Re: [Piglit] [PATCH] arb_gpu_shader5: Use array sizing to catch bad constant folding.

On Mon, Aug 26, 2013 at 10:26 PM, Ian Romanick wrote: > Have you tried this on other drivers? I've seen many cases where things > that seem like the should be errors get optimized out before the error gets > generated. While this is a known "bug" in those implementations, it's > generally better

Re: [Piglit] [PATCH] arb_gpu_shader5: Use array sizing to catch bad constant folding.

On 26 August 2013 22:26, Ian Romanick wrote: > On 08/23/2013 04:32 PM, Matt Turner wrote: > >> The GLSL spec doesn't guarantee that a linking error won't happen. >> --- >> .../built-in-functions/const-**bitCount.shader_test | 40 >> -- >> .../const-bitfieldExtract.**shader

Re: [Piglit] [PATCH] arb_gpu_shader5: Use array sizing to catch bad constant folding.

On 08/23/2013 04:32 PM, Matt Turner wrote: The GLSL spec doesn't guarantee that a linking error won't happen. --- .../built-in-functions/const-bitCount.shader_test | 40 -- .../const-bitfieldExtract.shader_test | 23 + .../const-bitfieldInsert.sha

[Piglit] [PATCH] gs: Test interaction between gl_PrimitiveIDIn and primitive restart.

Verified using the NVIDIA proprietary driver for linux. --- tests/all.tests| 9 + .../glsl-1.50/execution/geometry/CMakeLists.gl.txt | 1 + .../execution/geometry/primitive-id-restart.c | 265 + 3 files changed, 275 insertions(+) cr

[Piglit] [PATCH 01/13 v2] glsl: Refactor some code out of shader_runner.c

From: Ian Romanick These functions will be used in another file soon. v2: Extract string_match (suggested by Eric) and strcpy_to_space as well. Signed-off-by: Ian Romanick --- tests/shaders/CMakeLists.gl.txt| 2 +- tests/shaders/CMakeLists.gles2.txt | 2 +- tests/shaders/CMakeLists.gles

[Piglit] [PATCH 10/13 v2] glsl: Allow limit checking on negative constants

From: Ian Romanick GLSL ES 3.00 adds gl_MinProgramTexelOffset with a negative value. This constant must be -8 or less. v2: Select "<=" comparison based on "gl_Min" name instead of "gl_Max". The original commit message suggested this, and Eric seconded it. Signed-off-by: Ian Romanick --- tests

[Piglit] [PATCH 02/13 v2] glsl: Add program to test all shader built-in constants

From: Ian Romanick Using a simple text file, all shader built-in constants can be tested. This has several advantages over the previous methodology. 1. All of the built-in constants and their expected values are kept in one place per GLSL version. 2. The tests use the array-size idiom of many p

[Piglit] [PATCH 12/13 v2] glsl: Enhance built-in-constants to test extensions too

From: Ian Romanick v2: Use an array of pointers to char for required_extensions (using strndup) instead of the previous static buffer cleverness. Also replace the previous cleverness for generating the GLSL extension enables with something less clever and easier to understand. Both suggested by

Re: [Piglit] [PATCH 12/13] glsl: Enhance built-in-constants to test extensions too

On 08/26/2013 04:55 PM, Ian Romanick wrote: On 08/26/2013 03:32 PM, Eric Anholt wrote: Ian Romanick writes: From: Ian Romanick Signed-off-by: Ian Romanick --- tests/shaders/built-in-constants.c | 90 +- 1 file changed, 89 insertions(+), 1 deletion(-)

Re: [Piglit] [PATCH] cl: From API clRetain* tests, removed attempt to release already released and destroyed objects

On Thu, Aug 22, 2013 at 01:26:04PM +0300, Ville Korhonen wrote: > Problem with releasing already released objects originates from ICD > Loader, which requires that object is alive when calling OpenCL host > functions. Khronos ICD Loader specs > (http://www.khronos.org/registry/cl/extensions/khr/cl_

