On 05/11/2012 02:04 AM, Shuang He wrote:
Before:
time ./piglit-print-commands.py -t
'asmparsertest/ARBfp1.0/size_specifier-31.txt' -x
'asmparsertest/ARBfp1.0/size_specifier-31.txt'.+ -x valgrind/ tests/all.tests
asmparsertest/ARBfp1.0/size_specifier-31.txt :::
On 06/04/2012 01:54 PM, Kenneth Graunke wrote:
On 06/04/2012 11:37 AM, Pauli Nieminen wrote:
On Sat, Jun 02, 2012 at 12:56:49PM -0700, Kenneth Graunke wrote:
Pauli, do you plan on pushing this, or should I?
It probably takes some time until I arrange my ssh keys for
git.freedesktop.org
On 06/12/2012 04:02 PM, Chad Versace wrote:
piglit-util.h contained an #ifdef that defined C99 types for MSVC. Remove
that #ifdef block and let MSVC simply include the C99 headers located at
$PIGLIT_TOP/include/msvc/c99.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
On 07/11/2012 04:00 AM, jfons...@vmware.com wrote:
From: José Fonseca jfons...@vmware.com
Verified on NVIDIA.
Looks good to me! Thanks José.
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
On 07/20/2012 11:00 AM, Eric Anholt wrote:
Otherwise, we'd see massive spam of WARN tests, due to the frequency
of the bug (about 2% of testcases on my system).
I like it. For both patches:
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
)
}
}
+ piglit_present_results();
+
return pass ? PIGLIT_PASS : PIGLIT_WARN;
}
Ah, more lovely swap-then-probe code. Thanks for fixing that.
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit
On 09/12/2012 07:51 AM, Tom Stellard wrote:
On Thu, Sep 06, 2012 at 10:44:18PM +0200, Blaž Tomažič wrote:
On čet, 2012-09-06 at 09:16 -0700, Eric Anholt wrote:
Blaž Tomažič blaz.toma...@gmail.com writes:
On tor, 2012-09-04 at 07:14 -0700, Tom Stellard wrote:
Some of the cl-util files contain
On 09/21/2012 08:41 AM, Brian Paul wrote:
On 09/21/2012 03:47 AM, Kenneth Graunke wrote:
[snip]
+#define Elements(x) (sizeof(x)/sizeof(*(x)))
We have an ARRAY_SIZE() macro already.
Oops :) I thought we had one of those, but I must've been blind when I
went looking for it...
+
+enum
On 09/28/2012 01:48 PM, Chad Versace wrote:
I'm trying to transition Piglit from using GLUT to using Waffle. That
requires either killing or wrapping all GLUT code.
Define enum piglit_gl_visual in piglit-framework.h, of which each value
matches its corresponding GLUT visual enum.
Do the
that happens.
Nacked-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
://people.freedesktop.org/~chadversary/piglit.git].
Patches 1-9 are
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
Patches 10-11 are
Acked-by: Kenneth Graunke kenn...@whitecape.org
Patch 12 is
Nacked-by: Kenneth Graunke kenn...@whitecape.org
I'll try and look at the rest, but you don't need to wait for me
On 10/03/2012 04:50 PM, Chad Versace wrote:
On 10/03/2012 04:43 PM, Kenneth Graunke wrote:
On 09/28/2012 01:48 PM, Chad Versace wrote:
/**
+ * A bitmask of these enums specifies visual attributes for the test's window.
+ *
+ * Each enum has the same value of its corresponding GLUT enum
On 10/08/2012 03:21 PM, Chad Versace wrote:
The mesa commit 7dc0be8 changed the EGLConfig ID's for the Intel driver,
and piglit's GTF integration hardcodes the EGLConfig ID. (Apparently,
this implies that piglit's GTF integration only works the Intel driver...
hmm).
Most people don't have a
On 10/09/2012 07:57 AM, Chad Versace wrote:
On 10/08/2012 03:40 PM, Kenneth Graunke wrote:
On 10/08/2012 03:21 PM, Chad Versace wrote:
The mesa commit 7dc0be8 changed the EGLConfig ID's for the Intel driver,
and piglit's GTF integration hardcodes the EGLConfig ID. (Apparently,
this implies
---
tests/glslparsertest/glsl2/condition-08.vert | 8
1 file changed, 8 insertions(+)
create mode 100644 tests/glslparsertest/glsl2/condition-08.vert
diff --git a/tests/glslparsertest/glsl2/condition-08.vert
b/tests/glslparsertest/glsl2/condition-08.vert
new file mode 100644
index
the GLSL
specification and the C++98 specification.
