Brian Paul | 2 Jan 15:37 2012

Re: [PATCH] clear-accum: New test for i965 driver bug

On 12/30/2011 08:17 PM, Shuang He wrote:
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43783
> Signed-off-by: Shuang He<shuang.he@...>
> ---
>   tests/general/CMakeLists.gl.txt |    1 +
>   tests/general/clear-accum.c     |   51 +++++++++++++++++++++++++++++++++++++++
>   2 files changed, 52 insertions(+), 0 deletions(-)
>   create mode 100644 tests/general/clear-accum.c
>
> diff --git a/tests/general/CMakeLists.gl.txt b/tests/general/CMakeLists.gl.txt
> index f8c5ddf..90711c6 100644
> --- a/tests/general/CMakeLists.gl.txt
> +++ b/tests/general/CMakeLists.gl.txt
>  <at>  <at>  -124,5 +124,6  <at>  <at>  add_executable (object_purgeable-api-vbo object_purgeable-api-vbo.c object_purge
>   add_executable (occlusion_query occlusion_query.c)
>   add_executable (occlusion-query-discard occlusion-query-discard.c)
>   add_executable (quad-invariance quad-invariance.c)
> +add_executable (clear-accum clear-accum.c)
>
>   # vim: ft=cmake:
> diff --git a/tests/general/clear-accum.c b/tests/general/clear-accum.c
> new file mode 100644
> index 0000000..5145620
> --- /dev/null
> +++ b/tests/general/clear-accum.c
>  <at>  <at>  -0,0 +1,51  <at>  <at> 
> +/*
> + * Copyright © 2011 Intel Corporation
> + *
> + * Permission is hereby granted, free of charge, to any person obtaining a
(Continue reading)

Brian Paul | 2 Jan 15:37 2012

Re: [PATCH] map_buffer_range_error_check: Require OpenGL 1.5.

On 12/29/2011 06:21 PM, Vinson Lee wrote:
> glGenBuffers is available only if the GL version is 1.5 or greater.
>
> Signed-off-by: Vinson Lee<vlee@...>
> ---
>   .../map_buffer_range_error_check.c                 |    5 +++++
>   1 files changed, 5 insertions(+), 0 deletions(-)
>
> diff --git a/tests/spec/arb_map_buffer_range/map_buffer_range_error_check.c b/tests/spec/arb_map_buffer_range/map_buffer_range_error_check.c
> index 9abb3f9..87b891e 100644
> --- a/tests/spec/arb_map_buffer_range/map_buffer_range_error_check.c
> +++ b/tests/spec/arb_map_buffer_range/map_buffer_range_error_check.c
>  <at>  <at>  -45,6 +45,11  <at>  <at>  piglit_init(int argc, char *argv[])
>   	GLbitfield access = GL_MAP_READ_BIT | GL_MAP_WRITE_BIT;
>   	GLuint buffer;
>
> +	if (!GLEW_VERSION_1_5) {
> +		printf("Requires OpenGL 1.5\n");
> +		piglit_report_result(PIGLIT_SKIP);
> +	}
> +
>   	piglit_require_extension("GL_ARB_map_buffer_range");
>
>   	glGenBuffers(1,&buffer);

For this and the other patch:
Reviewed-by: Brian Paul <brianp@...>
Brian Paul | 2 Jan 15:48 2012

Re: [PATCH] [v6] Add test case of error check on glGetTexImage.