[Piglit] [PATCH] cl: Add cl-custom-flush-after-enqueue-kernel v2

From: Tom Stellard v2: - Add the license --- tests/all_cl.tests | 1 + tests/cl/custom/CMakeLists.cl.txt| 1 + tests/cl/custom/flush-after-enqueue-kernel.c | 118 +++ 3 files changed, 120 insertions(+) create mode 100644 tests/

[Piglit] [PATCH] cl: Add cl-custom-flush-after-enqueue-kernel

From: Tom Stellard This test checks that clEnqueueReadBuffer() does not begin copying before a kernel referencing its buffer has finished executing. --- tests/all_cl.tests | 1 + tests/cl/custom/CMakeLists.cl.txt| 1 + tests/cl/custom/flush-after-enqueue-k

[Piglit] [PATCH v3] KHR_debug: test ObjectLabel(), GetObjectLabel(), ObjectPtrLabel() and GetObjectPtrLabel()

v3: changed indentation to tabs, use test label constants and remove unused variable Signed-off-by: Timothy Arceri --- tests/all.tests |5 + tests/spec/CMakeLists.txt |1 + tests/spec/khr_debug/CMakeLists.gl.txt| 14 ++ tests/spec/khr_debug

Re: [Piglit] [PATCH 12/13] glsl: Enhance built-in-constants to test extensions too

On 08/26/2013 03:32 PM, Eric Anholt wrote: Ian Romanick writes: From: Ian Romanick Signed-off-by: Ian Romanick --- tests/shaders/built-in-constants.c | 90 +- 1 file changed, 89 insertions(+), 1 deletion(-) diff --git a/tests/shaders/built-in-constant

Re: [Piglit] [PATCH 08/16] glsl-1.40: Refector pass-through vertex shader out to its own file

On 08/26/2013 04:14 PM, Eric Anholt wrote: Ian Romanick writes: From: Ian Romanick Instead of replicating the same code over and over and over, just put it in one file that each shader test can reference. I've been meaning to kill "[vertex/fragment shader file]" support from shader_runner

Re: [Piglit] [PATCH 10/13] glsl: Allow limit checking on negative constants

On 08/26/2013 02:49 PM, Eric Anholt wrote: Ian Romanick writes: From: Ian Romanick GLSL ES 3.00 adds gl_MinProgramTexelOffset with a negative value. This constant must be -8 or less. For now, assume that all negative constants want a <= comparison. The alternative is to compare the name fo

Re: [Piglit] [PATCH 02/13] glsl: Add program to test all shader built-in constants

On 08/26/2013 03:34 PM, Eric Anholt wrote: Ian Romanick writes: From: Ian Romanick Using a simple text file, all shader built-in constants can be tested. This has several advantages over the previous methodology. 1. All of the built-in constants and their expected values are kept in one pla

Re: [Piglit] [PATCH 08/16] glsl-1.40: Refector pass-through vertex shader out to its own file

Ian Romanick writes: > From: Ian Romanick > > Instead of replicating the same code over and over and over, just put it > in one file that each shader test can reference. I've been meaning to kill "[vertex/fragment shader file]" support from shader_runner for a while. :( What I don't like about

Re: [Piglit] [PATCH 04/16] util: Allow non-ES2 code to draw_rect with non-fixed-function arrays

Ian Romanick writes: > From: Ian Romanick > > This will be used in future commits. > > Signed-off-by: Ian Romanick > Cc: Matt Turner > Cc: Paul Berry > --- > tests/util/piglit-util-gl-common.c | 86 > -- > tests/util/piglit-util-gl-common.h | 3 +- > 2 f

Re: [Piglit] [PATCH] glsl-3.30: Add a coverage test for bitcasting functions.

