FWIW, the skqp test is run constantly in mesa ci:
https://mesa-ci.01.org/mesa_master/builds/15138/results/174306816
Aditya Swarup writes:
> Commit 649c149dbb96ac367 removed the test for intel external sampler
> with dma only because of failure due to commit a5c39ed974402c -
> ("i965: Lift rest
It's preferable to push piglit tests first, and produce the CI test
failure. CI staff will track the failure, and attribute it to the
piglit patch. When the subsequent mesa patch is pushed, we track the
resolution. The process gives us artifacts in git that help to analyze
driver differences bet
Thanks for writing this test. I checked it in CI and pushed it.
-Mark
Danylo Piliaiev writes:
> Hello,
>
> Since the patch which fixes the issue got pushed in
> f9fd0cf4790cb2a530e75d1a2206dbb9d8af7cb2
> could this test be reviewed/pushed?
>
> Thanks!
>
> On 10/24/18 2:17 PM, Danylo Piliaiev w
Let me know when this is pushed, and I'll enable the tests in i965 CI.
Dylan Baker writes:
> Quoting Neil Roberts (2018-11-13 12:50:16)
>> This will make it easier to distinguish tests written for Vulkan. It
>> makes a bit of an inconsistency because the GL and CL tests don\u2019t have
>> any pr
Reviewed-by: Mark Janes
Tested-by: Mark Janes
Neil Roberts writes:
> This will make it easier to distinguish tests written for Vulkan. It
> makes a bit of an inconsistency because the GL and CL tests don’t have
> any prefix. Ideally maybe we would add a prefix for those too, but
&
Tested-by: Mark Janes
Sergii Romantsov writes:
> Intel ppgtt-model requires memory addresses to be aligned by page size.
> Test explores cases with cached memory-buckets and sizes above cached.
>
> CC: Kenneth Graunke
> CC: Lionel G Landwerlin
> CC: Chris Wilson
Reviewed-by: Mark Janes
Tested-by: Mark Janes
Dylan Baker writes:
> Otherwise these tests fail in odd ways when running from an installed
> instance.
>
> Cc: Alejandro Piñeiro
> Cc: Mark Janes
> Cc: Clayton Craft
> ---
> CMakeLists.txt | 2 +-
> 1 file changed,
Hi Neil,
We still get intermittent failures with this threshold, and have
historically disabled the multisample accuracy tests in CI.
Can we relax the threshold a bit more?
-Mark
Neil Roberts writes:
> The unlit and totally_lit stats are supposed to count the pixels where
> either a primitive
Dylan Baker writes:
> Quoting Marek Olšák (2018-04-10 14:22:10)
>> On Tue, Apr 10, 2018 at 2:15 PM, Dylan Baker wrote:
>>
>> Quoting Eric Anholt (2018-04-09 17:10:35)
>> > Marek Olšák writes:
>> >
>> > > Is this use case affected?
>> > >
>> > > piglit run --deqp-mustpas
to run the renamed
> tests in the master branch of the opensourced Khronos CTS tests at:
> https://github.com/KhronosGroup/VK-GL-CTS
>
> Cc: Mark Janes
> Cc: Dylan Baker
> Cc: Juan A. Suarez Romero
> Signed-off-by: Andres Gomez
> ---
> tests/khr_gles.py | 88
> +++
Tested-by: Mark Janes
Andres Gomez writes:
> After:
> https://github.com/KhronosGroup/VK-GL-CTS/commit/af8c22a343ee2c230488f6de71b36dc3070b2024
>
> OpenGL GL*-CTS case lists were renamed to KHR-GL*.
>
> We want to keep the existing profiles so we are able to keep running
&g
Tested-by: Mark Janes
Andres Gomez writes:
> Recently I added some changes in the cts_gl* profiles to adapt to
> renames in the opensourced Khronos CTS at:
> https://github.com/KhronosGroup/VK-GL-CTS
>
> Unfortunately, I didn't foresee that this change would break backward
Hi Ian,
This test fails on g33. Was that expected?
-Mark
Ian Romanick writes:
> From: Ian Romanick
>
> This allows fbo-blit-stretch to run on platforms that support
> ARB_framebuffer_object but not ARB_texture_non_power_of_two.