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
It doesn't actually test anything, and thus can never fail.
---
tests/all.tests | 1 -
1 file changed, 1 deletion(-)
diff --git a/tests/all.tests b/tests/all.tests
index 70dca35..b1002d4 100644
--- a/tests/all.tests
+++ b/tests/all.tests
@@ -112,7 +112,6 @@ glean['pointSprite'] =
On 12/30/2012 01:11 PM, Eric Anholt wrote:
jfons...@vmware.com writes:
From: José Fonseca jose.r.fons...@gmail.com
Not actually necessary. rgbTriStrip test used it but it was just a proof
of concept, never part of any test list.
This greatly simplifies building piglit for Windows.
I know
---
tests/all.tests| 1 -
tests/shaders/CMakeLists.gl.txt| 1 -
tests/shaders/glsl-fs-mix-constant.c | 81 --
tests/shaders/glsl-fs-mix-constant.frag| 7 ---
tests/shaders/glsl-fs-mix-constant.shader_test
Shader runner tests are easier to read at a glance than C tests.
---
tests/all.tests | 1 -
tests/shaders/CMakeLists.gl.txt | 1 -
tests/shaders/glsl-fs-mix.c | 92 ---
tests/shaders/glsl-fs-mix.frag| 6 ---
This series deletes roughly 20% of Piglit's import of Glean:
1. Benchmarking
Piglit is a not a benchmarking suite---it's a test suite. It has no
infrastructure for even running these tests, so there's no need to
keep them around.
2. Image output code
Jose nuked most of this,
Piglit is not a benchmarking suite, and this is a benchmark, not a test.
---
tests/glean/CMakeLists.gl.txt | 1 -
tests/glean/treadpixperf.cpp | 551 --
tests/glean/treadpixperf.h| 88 ---
3 files changed, 640 deletions(-)
delete mode 100644
Piglit is not a benchmarking suite, and these are benchmarks, not tests.
---
tests/glean/CMakeLists.gl.txt |1 -
tests/glean/tvtxperf.cpp | 1423 -
tests/glean/tvtxperf.h| 147 -
3 files changed, 1571 deletions(-)
delete mode 100644
Piglit is not a benchmarking suite, and this is a benchmark, not a test.
---
tests/glean/CMakeLists.gl.txt | 1 -
tests/glean/tchgperf.cpp | 317 --
tests/glean/tchgperf.h| 72 --
3 files changed, 390 deletions(-)
delete mode 100644
This allows some further simplifications as well: there's no need for
separate testFunctionality, runSubTests, and runOne functions.
---
tests/glean/tfbo.cpp | 153 +++
tests/glean/tfbo.h | 4 --
2 files changed, 9 insertions(+), 148
There never was any benchmarking code in this test; this was just cut
and pasted from somewhere else.
---
tests/glean/ttexture_srgb.h | 2 --
1 file changed, 2 deletions(-)
diff --git a/tests/glean/ttexture_srgb.h b/tests/glean/ttexture_srgb.h
index 1095565..8875b2d 100644
---
These don't appear to use the Timer functionality, so there's no need to
include it.
---
tests/glean/tshaderapi.cpp | 1 -
tests/glean/ttexcombine4.cpp | 1 -
tests/glean/ttexswizzle.cpp| 1 -
tests/glean/tvertarraybgra.cpp | 1 -
4 files changed, 4 deletions(-)
diff --git
This was previously used for benchmarking code. There isn't any now.
---
tests/glean/CMakeLists.gl.txt | 1 -
tests/glean/timer.cpp | 232 --
tests/glean/timer.h | 74 --
3 files changed, 307 deletions(-)
delete mode
These were used to try and settle the system before a benchmark, and
aren't terribly useful now that there are no benchmarks.
---
tests/glean/environ.cpp | 9 -
tests/glean/environ.h | 2 --
tests/glean/winsys.cpp | 8
tests/glean/winsys.h| 1 -
4 files changed, 20
I suspect most people simply use piglit-summary-html.py to compare
results across test runs. Piglit also supports comparing more than two
test runs.
In some cases, Glean's code did offer more detailed information such as
image comparisons (although not all tests implemented such features).
Jose removed the rgbTriStrip test and TIFF output, which was the only
remaining user of this code.