Reviewed-by: Matt Turner ___ Piglit mailing list Piglit@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/piglit

Re: [Piglit] [PATCH 00/13] Simplify GLSL minimum-maximum testing

Ian Romanick writes: > This series simplifies GLSL minimum-maximum testing. Future GLSL > versions can be tested by simply cut-and-pasting from the spec instead > of copying and modifying a bunch of .shader_test files. See patch 8 > (glsl-3.30: Verify the limits in section 7.4) and patch 9 > (g

Re: [Piglit] [PATCH 02/13] glsl: Add program to test all shader built-in constants

Ian Romanick writes: > From: Ian Romanick > > Using a simple text file, all shader built-in constants can be tested. > This has several advantages over the previous methodology. > > 1. All of the built-in constants and their expected values are kept in > one place per GLSL version. > > 2. The te

Re: [Piglit] [PATCH 12/13] glsl: Enhance built-in-constants to test extensions too

Ian Romanick writes: > From: Ian Romanick > > Signed-off-by: Ian Romanick > --- > tests/shaders/built-in-constants.c | 90 > +- > 1 file changed, 89 insertions(+), 1 deletion(-) > > diff --git a/tests/shaders/built-in-constants.c > b/tests/shaders/built-in

Re: [Piglit] Regression with eeec306bc2197a2fc0eeb064462a1244d39c16fa "framework: don't report failed cases as PASS"

On Mon, Aug 26, 2013 at 11:59:40AM -0700, Dylan Baker wrote: > Tom, > I'm a little confused, this was the exact intention of the patch series. > My understanding was that the intention was to fix incorrect reporting when multiple subtests had the same name. For example, with the old behavior, th

Re: [Piglit] [PATCH 10/13] glsl: Allow limit checking on negative constants

Ian Romanick writes: > From: Ian Romanick > > GLSL ES 3.00 adds gl_MinProgramTexelOffset with a negative value. > This constant must be -8 or less. For now, assume that all negative > constants want a <= comparison. The alternative is to compare the > name for 'gl_Min' instead of 'gl_Max'. I'

Re: [Piglit] Regression with eeec306bc2197a2fc0eeb064462a1244d39c16fa "framework: don't report failed cases as PASS"

Tom, I'm a little confused, this was the exact intention of the patch series. Dylan On Monday 26 August 2013 10:09:10 Tom Stellard wrote: > Hi Homer, > > This commit introduces a regression in piglit for reporting subtest > results. Now if one subtests fails, all subtests report failure, which

[Piglit] [PATCH 2/2] Use global function piglit_srgb_to_linear() to avoid code duplication

--- tests/spec/arb_sampler_objects/srgb-decode.c | 2 +- tests/texturing/tex-srgb.c | 2 +- tests/texturing/texwrap.c| 8 +--- 3 files changed, 3 insertions(+), 9 deletions(-) diff --git a/tests/spec/arb_sampler_objects/srgb-decode.c b/tests/spec/arb_sam

[Piglit] [PATCH 1/2] Use global function piglit_linear_to_srgb() to avoid code duplication

--- tests/general/framebuffer-srgb.c | 20 ++-- tests/spec/arb_framebuffer_srgb/srgb_conformance.c | 18 ++ 2 files changed, 4 insertions(+), 34 deletions(-) diff --git a/tests/general/framebuffer-srgb.c b/tests/general/framebuffer-srgb.c index a1

Re: [Piglit] [PATCH] glsl-3.30: Add a coverage test for bitcasting functions.

On 23 August 2013 16:57, Kenneth Graunke wrote: > This ensures that the bitcasting functions are available in 3.30 without > having to use #extension GL_ARB_shader_bit_encoding: require. > > Serious functional testing of those functions is left to Matt's thorough > generated tests for the GL_ARB_

Re: [Piglit] [PATCH] arb_gpu_shader5: Use array sizing to catch bad constant folding.

On 23 August 2013 16:32, Matt Turner wrote: > The GLSL spec doesn't guarantee that a linking error won't happen. > --- > .../built-in-functions/const-bitCount.shader_test | 40 > -- > .../const-bitfieldExtract.shader_test | 23 + > .../const-bitfield

[Piglit] [PATCH 8/9] GL 3.2: Test FramebufferTexture() with layered framebuffer attachments

Currently tests fail, using glGetFramebufferAttachmentParameteriv to check that the attachment is layered. The glError GL_INVALID_ENUM is returned from glGetFramebufferAttachmentParameteriv. It appears that GL_FRAMEBUFFER_ATTACHMENT_LAYERED may not be correctly working with the function --- .../gl

[Piglit] [PATCH 9/9] GL 3.2: Test FramebufferTexture() with buffer textures.