>
> Coneptually similar to 7b3f6d5.
>
> Signed-off-by: Ian Roman
Series is:
Tested-by: Mark Janes
Samuel Pitoiset writes:
> Since b741bc770 ("parser_utils: do not overwrite value when no
> digits are found "), the blue color has to be explicitly initialized
> to zero.
>
> Cc: Mark Janes
> Signed-off-by: Samuel Pitoiset
>
Tested-by: Mark Janes
Chad Versace writes:
> The test generates a mipmap by "copying" a client-array of pixels to
> miplevel 0 with glDrawPixels, then calling glGenerateMipmap. The test
> expects the rendered level 0 to match the client-array source. Dithering
> during
Recent versions of the GL CTS report X connection errors on standard
out. Support all variants by checking both stdout and stderr for the
error string.
---
framework/test/deqp.py | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/framework/test/deqp.py b/framework/test/deqp
Tested-by: Mark Janes
Francisco Jerez writes:
> This makes sure that reads and writes of the volatile-qualified images
> occur in the expected order. It's unclear whether the GLSL volatile
> qualifier was intended to have the required memory ordering
> implications as it
This patch allows the test to complete on a Braswell system with 1.5GB
memory. Without it, the OOM killer is dispatched.
Tested-by: Mark Janes
Kenneth Graunke writes:
> A 16384x16384 RGBA buffer takes 1GB of RAM. If the driver allocates
> mipmap levels, that increases our s
Tested-by: Mark Janes
Tapani Pälli writes:
> Signed-off-by: Tapani Pälli
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98769
> ---
> tests/util/CMakeLists.txt | 5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/tests/util/CMakeLists.txt b/test
This series breaks my compilation:
gcc -m64 -Wall -std=gnu99 -Werror=vla -Werror=pointer-arith
-Werror=variadic-macros -g
target_api/gl/tests/bugs/CMakeFiles/point-sprite.dir/point-sprite.c.o -o
bin/point-sprite -rdynamic lib/libpiglitutil_gl.so.0 -lGL
lib/libpiglitutil.so.0 -lrt -ldl -lx
Tested-by: Mark Janes
Reviewed-by: Mark Janes
Dylan Baker writes:
> This reduces memory consumption by a lot (about 60% on my system), which
> fixes some tests that fail sporadically on low memory systems.
>
> cc: Mark Janes
> Signed-off-by: Dylan Baker
> ---
> fra
The junit loader is unnecessarily strict with the input that it
accepts. It expects input generated by piglit, but can be made to
handle junit from other test suites like crucible.
v2: Remove unnecessary initializers
Conditionally parse command for piglit tests
---
framework/backends/junit.p
Dylan Baker writes:
> Quoting Mark Janes (2016-10-26 16:30:12)
>> The junit loader is unnecessarily strict with the input that it
>> accepts. It expects input generated by piglit, but can be made to
>> handle junit from other test suites like crucible.
>> ---
>
The junit loader is unnecessarily strict with the input that it
accepts. It expects input generated by piglit, but can be made to
handle junit from other test suites like crucible.
---
framework/backends/junit.py | 28
1 file changed, 20 insertions(+), 8 deletions(-)
; Mark and Michel,
> I'd appreciate if you could test this and make sure it fixes your respective
> problems.
Tested-by: Mark Janes
> Dylan Baker (3):
> piglit: Only catch parsed.func AttributeError
> framework: Properly load JSON into internal representation
>
Dan Kegel writes:
> Hrm. A solution that involved Jenkins would be way overkill.
> I'm just looking for the little core idiom for running just the
> test of tests that passed in the, um, past.
Even if you don't want to set up a CI with jenkins, you can refer to the
automation that I've written
. I don't need to run a single subtest.
Series is
Reviewed-by: Mark Janes
Tested-by: Mark Janes
Dylan Baker writes:
> This series refactors the JUnit backend so that the _write attribute is
> actually a callable class, instead of a method. This gives greater
> flexibility to h
Some minor nits on the comments below.