---
tests/glean/environ.cpp | 15 ---
tests/glean/environ.h | 11 ---
2 files changed, 26 deletions(-)
diff --git a/tests/glean/environ.cpp b/tests/glean/environ.cpp
index
At this point, Glean wasn't writing anything interesting anyway;
it was just clutter.
Since there's no need to specify a results directory on the command line
any longer, this patch also removes the -r option, making run tests
the default action.
This also allows us to simplify the Piglit runner
Now that result comparison is gone, there's no real need to store an
enum here...except for --listtests, it's always run. Simplify.
---
tests/glean/main.cpp| 24
tests/glean/options.cpp | 1 -
tests/glean/options.h | 4
3 files changed, 8 insertions(+), 21
This wasn't used for anything.
---
tests/glean/CMakeLists.gl.txt | 1 -
tests/glean/codedid.cpp | 146 --
tests/glean/codedid.h | 92 --
3 files changed, 239 deletions(-)
delete mode 100644 tests/glean/codedid.cpp
It turns out these frequently-used OpenGL operations weren't used
at all.
---
tests/glean/glutils.cpp | 157
tests/glean/glutils.h | 42 -
2 files changed, 199 deletions(-)
diff --git a/tests/glean/glutils.cpp
Both of these were unused.
---
tests/glean/geomutil.cpp | 156 ---
tests/glean/geomutil.h | 20 --
2 files changed, 176 deletions(-)
diff --git a/tests/glean/geomutil.cpp b/tests/glean/geomutil.cpp
index 01645f7..1852332 100644
---
-bounds-unrolled.shader_test
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
On 01/21/2013 03:01 PM, Paul Berry wrote:
diff --git
a/tests/spec/glsl-es-3.00/execution/varying-struct-arrays.shader_test
b/tests/spec/glsl-es-3.00/execution/varying-struct-arrays.shader_test
Seems like a good mean test...I'd hate to debug it, but worth having.
I didn't review the
This test fills the window with expected depth/stencil values, and uses
CopyTexImage2D to copy those to a texture. It then reads back the
texture and verifies that both depth and stencil were successfully
copied.
This is particularly useful because on Sandybridge and later hardware,
we use a
(ARB_texture_multisample is supposed and
+ redefines this behavior; skipping\n);
Ditto.
Otherwise, patch 1 is:
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
+ piglit_report_result(PIGLIT_SKIP);
+ }
+
glGetIntegerv(GL_MAX_SAMPLES, max_samples
like a good test. Thanks Chris!
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
+check_subtest(t);
+
+piglit_report_result(PIGLIT_PASS);
+}
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman
On 03/28/2013 01:01 AM, Kenneth Graunke wrote:
On 03/27/2013 12:12 AM, Ken Phillis Jr wrote:
I read over the patch, and I noticed a few things.
1) Which platform did you test on? My guess is Linux.
Linux with the AMD Catalyst driver, and Linux with my branch of Mesa
that implements this.
2
On 04/12/2013 04:39 PM, Jon Severinsson wrote:
---
framework/core.py |2 +-
tests/all.tests |6 ++
2 filer ändrade, 7 tillägg(+), 1 borttagning(-)
diff --git a/framework/core.py b/framework/core.py
index 744fda4..62b2ed9 100644
--- a/framework/core.py
+++ b/framework/core.py
Does anyone object to porting to Python 3.x (and dropping 2.x support)?
Some useful data points:
- Debian stable has 3.1, testing/unstable have 3.2
- Fedora 18 and Arch Linux have 3.3
- Windows installers for 3.3 are available on python.org
- Mac OS X support for 3.3 is also available on
Hi Jon,
There's a lot going on in this patch, for example:
1. Search and replace of, say, ConfigParser with configparser
2. try/expect blocks for using both Python 2/3 names for various things
3. Modifying the path all over the place.
4. Reorganizing imports
It'd be nice to split this into
From: Jon Severinsson j...@severinsson.net
This matches the Python 3.x name.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
framework/glsl_parser_test.py | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/framework/glsl_parser_test.py b/framework
From: Jon Severinsson j...@severinsson.net
First, we can try to import it via the Python 3.x name. If that fails,
we can try the Python 2.x name.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
framework/glsl_parser_test.py | 5 -
1 file changed, 4 insertions(+), 1 deletion
From: Jon Severinsson j...@severinsson.net
StringIO comes from the 'cStringIO' module in Python 2.x, and from the
'io' module in Python 3.x.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
framework/core.py | 5 -
framework/glsl_parser_test.py | 7 ++-
2 files
From: Jon Severinsson j...@severinsson.net
This matches the Python 3.x name.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
framework/threadpool.py | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/framework/threadpool.py b/framework/threadpool.py
From: Jon Severinsson j...@severinsson.net
First, we can try to import the Python 3.x name. If that fails, we can
try the Python 2.x name.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
framework/threadpool.py | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git
. This is the approach taken by the 2to3 command.