--- .../framebuffertexture-buffer-textures.c | 73 ++ 1 file changed, 73 insertions(+) create mode 100644 tests/spec/gl-3.2/layered-rendering/framebuffertexture-buffer-textures.c diff --git a/tests/spec/gl-3.2/layered-rendering/framebuffertexture-buffer-textures.c

[Piglit] [PATCH 7/9] GL 3.2: Test ReadPixels() works properly with layered framebuffers.

--- tests/spec/gl-3.2/layered-rendering/readpixels.c | 223 +++ 1 file changed, 223 insertions(+) create mode 100644 tests/spec/gl-3.2/layered-rendering/readpixels.c diff --git a/tests/spec/gl-3.2/layered-rendering/readpixels.c b/tests/spec/gl-3.2/layered-rendering/readpixel

[Piglit] [PATCH 6/9] GL 3.2: Test FRAMEBUFFER_ATTACHMENT_LAYERED was added.

Currently test fails. The glError GL_INVALID_ENUM is returned from glGetFramebufferAttachmentParameteriv. It appears that GL_FRAMEBUFFER_ATTACHMENT_LAYERED may not be correctly working with the function --- .../get-framebuffer-attachment-parameter.c | 210 + 1 file cha

Re: [Piglit] [PATCH 1/3] GLSL 1.50: Test that geometry shader must specify input type, output type and maximum output vertices

On 23 August 2013 15:13, Nicholas Mack wrote: > --- > .../linker/gs-must-specify-input-type.shader_test | 52 > + > .../gs-must-specify-max-vertices.shader_test | 53 > ++ > .../linker/gs-must-specify-output-type.shader_test | 53 > +

Re: [Piglit] [PATCH 3/3] GLSL 1.50: Test that linking fails if there is a geom shader without a vert shader

On 23 August 2013 15:13, Nicholas Mack wrote: > --- > .../glsl-1.50/linker/gs-without-vs.shader_test | 39 > ++ > 1 file changed, 39 insertions(+) > create mode 100644 tests/spec/glsl-1.50/linker/gs-without-vs.shader_test > > diff --git a/tests/spec/glsl-1.50/linker/gs-w

Re: [Piglit] [PATCH 2/3] GLSL 1.50: Test that differently specified input/output layouts results in link error

On 23 August 2013 15:13, Nicholas Mack wrote: > --- > ...ax-vertices-multiple-shader-objects.shader_test | 64 > ++ > ...output-type-multiple-shader-objects.shader_test | 64 > ++ > 2 files changed, 128 insertions(+) > create mode 100644 > tests/spec/glsl

[Piglit] [PATCH 5/9] GL 3.2: Test layered framebuffer status

--- .../layered-rendering/framebuffer-layer-complete.c | 168 + 1 file changed, 168 insertions(+) create mode 100644 tests/spec/gl-3.2/layered-rendering/framebuffer-layer-complete.c diff --git a/tests/spec/gl-3.2/layered-rendering/framebuffer-layer-complete.c b/tests/spec/g

[Piglit] [PATCH 04/16] util: Allow non-ES2 code to draw_rect with non-fixed-function arrays

From: Ian Romanick This will be used in future commits. Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- tests/util/piglit-util-gl-common.c | 86 -- tests/util/piglit-util-gl-common.h | 3 +- 2 files changed, 56 insertions(+), 33 deletions(-

[Piglit] [PATCH 4/9] GL 3.2: Test layered framebuffers clear the color attachment properly.