Reviewed-by: Mark Janes
Tested-by: Mark Janes
Dylan Baker writes:
> This adds the necessary bits to make dEQP-EGL work.
>
> Signed-off-by: Dylan Baker
> CC: Mark Janes
> ---
> piglit.conf.example | 10 ++
> tes
) strtoll(*text, &endptr, 0);
> - if (errno == ERANGE) {
> + long long value = strtoll(*text, &endptr, 0);
> + if (errno == ERANGE || (unsigned long long) value >
> 0xull) {
^^^
deqp.iter_deqp_test_cases(
>> - deqp.gen_caselist_txt(_CTS_BIN, 'ESEXT-CTS-cases.txt',
>> - _EXTRA_ARGS)),
>> +deqp.gen_caselist_txt(_CTS_BIN, 'GL45-CTS-cases.txt',
>> _EXTRA_ARGS)),
>> ),
Marek Olšák writes:
> On Fri, May 27, 2016 at 3:18 AM, Mark Janes wrote:
>> Marek Olšák writes:
>>
>>> On Mon, Apr 18, 2016 at 6:45 PM, Dylan Baker
>>> wrote:
>>>> Quoting Marek Olšák (2016-04-16 15:16:34)
>>>>> Hi,
>>>&
Marek Olšák writes:
> On Mon, Apr 18, 2016 at 6:45 PM, Dylan Baker wrote:
>> Quoting Marek Olšák (2016-04-16 15:16:34)
>>> Hi,
>>>
>>> This makes shader_runner very fast. The expected result is 40%
>>> decrease in quick.py running time, or a 12x faster piglit run if you
>>> run shader tests alon
Reviewed-by: Mark Janes
Dylan Baker writes:
> In commit b0d05323e code was added to produce clearer error messages for
> tests who's status changed from crash to fail and vice versa. There is a
> not so subtle bug in that patch, it adds a "crash" element, but that
>
Reviewed-by: Mark Janes
Dylan Baker writes:
> glslparsertest has a syntax for the 'require_extensions' option that is
> "!", when this is specified it means "This extensions is
> not supported". There is a bug in the fast skip code in
> glsl_
Reviewed-by: Mark Janes
Dylan Baker writes:
> A pretty common error it seems is to get a failure to open display when
> using X11. This adds an extra method to the DEQPBaseTest class that
> reruns the test is it finds that error in the stderr. This affects only
> deqp based suite
I don't need to see a v2
Acked-by: Mark Janes
Dylan Baker writes:
> Quoting Mark Janes (2016-03-15 17:40:16)
>> Dylan Baker writes:
>>
>> > Right now the JUnit backend knows what to do with a failure when it
>> > expects one (or a crash), but no
Mark Janes writes:
> This patch doesn't fix the issue I see on HSW with Resource errors.
I take back this statement. I was confused because the tests went from
fail to crash, but there was no indication of this discrepancy in the
output.
Thanks for your recent patch to make this mo
Dylan Baker writes:
> Right now the JUnit backend knows what to do with a failure when it
> expects one (or a crash), but not what to do when it expects a failure
> and gets a crash (or vice versa).
>
> This patch teaches it what to do in that case, mark the test as a
> failure and give a message
This patch doesn't fix the issue I see on HSW with Resource errors.
Dylan Baker writes:
> This is a previously unseen error type generated on by the Vulkan CTS
> (which is dEQP based).
>
> Signed-off-by: Dylan Baker
> ---
>
> Mark, I think this will fix the that jenkins bug we're seeing.
>
> f
Reviewed-by: Mark Janes
Kenneth Graunke writes:
> The test was passing a pointer to an uninitialized integer value as the
> glShaderSource length parameter. This is meant to be an array
> containing the length of each shader string.
>
> Instead, just pass NULL, which makes th
Alejandro Piñeiro writes:
> On 15/01/16 13:07, Mark Janes wrote:
>> This patch causes piglit.spec.ext_framebuffer_multisample.accuracy to
>> fail on g33 and g965. Is that what you expected?
>
> No, I was not expecting any regression. It seemed a trivial change,
> ba
This patch causes piglit.spec.ext_framebuffer_multisample.accuracy to
fail on g33 and g965. Is that what you expected?