For more information, see:
http://stackoverflow.com/questions/12172791/changes-in-import-statement-python3
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
framework/core.py | 8
framework/exectest.py | 2
On 04/17/2013 07:37 AM, Brian Paul wrote:
On 04/17/2013 03:48 AM, Kenneth Graunke wrote:
Does anyone object to porting to Python 3.x (and dropping 2.x support)?
Some useful data points:
- Debian stable has 3.1, testing/unstable have 3.2
- Fedora 18 and Arch Linux have 3.3
- Windows installers
On 04/19/2013 09:16 PM, Dylan Baker wrote:
Optparse has been deprecated in favor of argparse upstream (PEP 389
http://www.python.org/dev/peps/pep-0389/), and since we are using
argparse already in the other modules, it makes sense to replace
optparse and use only one parsing method in all
.
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
Signed-off-by: Vinson Lee v...@freedesktop.org
---
tests/spec/amd_performance_monitor/api.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/tests/spec/amd_performance_monitor/api.c
b/tests/spec
On 04/30/2013 10:34 AM, Dylan Baker wrote:
From: Kenneth Graunke kenn...@whitecape.org
Previously, we indented groups of tests by adding 10 columns
of whitespace to the table and adding colspan. This is pretty
ugly; we really should use CSS for things like this.
The new CSS-based approach
to some EGL platform, this problem
caused GLES tests to skip.
Enables 76 GLES tests on my system.
I presume GLX tests keep working with this change (just those tests use
GLX)?
If so,
Acked-by: Kenneth Graunke kenn...@whitecape.org
CC: Eric Anholt e...@anholt.net
Signed-off-by: Chad Versace
On 05/19/2013 07:30 AM, Ken Phillis Jr wrote:
On Sun, May 19, 2013 at 1:07 AM, Kenneth Graunke kenn...@whitecape.org wrote:
On 05/18/2013 09:35 PM, Dylan Baker wrote:
My problem with the current list format is its too complex, and is
trying to solve nonexistent problems. There is no reason
On 05/17/2013 09:31 AM, Dylan Baker wrote:
This new summary class (currently NewSummary), is a complete redesign of
the summary generation system. Much of the code in summary was not
commented, and also not clear as to what it was doing, making it very
difficult to maintain and even more
be allowed.
Assuming you make this change, this patch is:
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
args = parser.parse_args()
-core.checkDir(args.summaryDir, not args.overwrite)
+# If args.list and args.resultsFiles are empty, then give an error
+if not args.list
On 05/17/2013 09:31 AM, Dylan Baker wrote:
With 10,000+ tests all living in the same folder it can be hard to
identify a single file, especially when trying to work with the HTML
generation itself. This patch changes the behavior so that each test
has a directory tree for the group results with
', 'all'],
Shouldn't 'skip' be an option too? By 'skip' I mean tests that were
executed but explicitly output 'skip' (say due to missing extensions),
not missing tests that weren't present in a particular test run.
With those changes, this would get a:
Reviewed-by: Kenneth Graunke kenn
On 05/17/2013 09:32 AM, Dylan Baker wrote:
Previously every page (fixes, regressions, etc) was generated regardless
of whether there was anything on that page. That is simply silly, this
patch does a quick check and only generates pages that actually have
content, and add those pages to the link
On 05/17/2013 09:32 AM, Dylan Baker wrote:
When comparing two test results the current behavior sets a status of
skip for tests that were not run in one set of results. This is
confusing and silly, so instead give them a status of N/A which gives a
clear understanding of what skipped, and what
This should fail as there aren't profiles in GLSL 1.10. It currently
causes assertion failures in Mesa.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/spec/glsl-1.10/compiler/invalid-version-es.vert | 11 +++
1 file changed, 11 insertions(+)
create mode 100644 tests
as expected.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/spec/glsl-1.50/compiler/profiles/core-profile-define.frag| 7 +++
.../spec/glsl-1.50/compiler/profiles/version-150-bad-profile.frag | 7 +++
.../spec/glsl-1.50/compiler/profiles/version-150-core-profile.frag | 7
This is explicitly forbidden: Geometry shader input blocks [...]. All
other input and output block arrays must specify an array size.