Currently the test only clears the first layer of the layered attachment, when glClear and glClearBuffer should clear all layers of the leyered framebuffer attachment. --- tests/spec/gl-3.2/layered-rendering/clear-color.c | 215 ++ 1 file changed, 215 insertions(+) create mode

[Piglit] [PATCH 1/9] Add layered-rendering tests to CMakeLists

Create new subdirectory: layered-rendering/ Add Tests: layered-rendering/blit.c layered-rendering/clear-color.c layered-rendering/framebuffer-layer-complete.c layered-rendering/framebuffertexture.c layered-rendering/framebuffertexture-buffer-textures.c layered-rendering/get-framebuffer-attachment-

[Piglit] [PATCH 2/9] Add layered-rendering tests to all.tests

Tests Added: blit.c clear-color.c framebuffer-layer-complete.c framebuffertexture.c framebuffertexture-buffer-textures.c get-framebuffer-attachment-parameter.c readpixel.c --- tests/all.tests | 8 1 file changed, 8 insertions(+) diff --git a/tests/all.tests b/tests/all.tests index 6bf9f3

[Piglit] [PATCH 3/9] GL 3.2: Test layered framebuffers blit properly.

--- tests/spec/gl-3.2/layered-rendering/blit.c | 352 + 1 file changed, 352 insertions(+) create mode 100644 tests/spec/gl-3.2/layered-rendering/blit.c diff --git a/tests/spec/gl-3.2/layered-rendering/blit.c b/tests/spec/gl-3.2/layered-rendering/blit.c new file mode

Re: [Piglit] [PATCH 00/13] Simplify GLSL minimum-maximum testing

On 08/26/2013 11:30 AM, gurketsky wrote: On 26.08.2013 19:33, Ian Romanick wrote: This series simplifies GLSL minimum-maximum testing. Future GLSL versions can be tested by simply cut-and-pasting from the spec instead of copying and modifying a bunch of .shader_test files. See patch 8 (glsl-3.

[Piglit] [PATCH 10/16] glsl-1.40: Remove redundant assignment to texcoords

From: Ian Romanick Signed-off-by: Ian Romanick --- tests/spec/glsl-1.40/execution/glsl-fs-shadow2DRect.shader_test | 2 +- tests/spec/glsl-1.40/execution/glsl-fs-shadow2DRectProj.shader_test | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/spec/glsl-1.40/executio

[Piglit] [PATCH 16/16] arb_shading_language_420pack: Refector pass-through fragment shader out to its own file

From: Ian Romanick Instead of replicating the same code over and over and over, just put it in one file that each shader test can reference. Signed-off-by: Ian Romanick Cc: Matt Turner --- .../execution/aggregate-initializer-matrix.shader_test| 8 ++-- .../execution/aggregate-

[Piglit] [PATCH 15/16] arb_shading_language_420pack: Modify tests to use draw rect

From: Ian Romanick Now that "draw rect" works in core profiles, use it! Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- .../execution/aggregate-initializer-matrix.shader_test | 9 + .../execution/aggregate-initializer-sized-array.shader_test | 9 +---

[Piglit] [PATCH 12/16] glsl-1.50: Modify tests that use passthrough.vs to use draw rect

From: Ian Romanick Now that "draw rect" works in core profiles, use it! Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- .../execution/geometry/generate-zero-primitives.shader_test | 9 + tests/spec/glsl-1.50/execution/passthrough.vs | 4 ++-- tests/spec

[Piglit] [PATCH 14/16] arb_shading_language_420pack: Refector pass-through vertex shader out to its own file

From: Ian Romanick Instead of replicating the same code over and over and over, just put it in one file that each shader test can reference. Signed-off-by: Ian Romanick Cc: Matt Turner --- .../execution/const-var-non-constant-expr.shader_test| 9 ++--- .../execution/implicit-r

[Piglit] [PATCH 13/16] glsl-1.50: Modify tests that don't use passthrough.vs to use draw rect

From: Ian Romanick Now that "draw rect" works in core profiles, use it! Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- .../glsl-1.50/execution/geometry-basic.shader_test | 13 +++- .../clip-distance-in-explicitly-sized.shader_test | 13 +++- .../interface-blo

[Piglit] [PATCH 09/16] glsl-1.40: Modify test to use draw rect

From: Ian Romanick Now that "draw rect" works in core profiles, use it! Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- .../execution/gl_DepthRangeParameters.shader_test | 9 + .../glsl-1.40/execution/glsl-fs-shadow2DRect.shader_test | 15 --- ...