Alejandro Piñeiro writes:
> attach_color_texture uses TEXTURE_RECTANGLE, so it should check
> for GL_ARB_texture_rectangle.
>
> attach_multisample_color_texture uses multisampl
Reviewed-by: Mark Janes
Daniel Vetter writes:
> In
>
> commit 5cbc1834fd47e7f475afc85e1638762e5b221a81
> Author: Dylan Baker
> Date: Mon Dec 14 15:34:11 2015 -0800
>
> framework/test/base.py: Handle fail cases for tests.
>
> the Test baseclass was made
Reviewed-by: Mark Janes
baker.dyla...@gmail.com writes:
> From: Dylan Baker
>
> I'm going to admit I'm a bit puzzled how this could have slipped through
> without being caught (I'm guessing an unrelated change uncovered this).
> But basically if a test doesn
There are a few things with this patch that complicate my use of junit:
1) the stdout for the parent test is in the stored in the testsuite tag
in the junit xml. However, I can't access this data from the
jenkins ui. If a subtest fails, I won't be able to write up a bug
from the Jen
Reviewed-by: Mark Janes
Tapani Palli writes:
> Signed-off-by: Tapani Pälli
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93093
> ---
> tests/spec/arb_draw_indirect/draw-arrays-shared-binding.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/tests/s
Emil Velikov writes:
> On 25 September 2015 at 16:40, Mark Janes wrote:
>> With the push of SSBO support in mesa, one of these tests continues to
>> fail, but only on broadwell:
>>
>> piglit.spec.arb_shader_storage_buffer_object.array-ssbo-binding
>>
>>
ames
The transformation is not reversible, eg:
test/foo.bar/baz -> test.foo.bar.baz -> test/foo/bar/baz
This doesn't affect my use case, though.
Reviewed-by: Mark Janes
> @@ -264,6 +267,15 @@ def _load(results_file):
> result.command = out[0]
> resu
Reviewed-by: Mark Janes
baker.dyla...@gmail.com writes:
> From: Dylan Baker
>
> This calculates the group totals on load (this is stored in the json),
> which is required to get proper totals information.
> ---
> framework/backends/junit.py | 2 ++
Jose Fonseca writes:
> On 09/10/15 01:21, baker.dyla...@gmail.com wrote:
>> This series updates the junit backend to allow it to properly load junit
>> and convert it back into piglit's internal representation, thus allowing
>> it to be summarized using the piglit summary tools
>>
>> There is sti
Dylan Baker writes:
> On Thu, Oct 8, 2015 at 6:09 PM, Mark Janes wrote:
>
>> baker.dyla...@gmail.com writes:
>>
>> > This series updates the junit backend to allow it to properly load junit
>> > and convert it back into piglit's internal representatio
baker.dyla...@gmail.com writes:
> This series updates the junit backend to allow it to properly load junit
> and convert it back into piglit's internal representation, thus allowing
> it to be summarized using the piglit summary tools
>
> There is still some work that needs to be done beyond this,
Series is
Reviewed-by: Mark Janes
dylanx.c.ba...@intel.com writes:
> From: Dylan Baker
>
> This class will provide a new interface for time, one that stores the
> start and stop times, rather than the total time. It then provides
> methods for getting the total and the delta
Reviewed-by: Mark Janes
Tested-by: Mark Janes
Ian Romanick writes:
> From: Ian Romanick
>
> Section 4.3 (Storage Qualifiers) of the OpenGL ES 1.00.17 spec says:
>
> "Declarations of globals without a storage qualifier, or with just
> the const qualifier, may i
patch 1 is
Reviewed-by: Mark Janes
Tested-by: Mark Janes
Dylan Baker writes:
> Like command, this somehow didn't get moved to the new TestResult class.