On 12/29/2011 02:54 AM, Jian Zhao wrote:
> Add test case on error check on glGetTexImage with types(like
> GL_UNSIGNED_INT_10F_11F_11F_REV) that only work with format GL_RGB, but format
> isn't GL_RGB the result should be GL_INVALID_OPERATION.
>
> As glGetTexImage in OpenGL 4.2 Reference Pages:
> "GL_INVALID_OPERATION is returned if type is one of GL_UNSIGNED_BYTE_3_3_2,
> GL_UNSIGNED_BYTE_2_3_3_REV, GL_UNSIGNED_SHORT_5_6_5,
> GL_UNSIGNED_SHORT_5_6_5_REV, or GL_UNSIGNED_INT_10F_11F_11F_REV and format is
> not GL_RGB".
>
> v2: change the test name to be more specific and change some code style to be
> easier to read (per Brian's suggest).
>
> v3: move the case to the spec directory and rename it to
> getteximage_error_check.c. (per Brian's suggest)
>
> v4: add all the formats that only work with GL_RGB format, rename it more
> specific and some minor changes. (per Ian's suggest)
>
> v5: some changes in code styles, like "One space on either side of a binary operator or =". (per Ian's suggest)
>
> v6: fix all the indention issues. (per yuanhan's suggest)
>
> Signed-off-by: Jian Zhao<jian.j.zhao@...>
> Reviewed-by: Brian Paul<brianp@...>
> Reviewed-by: Yuanhan Liu<yuanhan.liu@...>

Looks good. (already reviewed-by).

(Continue reading)

Kenneth Graunke | 2 Jan 22:59 2012

[PATCH] framework: Don't override result to 'fail' on returncode != 0.

We do want to override interpretResult() for crashes, but not for generic
failures.  So default to 'fail', call interpretResult(), then check for
crashes.

Signed-off-by: Kenneth Graunke <kenneth@...>
---
 framework/exectest.py |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/framework/exectest.py b/framework/exectest.py
index d783fa5..04b0d3e 100644
--- a/framework/exectest.py
+++ b/framework/exectest.py
 <at>  <at>  -78,6 +78,7  <at>  <at>  class ExecTest(Test):

 			results = TestResult()

+			results['result'] = 'fail'
 			out = self.interpretResult(out, results)

 			if proc.returncode == -5:
 <at>  <at>  -95,7 +96,6  <at>  <at>  class ExecTest(Test):
 				# Windows EXCEPTION_INT_DIVIDE_BY_ZERO
 				results['result'] = 'crash'
 			elif proc.returncode != 0:
-				results['result'] = 'fail'
 				results['note'] = 'Returncode was %d' % (proc.returncode)

 			env = ''
--

-- 
(Continue reading)

Brian Paul | 2 Jan 23:09 2012

[PATCH] fbo-blit-d24s8: don't try to blit betweendifferent Z/stencil depths

The docs for glBlitFramebuffer say GL_INVALID_OPERATION is generated
if the source/dest depth/stencil buffer formats do no match.  The spec
is a bit vague on what's meant by formats (do just the component depths
need to match, or must the hw formats match?) but clearly if the depths
don't match, the formats can't match.

This fixes a bogus error that's generated when testing swrast which
usually uses a 16-bit depth buffer for the window and a 24-bit Z
renderbuffer for the FBO.

v2: just report PIGLIT_SKIP if the bit depths don't match (per Eric)
---
 tests/fbo/fbo-blit-d24s8.c |   18 ++++++++++++++++++
 1 files changed, 18 insertions(+), 0 deletions(-)

diff --git a/tests/fbo/fbo-blit-d24s8.c b/tests/fbo/fbo-blit-d24s8.c
index 15b9def..91d75dd 100644
--- a/tests/fbo/fbo-blit-d24s8.c
+++ b/tests/fbo/fbo-blit-d24s8.c
 <at>  <at>  -154,6 +154,7  <at>  <at>  run_test(void)
 	int y1 = PAD * 2 + SIZE;
 	int y2 = PAD * 3 + SIZE * 2;
 	GLenum err;
+	GLint win_depth_bits, fbo_depth_bits, win_stencil_bits, fbo_stencil_bits;

 	glViewport(0, 0, piglit_width, piglit_height);
 	piglit_ortho_projection(piglit_width, piglit_height, GL_FALSE);
 <at>  <at>  -171,6 +172,23  <at>  <at>  run_test(void)

 	fbo = make_fbo(fbo_width, fbo_height);
(Continue reading)