[Piglit] [PATCH 03/16] util: Unify desktop OpenGL and OpenGL ES draw_rect functions

From: Ian Romanick After the previous commit, the code for each was identical. Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- tests/util/piglit-util-gl-common.c | 95 tests/util/piglit-util-gl.c| 96 tests/ut

[Piglit] [PATCH 11/16] glsl-1.50: Refector pass-through vertex shader out to its own file

From: Ian Romanick Instead of replicating the same code over and over and over, just put it in one file that each shader test can reference. Signed-off-by: Ian Romanick --- .../execution/geometry/generate-zero-primitives.shader_test | 11 ++- tests/spec/glsl-1.50/execution/passthroug

[Piglit] [PATCH 07/16] shader_runner: Conditionally use piglit_draw_rect_shader

From: Ian Romanick If the shader uses the attribute piglit_vertex with the command "draw rect", assume that it wants to use generic vertex attributes. With the preceeding changes to piglit_draw_rect_from_arrays, "draw rect" should now function a desktop OpenGL core profile. Signed-off-by: Ian R

[Piglit] [PATCH 02/16] util: Refactor desktop OpenGL draw_rect code to use piglit_draw_rect_from_arrays

From: Ian Romanick Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- tests/util/piglit-util-gl.c | 24 +++- 1 file changed, 3 insertions(+), 21 deletions(-) diff --git a/tests/util/piglit-util-gl.c b/tests/util/piglit-util-gl.c index 3dbeec8..cf6e599 100644 -

[Piglit] [PATCH 06/16] util: Use a VAO and VBO for generic attributes in piglit_draw_rect_from_arrays

From: Ian Romanick This will allow users of piglit_draw_rect_from_arrays to function in desktop OpenGL core profile. Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- tests/util/piglit-util-gl-common.c | 53 -- 1 file changed, 51 insertions(+)

[Piglit] [PATCH 05/16] util: Add piglit_draw_rect_shader function

From: Ian Romanick This is just like piglit_draw_rect, but it always uses generic attribute arrays. This will be used in future commits. Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- tests/util/piglit-util-gl-common.c | 28 tests/util/piglit-uti

[Piglit] [PATCH 01/16] util: Move draw_arrays to a public location

From: Ian Romanick Future commits will re-use this code in desktop OpenGL. Signed-off-by: Ian Romanick Cc: Matt Turner Cc: Paul Berry --- tests/util/piglit-util-gl-common.c | 53 + tests/util/piglit-util-gl-common.h | 1 + tests/util/piglit-util-gles.c |

[Piglit] [PATCH 08/16] glsl-1.40: Refector pass-through vertex shader out to its own file

From: Ian Romanick Instead of replicating the same code over and over and over, just put it in one file that each shader test can reference. Signed-off-by: Ian Romanick --- .../glsl-1.40/execution/gl_DepthRangeParameters.shader_test | 11 ++- tests/spec/glsl-1.40/passthrough.vs

[Piglit] [PATCH 00/16] draw rect in core profile

The first 7 patches make 'draw rect' work in core profile shader tests. The remaining patches modify existing tests to use 'draw rect' instead of 'draw arrays'. There are additional tests that could be converted, but I basically ran out of motivation... and started shaving other yaks. I started t

Re: [Piglit] [PATCH 00/13] Simplify GLSL minimum-maximum testing

On 26.08.2013 19:33, Ian Romanick wrote: This series simplifies GLSL minimum-maximum testing. Future GLSL versions can be tested by simply cut-and-pasting from the spec instead of copying and modifying a bunch of .shader_test files. See patch 8 (glsl-3.30: Verify the limits in section 7.4) and

Re: [Piglit] [PATCH 2/2] GLSL 1.50: Test that struct varyings are passed correctly from GS to FS

On 23 August 2013 11:02, Nicholas Mack wrote: > --- > .../varying-struct-basic-vs-gs-fs.shader_test | 141 > + > 1 file changed, 141 insertions(+) > create mode 100644 > tests/spec/glsl-1.50/execution/varying-struct-basic-vs-gs-fs.shader_test > > diff --git > a/tests/sp

