lee1-0/+4
2009-10-01glsl1: Add log2 precision test.Vinson Lee1-0/+17
2009-09-30glsl1: Add exp2 precision test.Vinson Lee1-0/+18
2009-09-29glsl1: sqrt(vec2) testBrian Paul1-0/+16
2009-09-29glsl1: added sqrt(vec4) testBrian Paul1-0/+17
2009-09-24glsl1: added deeply if-nested return testBrian Paul1-0/+21
2009-09-21docs: document --quick and other updatesBrian Paul1-0/+19
2009-09-21glean: when --quick is specified but not --visuals, only test one visualBrian Paul3-5/+14
2009-09-21options: added maxVisuals fieldBrian Paul2-0/+3
2009-09-21dsfilt: add filter() method with maxConfigs paramBrian Paul2-2/+18
2009-09-02pbo: assorted clean-upsBrian Paul1-31/+22
2009-09-02pbo: add missing texture disable and glColor callsBrian Paul1-0/+2
We were getting lucky before. When debugging, disabling some of the sub-tests caused later tests to fail because of mis-set state.
2009-09-02fragProg1: add tests for ADD/SUB with saturationBrian Paul1-0/+31
2009-08-20glsl1: add in-place swizzle testBrian Paul1-0/+15
This is intended to catch failed dependency checking in SOA implementations.
2009-08-20depthStencil: added cast in pixels/second computation to avoid int overflowBrian Paul1-1/+2
This fixes the negative rates that were sometimes reported.
2009-08-06glsl1: s/half/Half/Brian Paul1-8/+8
'half' is a reserved word, don't use it for a function name.
2009-07-29glsl1: add additional GLSL 1.20 array declaration and constructor testsBrian Paul1-0/+80
Mesa doesn't handle these properly yet so they're currently disabled.
2009-07-13fbo: s/GL_LINE/GL_LINES/Vinson Lee1-3/+3
Fix typo of GL_LINES.
2009-07-12tvertprog1: Use gl[En|Dis]ableClientState for GL_VERTEX_ARRAY.Vinson Lee1-2/+2
GL_VERTEX_ARRAY is a client-side capability.
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-22 07:09:49 +0000