> Generally this isn't a problem since this is a very uncommon path, but
> in some cases this does get hit, and it would
Reviewed-by: Mark Janes
Dylan Baker writes:
> This adds support for running Khronos' deqp-based conformance suite with
> piglit. This gives access to all of the powerful features of pigilt,
> per-process tests, the junit and json backends, the summary tools, and
> the familia
Emil Velikov writes:
> On 25 September 2015 at 16:40, Mark Janes wrote:
>> With the push of SSBO support in mesa, one of these tests continues to
>> fail, but only on broadwell:
>>
>> piglit.spec.arb_shader_storage_buffer_object.array-ssbo-binding
A subsequent run o
With the push of SSBO support in mesa, one of these tests continues to
fail, but only on broadwell:
piglit.spec.arb_shader_storage_buffer_object.array-ssbo-binding
Standard Output
/tmp/build_root/m64/lib/piglit/bin/arb_shader_storage_buffer_object-array-ssbo-binding
-auto -fbo
pigl
Tested-by: Mark Janes
Timothy Arceri writes:
> Samples must be greater than 0 and color formats must match.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91670
> ---
> tests/spec/gl-3.2/layered-rendering/framebuffertexture.c | 4 ++--
> 1 file changed, 2 insertio
Tested-by: Mark Janes
Tapani Pälli writes:
> Move IS_PER_PATCH as part of tessellation queries, require
> GL_ARB_compute_shader for compute shader enum queries.
>
> Strictly speaking, spec would allow to query GL_REFERENCED_BY_COMPUTE_SHADER
> without having compute sh
Tested-by: Mark Janes
Francisco Jerez writes:
> Fixes a rounding instability that would cause shader_runner to probe
> pixels offset by one for some points close to the right edge of the
> window on systems using x87 floating point arithmetic with certain
> compiler versions (the
JUnit has no concept of "warning". It supports the following
statuses:
- skip
- success
- fail
- error
A test which emits a warning is more accurately represented as
"success" in JUnit.
v2: Continue to report failure for "dmesg-warn", which is more serious
than "warn". (from Ilia Mirki
Ilia Mirkin writes:
> On Fri, Jun 26, 2015 at 2:44 PM, Dylan Baker wrote:
>>> That's fine, but then that's the justification, not "junit interprets
>>> warn as fail". Note that some (regular) piglit tests will also emit
>>> "warn", and they don't mean "fail". I believe that junit needs to be
>>>
45, G965:
!opengl 1_1.teximage-colors gl_r32f
!opengl 1_1.teximage-colors gl_rg32f
!opengl 1_1.teximage-colors gl_rgb32f
!opengl 1_1.teximage-colors gl_rgba32f
No output was provided for these warnings, so they may have have
previously been dmesg-warn.
>
> On Fri, Jun 26, 2015 at 1:40
JUnit has no concept of "warning". It supports the following
statuses:
- skip
- success
- fail
- error
dEQP has been found to intermittently emit warnings for passed tests,
and this status is accurately represented in piglit json. However,
current JUnit transforms them into failures.
A tes
Ilia Mirkin writes:
> On Thu, Jun 25, 2015 at 8:52 PM, Mark Janes wrote:
>> Many tests in the functional.fbo.completeness.attachment_combinations
>> report "QualityWarning" intermittently. This particular warning is
>> emitted because dEQP attempted to cr
Many tests in the functional.fbo.completeness.attachment_combinations
report "QualityWarning" intermittently. This particular warning is
emitted because dEQP attempted to create an unsupported framebuffer.
"warning" is interpreted as "failure" in junit, but this output from
dEQP is clearly not a
Reviewed-by: Mark Janes
Dylan Baker writes:
> Currently we blindly pass deqp extra args to each test case, and only to
> the test case.
>
> There are two cases where this is problematic:
> 1) if the extra args need to be passed to the deqp command that
>generates the t
Tested-by: Mark Janes
Timothy Arceri writes:
> In GLES vertex shader outputs cant be arrays of structs or structs that
> contain arrays or structs.