Brian Paul | 2 Jan 23:09 2012

[PATCH] incomplete-texture: test sampling from an incomplete texture object

v2: take a command line parameter for fixed vs. ARB_fp vs. GLSL mode.
---
 tests/all.tests                      |    4 +
 tests/texturing/CMakeLists.gl.txt    |    1 +
 tests/texturing/incomplete-texture.c |  310 ++++++++++++++++++++++++++++++++++
 3 files changed, 315 insertions(+), 0 deletions(-)
 create mode 100644 tests/texturing/incomplete-texture.c

diff --git a/tests/all.tests b/tests/all.tests
index 8dab27b..dba691c 100644
--- a/tests/all.tests
+++ b/tests/all.tests
 <at>  <at>  -660,6 +660,10  <at>  <at>  add_plain_test(texturing, 'gen-nonzero-unit')
 add_plain_test(texturing, 'gen-texsubimage')
 add_plain_test(texturing, 'getteximage-formats')
 add_plain_test(texturing, 'getteximage-simple')
+texturing['incomplete-texture-fixed'] = PlainExecTest(['incomplete-texture', '-auto', 'fixed'])
+texturing['incomplete-texture-arb_fp'] = PlainExecTest(['incomplete-texture', '-auto', 'arb_fp'])
+texturing['incomplete-texture-glsl'] = PlainExecTest(['incomplete-texture', '-auto', 'glsl'])
+
 add_plain_test(texturing, 'levelclamp')
 add_plain_test(texturing, 'lodbias')
 add_plain_test(texturing, 'lodclamp')
diff --git a/tests/texturing/CMakeLists.gl.txt b/tests/texturing/CMakeLists.gl.txt
index 3945467..6e12cc0 100644
--- a/tests/texturing/CMakeLists.gl.txt
+++ b/tests/texturing/CMakeLists.gl.txt
 <at>  <at>  -31,6 +31,7  <at>  <at>  add_executable (gen-teximage gen-teximage.c)
 add_executable (gen-texsubimage gen-texsubimage.c)
 add_executable (getteximage-simple getteximage-simple.c)
(Continue reading)

Paul Berry | 2 Jan 23:32 2012
Picon

[PATCH 1/2] Modify piglit_check_gl_error() to return a boolean.

Previously piglit_check_gl_error() would terminate the test if an
error occurred.  However, there are many cases when we would prefer to
simply report the error, and continue testing.

Where it was convenient, I've replaced calls of the form

piglit_check_gl_error(..., PIGLIT_FAIL);

with

pass = piglit_check_gl_error(...) && pass;

So that the test can continue executing.

In other cases, I replaced it with

if (piglit_check_gl_error(...))
        piglit_report_result(PIGLIT_FAIL);

Which is equivalent to the old functionality.
---
 tests/general/roundmode-pixelstore.c               |    4 +-
 tests/hiz/hiz-util.c                               |   25 +++--
 .../api/bindfragdataindexed-invalid-parameters.c   |   24 +++--
 tests/spec/arb_blend_func_extended/api/blend-api.c |   27 ++++--
 .../arb_blend_func_extended/api/getfragdataindex.c |   12 ++-
 tests/spec/arb_draw_buffers/state_change.c         |    6 +-
 ...ufferRenderbuffer-GL_DEPTH_STENCIL_ATTACHMENT.c |    4 +-
 ...mebufferTexture2D-GL_DEPTH_STENCIL_ATTACHMENT.c |    4 +-
 .../map_buffer_range_error_check.c                 |   30 ++++--
(Continue reading)

Paul Berry | 2 Jan 23:32 2012
Picon

[PATCH 2/2] Transform feedback: Test error conditions.

This test verifies that the implementation correctly detects all of
the errors specified in the "Errors" section of the
EXT_transform_feedback spec, with the following exceptions (which seem
like they should be handled in separate tests):

- Errors related to BeginQuery and EndQuery.

- Error due to mismatch of transform feedback mode and drawing mode
  (e.g. drawing GL_LINES when transform feedback is expecting
  GL_TRIANGLES).