Re: [Piglit] [PATCH 1/2] GLSL 1.50: Test that geometry shader output may be a struct

On 23 August 2013 11:02, Nicholas Mack wrote: > --- > tests/spec/glsl-1.50/compiler/output-struct.geom | 52 > > 1 file changed, 52 insertions(+) > create mode 100644 tests/spec/glsl-1.50/compiler/output-struct.geom > This patch is: Reviewed-by: Paul Berry __

[Piglit] [PATCH 11/13] glsl-es-3.30: Verify the limits in section 7.3

From: Ian Romanick minimum-maximums.txt was created by copy-and-pasting directly from section 7.3 followed by 's/ = / /;s/;//'. This is why the names are not listed in alphabetic order. (In the OpenGL ES 3.00 spec, the built-in constants are in section 7.3 instead of section 7.4.) Signed-off-b

[Piglit] [PATCH 05/13] glsl-1.30: Convert minimum-maximums tests to use built-in-constants program

From: Ian Romanick minimum-maximums.txt was generated by: cd tests/spec/glsl-1.30 echo 1.30 > minimum-maximums.txt grep gl_Max execution/maximums/gl_Max* |\ sort |\ sed 's/[^(]\+[(]//;s/ >= / /;s/[)]$//' >> minimum-maximums.txt Signed-off-by: Ian Romanick --- tests

[Piglit] [PATCH 08/13] glsl-3.30: Verify the limits in section 7.4

From: Ian Romanick minimum-maximums.txt was created by copy-and-pasting directly from section 7.4 followed by 's/ = / /;s/;//'. This is why the names are not listed in alphabetic order. Signed-off-by: Ian Romanick --- tests/all.tests | 4 tests/spec/glsl-3.30/m

[Piglit] [PATCH 04/13] glsl-1.20: Convert minimum-maximums tests to use built-in-constants program

From: Ian Romanick minimum-maximums.txt was generated by: cd tests/spec/glsl-1.20 echo 1.20 > minimum-maximums.txt grep gl_Max execution/maximums/gl_Max* |\ sort |\ sed 's/[^(]\+[(]//;s/ >= / /;s/[)]$//' >> minimum-maximums.txt Signed-off-by: Ian Romanick --- tests

[Piglit] [PATCH 12/13] glsl: Enhance built-in-constants to test extensions too

From: Ian Romanick Signed-off-by: Ian Romanick --- tests/shaders/built-in-constants.c | 90 +- 1 file changed, 89 insertions(+), 1 deletion(-) diff --git a/tests/shaders/built-in-constants.c b/tests/shaders/built-in-constants.c index 9e5b940..fb955ae 100644

[Piglit] [PATCH 13/13] arb_shading_language_420pack: Verify the limits in section 7.4

From: Ian Romanick The GL_ARB_shading_language_420pack says: "This is a language feature only extension formed from changes made to version 4.20 of GLSL. It includes: ... * Built-in constants for gl_MinProgramTexelOffset and gl_MaxProgramTexelOffset." Signed-off-by: Ian

[Piglit] [PATCH 03/13] glsl-1.10: Convert minimum-maximums tests to use built-in-constants program

From: Ian Romanick minimum-maximums.txt was generated by: cd tests/spec/glsl-1.10 echo 1.10 > minimum-maximums.txt grep gl_Max execution/maximums/gl_Max* |\ sort |\ sed 's/[^(]\+[(]//;s/ >= / /;s/[)]$//' >> minimum-maximums.txt Signed-off-by: Ian Romanick --- tests

[Piglit] [PATCH 01/13] glsl: Refactor some code out of shader_runner.c

From: Ian Romanick These functions will be used in another file soon. Signed-off-by: Ian Romanick --- tests/shaders/CMakeLists.gl.txt| 2 +- tests/shaders/CMakeLists.gles2.txt | 2 +- tests/shaders/CMakeLists.gles3.txt | 2 +- tests/shaders/parser_utils.c | 49