The vertex and fragment shader tests pass on Mesa. The GS test has not
been tested.
Cc: Jordan Justen jljus...@gmail.com
Signed-off-by: Kenneth Graunke kenn
On 06/07/2013 10:51 PM, Matt Turner wrote:
[snip]
Both are
Reviewed-by: Matt Turner matts...@gmail.com
Does glslparsertest test need to know about core/compatibility options
for glsl_version?
If we want to have compatibility mode tests, then yes. But since we
aren't planning on
On 06/08/2013 03:21 AM, Fabian Bieler wrote:
On Sat, Jun 8, 2013, at 08:28 AM, Kenneth Graunke wrote:
This is explicitly forbidden: Geometry shader input blocks [...]. All
other input and output block arrays must specify an array size.
The vertex and fragment shader tests pass on Mesa
remove API-only tests (though there are a much
smaller number of those and they run fairly quickly).
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/gpu.tests | 27 +++
1 file changed, 27 insertions(+)
create mode 100644 tests/gpu.tests
This may be kind
On 06/14/2013 06:28 AM, Ken Phillis Jr wrote:
On Thu, Jun 13, 2013 at 4:48 PM, Kenneth Graunke kenn...@whitecape.org wrote:
Often, I make changes to Mesa which only affect tests that actually
execute on the GPU. For such changes, there's no real benefit to
running compiler-only/front-end tests
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/spec/arb_seamless_cube_map/CMakeLists.gl.txt | 1 +
.../arb_seamless_cube_map/initially-disabled.c | 51 ++
2 files changed, 52 insertions(+)
create mode 100644 tests/spec/arb_seamless_cube_map/initially
v2: Actually add the new test to all.tests.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/all.tests| 1 +
tests/spec/arb_seamless_cube_map/CMakeLists.gl.txt | 1 +
.../arb_seamless_cube_map/initially-disabled.c | 51
This must've been cut and pasted from a vertex shader by accident.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
Cc: Nicholas Mack nichm...@gmail.com
---
tests/spec/glsl-1.50/compiler/layout-global-only.frag | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/spec
These are the tests from glsl-1.50/compiler/profiles with the version
numbers changed. It seemed worth testing for 3.30 as well.
---
.../compiler/profiles/core-profile-default.vert | 16
.../glsl-3.30/compiler/profiles/core-profile-define.frag | 7 +++
This is copy and pasted from glsl-1.50/compiler/version-macro.frag with
the version changed.
Since this is the first test for GLSL 3.30, this patch also hooks up the
directory so all.tests scans it for parsertests and shader runner tests.
---
tests/all.tests | 8
;
+
+ piglit_present_results();
+
+ return pass ? PIGLIT_PASS : PIGLIT_FAIL;
+}
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
On 08/23/2013 12:51 PM, Anuj Phogat wrote:
On Fri, Aug 23, 2013 at 11:46 AM, Kenneth Graunke kenn...@whitecape.org wrote:
These are the tests from glsl-1.50/compiler/profiles with the version
numbers changed. It seemed worth testing for 3.30 as well.
---
.../compiler/profiles/core-profile
-alphatest.c | 153 ++
3 files changed, 155 insertions(+)
create mode 100644 tests/fbo/fbo-mrt-alphatest.c
Reviewed-by: Kenneth graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
http
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_shader_bit_encoding extension.
Signed-off-by: Kenneth
the generator script (and
should've been checked in earlier), so I'm OK with it.
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
On 08/28/2013 11:45 AM, Ian Romanick wrote:
On 08/23/2013 11:46 AM, Kenneth Graunke wrote:
This is copy and pasted from glsl-1.50/compiler/version-macro.frag with
the version changed.
Since this is the first test for GLSL 3.30, this patch also hooks up the
directory so all.tests scans
These follow the pattern of the 1.50 and 3.30 tests.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/spec/glsl-1.10/compiler/version-macro.frag| 14 ++
tests/spec/glsl-1.20/compiler/version-macro.frag| 14 ++
tests/spec/glsl-1.30/compiler/version
Again, following the pattern of the other __VERSION__ macro tests.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/spec/glsl-1.10/compiler/version-macro-default.frag | 13 +
1 file changed, 13 insertions(+)
create mode 100644 tests/spec/glsl-1.10/compiler/version
The old tests verified that the value of __VERSION__ was correct, but
didn't verify that __VERSION__ is an actual #define, as opposed to a
built-in integer constant variable.