>
> The GLSL ES 3.0 spec says:
>
> "Vertex output variables output per-vertex data and are declared
> using the ou
Tested-by: Mark Janes
Chad Versace writes:
> The following commit regressed lookup of some new OpenGL symbols:
>
> commit cb334db3f2bb7ddf9b2f43641f898896f99c69b1
> Author: Daniel Kurtz
> Date: Thu Oct 16 21:36:01 2014 +0800
> Subject: dispatch: Use dl
Tested-by: Mark Janes
Andres Gomez writes:
> Old glsl-deriv-varyings test has been migrated to
> shader_runner and, also, splitted in four
> different tests:
> * glsl-derivs-varyings
> * glsl-derivs-sign
> * glsl-derivs-abs
> * glsl-derivs-abs-sign
>
> In addition,
Reviewed-by: Mark Janes
Jose Fonseca writes:
> If `api` appears in the middle of the test classname, like it happens
> with all `spec/glsl-1.20/api/getactiveattrib *` tests, then Jenkins will
> intercept it making it impossible to see individual tests results.
> Therefore, j
Adding the list.
Mark Janes writes:
> Hi Chad,
>
> This patch was recently pushed, and it caused piglit failures for
> arb_pipeline_statistics_query:
>
>
> Standard Output
>
> /tmp/build_root/m32/lib/piglit/bin/arb_pipeline_statistics_q
Series is
Reviewed-by: Mark Janes
Dylan Baker writes:
> One of the more interesting problems for piglit, especially with
> hardware bring-up where bad code may result in the system crashing, is
> to know what test actually crashed the system.
>
> This series is meant to addre
Reviewed-by: Mark Janes
Timothy Arceri writes:
> Test results:
>
> AMD Radeon HD 6670 - Catalyst 14.501.1003 OpenGL 4.4
>
> interface-block-instance-name-input-read-only.frag - pass
> interface-block-input-read-only.frag - pass
> interface-block-instance-name-uniform-
Reviewed-by: Mark Janes
Dylan Baker writes:
> This fixes several tests on platforms that support
> arb_transform_feedback3.
>
> Signed-off-by: Dylan Baker
> ---
> tests/all.py | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tests/all.py b/t
series Reviewed-by: Mark Janes
Matt Turner writes:
> By flattening the if statements we cut 288 instructions from the
> fragment shader (806 -> 518, or 35%), and we remove all register
> spilling on i965.
>
> Reduces runtime of "varying-packing-simple int separate"
distribution,
> and fail when performance goes below a given percentile.
I'm curious to know which benchmarks you use. We experience variability
in a lot of benchmarks, especially if they are cpu-heavy and running on
an under-powered system. It would be great to have workloads that
produc
over 350k unskipped tests for each
build. We filter out skipped tests due to the memory consumption on
jenkins when displaying this many tests.
I am interested in learning more about your test system, and sharing
lessons learned / techniques.
-Mark
Reviewed-by: Mark Janes
Jose Fonseca
series is
Reviewed-by: Mark Janes
Timothy Arceri writes:
> Uniform interface blocks can have different instance names across shaders.
> These tests check that array rules are applied correctly.
>
> Test results:
>
> Intel Ivybridge - Mesa 10.6
>
> uniform-block-arra
k_mixed_flatten(tests, testlist):
> nt.assert_dict_equal(profile_.test_list, baseline)
>
>
> -def generate_prepare_test_list_test_test_matches():
> -""" Generate tests for TestProfile.perpare_test_list filtering """
typo:
In junit.py, the testname variable is used by a closure within the
write_test() scope. Modifying it breaks the filtering of expected
failures.
---
framework/backends/junit.py | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/framework/backends/junit.py b/framework/backe
Jose Fonseca writes:
> On 20/02/15 20:50, Mark Janes wrote:
>> We support using the test names as displayed in jenkins to start bisect
>> jobs. If a test fails, you can copy/paste the name into a job, and it
>> will find the commit that broke it.
>>
>> With this
We support using the test names as displayed in jenkins to start bisect
jobs. If a test fails, you can copy/paste the name into a job, and it
will find the commit that broke it.
With this change, special-cased test names can't be used with
--include-tests anymore.