[Piglit] [PATCH 09/13] glsl-es-1.00: Convert minimum-maximums tests to use built-in-constants program

From: Ian Romanick minimum-maximums.txt was generated by: cd tests/spec/glsl-es-1.00 echo 1.00 > minimum-maximums.txt grep gl_Max execution/maximums/gl_Max* |\ sort |\ sed 's/[^(]\+[(]//;s/ >= / /;s/[)]$//' >> minimum-maximums.txt Signed-off-by: Ian Romanick --- te

[Piglit] [PATCH 07/13] glsl-1.50: Convert minimum-maximums tests to use built-in-constants program

From: Ian Romanick minimum-maximums.txt was created by copy-and-pasting directly from section 7.4 followed by 's/ = / /;s/;//'. This is why the names are not listed in alphabetic order. The old tests did not verify all of the constants. Signed-off-by: Ian Romanick Cc: Nicholas Mack --- .../

[Piglit] [PATCH 06/13] glsl-1.40: Convert minimum-maximums tests to use built-in-constants program

From: Ian Romanick minimum-maximums.txt was generated by: cd tests/spec/glsl-1.40 echo 1.40 > minimum-maximums.txt grep gl_Max execution/maximums/gl_Max* |\ sort |\ sed 's/[^(]\+[(]//;s/ >= / /;s/[)]$//' >> minimum-maximums.txt Signed-off-by: Ian Romanick --- tests

[Piglit] [PATCH 00/13] Simplify GLSL minimum-maximum testing

This series simplifies GLSL minimum-maximum testing. Future GLSL versions can be tested by simply cut-and-pasting from the spec instead of copying and modifying a bunch of .shader_test files. See patch 8 (glsl-3.30: Verify the limits in section 7.4) and patch 9 (glsl-es-3.30: Verify the limits in

[Piglit] [PATCH 02/13] glsl: Add program to test all shader built-in constants

From: Ian Romanick Using a simple text file, all shader built-in constants can be tested. This has several advantages over the previous methodology. 1. All of the built-in constants and their expected values are kept in one place per GLSL version. 2. The tests use the array-size idiom of many p

[Piglit] [PATCH 10/13] glsl: Allow limit checking on negative constants

From: Ian Romanick GLSL ES 3.00 adds gl_MinProgramTexelOffset with a negative value. This constant must be -8 or less. For now, assume that all negative constants want a <= comparison. The alternative is to compare the name for 'gl_Min' instead of 'gl_Max'. Signed-off-by: Ian Romanick --- te

[Piglit] Regression with eeec306bc2197a2fc0eeb064462a1244d39c16fa "framework: don't report failed cases as PASS"

Hi Homer, This commit introduces a regression in piglit for reporting subtest results. Now if one subtests fails, all subtests report failure, which is incorrect. I've looked through the changes in this commit, and I don't understand which change introduced the regression, could you take a look?

Re: [Piglit] [PATCH] gs: Test EndPrimitive() functionality.

On 23 August 2013 12:17, Kenneth Graunke wrote: > On 08/11/2013 10:32 PM, Paul Berry wrote: > >> --- >> tests/all.tests| 9 + >> tests/spec/glsl-1.50/**execution/CMakeLists.txt | 1 + >> .../glsl-1.50/execution/**geometry/CMakeLists.gl.txt | 13 +

Re: [Piglit] [PATCH v2] KHR_debug: test ObjectLabel(), GetObjectLabel(), ObjectPtrLabel() and GetObjectPtrLabel()

On 08/24/2013 06:35 PM, Timothy Arceri wrote: v2: fixed whitespaces, broke tests into subtests, added missing build dir, and tidied up fail messages Signed-off-by: Timothy Arceri --- tests/all.tests |5 + tests/spec/CMakeLists.txt |1 + test

[Piglit] [v8] tests/spec: EXT_image_dma_buf_import sample argb

Tests that one can sample the given buffer without other mipmap levels by setting texture filtering accordingly. v2: - compile only on platforms that have drm (Eric) - use standard drm definitions for fourcc instead of duplicated local (Daniel, Eric) - removed irrelevant quote of the