- Errors due to exceeding MAX_TRANSFORM_FEEDBACK_SEPARATE_ATTIBS in
  a call to TransformFeedbackVaryings, GetTransformFeedbackVarying,
  or
  Get{Integer,Boolean}Indexedv(GL_TRANSFORM_FEEDBACK_BUFFER_BINDING).
---
 tests/all.tests                                    |   18 +
 .../spec/ext_transform_feedback/CMakeLists.gl.txt  |    1 +
 tests/spec/ext_transform_feedback/api-errors.c     |  377 ++++++++++++++++++++
 3 files changed, 396 insertions(+), 0 deletions(-)
 create mode 100644 tests/spec/ext_transform_feedback/api-errors.c

diff --git a/tests/all.tests b/tests/all.tests
index 9002ada..0290f4c 100644
--- a/tests/all.tests
+++ b/tests/all.tests
 <at>  <at>  -1392,6 +1392,24  <at>  <at>  ext_timer_query['time-elapsed'] = concurrent_test('ext_timer_query-time-elapsed'

 ext_transform_feedback = Group()
 spec['EXT_transform_feedback'] = ext_transform_feedback
(Continue reading)

Shuang He | 3 Jan 02:27 2012
Picon

Re: [PATCH] clear-accum: New test for i965 driver bug

On 2012/1/2 22:37, Brian Paul wrote:
> On 12/30/2011 08:17 PM, Shuang He wrote:
>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43783
>> Signed-off-by: Shuang He<shuang.he@...>
>> ---
>>   tests/general/CMakeLists.gl.txt |    1 +
>>   tests/general/clear-accum.c     |   51 
>> +++++++++++++++++++++++++++++++++++++++
>>   2 files changed, 52 insertions(+), 0 deletions(-)
>>   create mode 100644 tests/general/clear-accum.c
>>
>> diff --git a/tests/general/CMakeLists.gl.txt 
>> b/tests/general/CMakeLists.gl.txt
>> index f8c5ddf..90711c6 100644
>> --- a/tests/general/CMakeLists.gl.txt
>> +++ b/tests/general/CMakeLists.gl.txt
>>  <at>  <at>  -124,5 +124,6  <at>  <at>  add_executable (object_purgeable-api-vbo 
>> object_purgeable-api-vbo.c object_purge
>>   add_executable (occlusion_query occlusion_query.c)
>>   add_executable (occlusion-query-discard occlusion-query-discard.c)
>>   add_executable (quad-invariance quad-invariance.c)
>> +add_executable (clear-accum clear-accum.c)
>>
>>   # vim: ft=cmake:
>> diff --git a/tests/general/clear-accum.c b/tests/general/clear-accum.c
>> new file mode 100644
>> index 0000000..5145620
>> --- /dev/null
>> +++ b/tests/general/clear-accum.c
>>  <at>  <at>  -0,0 +1,51  <at>  <at> 
(Continue reading)

Shuang He | 3 Jan 04:08 2012
Picon

[PATCH] clear-accum: New test for i965 driver bug

v2: Verify result of clearing accum buffer (Brian Paul's suggestion)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43783
Signed-off-by: Shuang He <shuang.he@...>
---
 tests/general/CMakeLists.gl.txt |    1 +
 tests/general/clear-accum.c     |   60 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 61 insertions(+), 0 deletions(-)
 create mode 100644 tests/general/clear-accum.c

diff --git a/tests/general/CMakeLists.gl.txt b/tests/general/CMakeLists.gl.txt
index f8c5ddf..90711c6 100644
--- a/tests/general/CMakeLists.gl.txt
+++ b/tests/general/CMakeLists.gl.txt
 <at>  <at>  -124,5 +124,6  <at>  <at>  add_executable (object_purgeable-api-vbo object_purgeable-api-vbo.c object_purge
 add_executable (occlusion_query occlusion_query.c)
 add_executable (occlusion-query-discard occlusion-query-discard.c)
 add_executable (quad-invariance quad-invariance.c)
+add_executable (clear-accum clear-accum.c)

 # vim: ft=cmake:
diff --git a/tests/general/clear-accum.c b/tests/general/clear-accum.c
new file mode 100644
index 0000000..03be115
--- /dev/null
+++ b/tests/general/clear-accum.c
 <at>  <at>  -0,0 +1,60  <at>  <at> 
+/*
+ * Copyright © 2011 Intel Corporation
+ *
(Continue reading)


Gmane