The new approach checks both.
Suggested-by: Ian Romanick ian.d.roman...@intel.com
Signed-off-by: Kenneth Graunke kenn
These have been superceded by tests/spec/glsl*/compiler/version-macro*.
The parser tests already ensure the value is correct (and fail to
compile if it isn't), so there's no additional value in having an
execution test.
---
tests/shaders/glsl-version-define-110.shader_test | 19
the
command line options in the PIGLIT_GL_TEST_CONFIG_BEGIN block to avoid
this problem.
Signed-off-by: Ian Romanick ian.d.roman...@intel.com
Cc: Kenneth Graunke kenn...@whitecape.org
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing
If the last counter returned is a 32-bit value, reading a uint64_t might
go past the end of the buffer. So, delay reading the 64-bit value until
we know the type is 64-bit.
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
---
tests/spec/amd_performance_monitor/measure.c | 2 +-
1 file
On 09/16/2013 11:08 AM, Marek Olšák wrote:
---
piglit-run.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/piglit-run.py b/piglit-run.py
index e8e11b7..7945b21 100755
--- a/piglit-run.py
+++ b/piglit-run.py
@@ -62,7 +62,7 @@ def main():
(which always fail, and probably will continue to fail for the
forseeable future) in quick.tests makes a lot of sense.
I'm glad to see that you kept streaming-texture-leak. It is pretty
slow, but it at least passes and has caught bugs in the past.
Reviewed-by: Kenneth Graunke kenn
.
---
tests/all.tests | 5
tests/gtf.tests | 85
-
2 files changed, 90 deletions(-)
delete mode 100644 tests/gtf.tests
Makes sense to me.
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
(textureSize.c:68)
Signed-off-by: Kenneth Graunke kenn...@whitecape.org
Cc: Chris Forbes chr...@ijw.co.nz
---
tests/texturing/shaders/textureSize.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/tests/texturing/shaders/textureSize.c
b/tests/texturing/shaders/textureSize.c
index f010d9c
was going to say you should add his copyright, but you've rewritten
most of the code anyway, so I don't think you really need to.
Either way, this looks great.
Reviewed-and-tested-by: Kenneth Graunke kenn...@whitecape.org
+from Queue import Queue
+from threading import Thread
+
+
+class Worker(Thread
On 09/27/2013 02:39 PM, Dylan Baker wrote:
This code tests the way Summary.py assigns statuses to tests.
Specifically it tests regressions, fixes, changes, and problems,
assuring that only the correct permutations of these tests are added to
the list.
I pulled your latest 'status-update-v2'
()})
+ ['result'] for i in
self.results]))}
Ditto.
# Print the summary
print summary:
This is way nicer.
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit
(+), 36 deletions(-)
I'm really happy to see this. This means I get changes.html back, even
for single results, which is what I want. I have that page bookmarked,
and always go there, then click on 'problems' or whatever page I really
want to look at.
Reviewed-by: Kenneth Graunke kenn
,
+# changes and skips
Typos: Craete, regerssions
Otherwise,
Reviewed-by: Kenneth Graunke kenn...@whitecape.org
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
On 10/15/2013 03:58 PM, Dylan Baker wrote:
This was broken by a previous commit 06b185187, this patch corrects
that oversight allowing exclude to continue working.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
piglit-summary-html.py | 15 ---
1 file changed, 12
,
- {GL_EXT_texture_mirror_clamp, NULL}},
+ {GL_EXT_texture_mirror_clamp, NULL, NULL}},
{0}
};
Reviewed-and-tested-by: Kenneth Graunke kenn...@whitecape.org
And...pushed, thanks again for doing this!
___
Piglit mailing list
Piglit
A number of people (at least me, cworth, and ajax) have been hitting a
bug where tests randomly fail with a Bad file descriptor error. It
only occurs with Python 2.7.4 and 2.7.5. Other people have strangely
not been affected.
I did some searching and discovered this Python bug:
On 10/15/2013 05:32 PM, Ian Romanick wrote:
From: Ian Romanick ian.d.roman...@intel.com
The utility is that all.tests can do 'sometest -list-subtests' to
automatically get the list of subtests to run. Since the syntax for
selecting subtests is regularized, this works quite well.
1 - 100 of 321 matches
Mail list logo