Instead of adding an underscore
Hi Eduardo,
I track piglit regressions for the mesa team at Intel. It would really
help me if you put a note in your commit message when a new test is
known to fail on any platform.
thanks,
Mark
Eduardo Lima Mitev writes:
> As a heads-up, with this patch piglit fails the test for current Mes
${Waffle_LDFLAGS}
> )
> IF(PIGLIT_BUILD_GLX_TESTS)
> list(APPEND UTIL_GL_LIBS
> --
> 2.2.1
>
> ___
> Piglit mailing list
> Piglit@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/piglit
Tested-by: Mark Janes
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
Tested-by: Mark Janes
Jason Ekstrand writes:
> Since GL_ABGR_EXT was extension number 1 to the GL spec, it didn't take
> packed formats into account. As far as I can tell from the way the packed
> formats extensions are written, packed formats with GL_ABGR_EXT isn't
&g
n(
> 'glsl-{}'.format(version), 'execution', 'texelFetchOffset',
> '{}-{}'.format(stage, sampler))] = PiglitGLTest(
> -['texelFetch', 'offset,' '140', stage, sampler],
> +
Reviewed-by: Mark Janes
Dylan Baker writes:
> bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=87513
> Signed-off-by: Dylan Baker
> ---
> generated_tests/generate-cl-store-tests.py | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/genera
gt; module.
>
> v2: - remove accidentally included hunk (Mark)
> v3: - Actually include the modules directory in the commit (Mark)
Reviewed-by: Mark Janes
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
It looks like this patch omits the utils.py module which is being
called.
This would have been hard to spot in a build-test, because the pyc from
your previous patch was still in place.
-Mark
Dylan Baker writes:
> Patch b59ff71eb was supposed to fix os.makedirs exceptions, but falls
> short bec
It seems like you unintentionally included an extra file:
generated_tests/genclbuiltins.py
The deltas are not related to os.makedirs.
Dylan Baker writes:
> generated_tests/genclbuiltins.py | 54
> +-
> diff --git a/generated_tests/genclbuiltins.py
> b/ge
Thanks, this is a superior implementation imho.
-Mark
Dylan Baker writes:
> +
> +def safe_makedirs(dirs):
> +"""A safe function for creating a directory tree.
> +
> +This function wraps os.makedirs, and provides a couple of sanity checks,
> +first, if the directory already exists it
Thanks for making this fix.
Dylan Baker writes:
> -os.makedirs(dirname)
> +try:
> +os.makedirs(dirname)
> +except OSError as e:
> +if e.errno == 17: # file exists
> +pass
> +raise
This works
> v2: Try to handle both gl and gles*.
>
> Cc: Mark Janes
> Reported-by: Mark Janes
> Signed-off-by: Emil Velikov
> ---
> tests/util/CMakeLists.txt | 5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/tests/util/CMakeLists.txt b/tests/util/CMakeLists.txt
> ind
instruction that should be used,
instead of linking the gles tests against GL?
-Mark
Emil Velikov writes:
> Similar to the earlier patch for waffle-less (glut) builds. The need
> for this patch became apparent as we removed the libGL link dependency
> in waffle.
>
> Cc: Mark Janes
>
Dylan Baker writes:
> On Friday, November 21, 2014 07:37:20 PM Jose Fonseca wrote:
>> On 20/11/14 12:06, jfons...@vmware.com wrote:
>> > From: José Fonseca
>> >
>> > Showing the command line of the test is quite useful, specially when
>> > diagnosing failures.
>
> If Marks okay with it.
I like
Ben Widawsky writes:
> On Tue, Nov 18, 2014 at 02:48:09PM -0800, Mark Janes wrote:
>> Ilia Mirkin writes:
>>
>> > FWIW, on nv50 and nvc0, EXT_timer_query time-elapsed and
>> > ARB_timer_query query GL_TIMESTAMP reliably fail, and the rest
>> > r
Ilia Mirkin writes:
> FWIW, on nv50 and nvc0, EXT_timer_query time-elapsed and
> ARB_timer_query query GL_TIMESTAMP reliably fail, and the rest
> reliably pass. I don't remember ever seeing inconsistent behaviour.
> FWIW I don't really know what these extensions do, or what the tests
> check for,
EXT_timer_query and ARB_timer_query tests fail intermittently, causing
confusion for developers running quick.py to find regressions. These
tests have always been intermittent, and people generally know to
ignore them when they fail.
However, if everyone ignores a test, there is no point in runni
1 - 100 of 107 matches
Mail list logo
