Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/test_monitoring.py | 178 +++
unittests/monitoring_tests.py | 186 -
2 files changed, 178 insertions(+), 186 deletions(-)
create mode
.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/genclbuiltins.py | 80
1 file changed, 39 insertions(+), 41 deletions(-)
diff --git a/generated_tests/genclbuiltins.py b/generated_tests/genclbuiltins.py
index f907ee8..338914c
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
.../summary/test_console.py} | 135 +
1 file changed, 55 insertions(+), 80 deletions(-)
rename unittests/{summary_console_tests.py =>
framework/summary/test_console.py} (56%)
d
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/generators/test_glsl.py | 319 +-
1 file changed, 111 insertions(+), 208 deletions(-)
diff --git a/unittests/generators/test_glsl.py
b/unittests/generators/test_glsl.py
index 84802c2..6
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/summary/__init__.py| 0
unittests/framework/summary/test_common.py | 447 +
unittests/summary_common_tests.py | 424 ---
3 files changed, 447 inse
These tests weren't good tests at all, they didn't test useful
functionality.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/run_parser_tests.py | 264 --
1 file changed, 264 deletions(-)
delete mode 100644 uni
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
.../summary/test_html.py} | 21 +
1 file changed, 5 insertions(+), 16 deletions(-)
rename unittests/{summary_html_tests.py => framework/summary/test_html.py}
(73%)
diff --git a/
This adds a significant number of comments to the tests, it also makes a
very small change to the various dmesg classes, it adds a __repr__
method to each of them to make testing the classes easier.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
framework/dmesg.py
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/test_options.py | 221
unittests/options_tests.py | 208 -
2 files changed, 221 insertions(+), 208 deletions(-)
create mode 100644 uni
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/test/test_piglit_test.py | 164 +++
unittests/piglit_test_tests.py | 140 ---
2 files changed, 164 insertions(+), 140 deletions(-)
create mode 100644 uni
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
.../{integration_tests.py => suites/test_integration.py} | 15 +++
1 file changed, 7 insertions(+), 8 deletions(-)
rename unittests/{integration_tests.py => suites/test_integration.py} (91%)
diff --git
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
tox.ini | 1 +
unittests/base_tests.py | 438 --
unittests/framework/test/test_base.py | 412
3 files changed, 413 inse
version to 1.50, like the
shader_image_load_store generator did.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/gen_extensions_defined.py| 2 +-
generated_tests/gen_shader_image_load_store_tests.py | 17 -
2 files changed, 1 insertion(
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/deqp_tests.py | 282
unittests/framework/test/test_deqp.py | 296 ++
2 files changed, 296 insertions(+), 282 deletions(-)
delete mode
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/test_profile.py | 370 +++
unittests/framework/utils.py| 45 +
unittests/profile_tests.py | 376
3 files changed, 415 inse
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
tox.ini | 5 +-
unittests/core_tests.py | 289 ---
unittests/framework/skip.py | 47 +++
unittests/framework/test_core.py
This port uses a lot less code, is much simpler, and is able to mark
tests as slow or very_slow (which is unfortunately necessary).
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
tox.ini | 24 +++---
unittests/generators/test_generators.py
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
tox.ini| 4 +-
unittests/suites/__init__.py | 0
unittests/{test_lists.py => suites/test_native.py} | 48 +++---
3 files changed, 26 insertio
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/test_log.py | 193
unittests/log_tests.py | 171 ---
2 files changed, 193 insertions(+), 171 deletions(-)
create mode 100644 uni
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/{opencv_tests.py => framework/test/test_opencv.py} | 8 +++-
1 file changed, 3 insertions(+), 5 deletions(-)
rename unittests/{opencv_tests.py => framework/test/test_opencv.py} (89%)
diff --git a/unittests/openc
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/test/test_shader_test.py | 194 +
unittests/shader_test_tests.py | 205 ---
2 files changed, 194 insertions(+), 205 deletions(-)
create mode
This patch adds a couple of new tests that were not being generated for
some reason, and fixes a test that wasn't testing what it was supposed
to. These tests fail of course, so they've been marked as expected fail
and will be fixed later.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
.../test/test_gleantest.py}| 67 ++
1 file changed, 31 insertions(+), 36 deletions(-)
rename unittests/{gleantest_tests.py => framework/test/test_gleantest.py} (70%)
diff --git a/
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/generators/test_types.py | 33 +++--
1 file changed, 11 insertions(+), 22 deletions(-)
diff --git a/unittests/generators/test_types.py
b/unittests/generators/test_types.py
index f60282c..3
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/framework/test_grouptools.py | 132 +
unittests/grouptools_tests.py | 130
2 files changed, 132 insertions(+), 130 deletions(-)
create mode
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
tox.ini| 11 +-
unittests/base_tests.py| 2 +-
unittests/framework/__init__.py| 0
unittests/framework/test_status.py | 139 +
unittests/status_tests.py
Acquiring the VM's to run the tests is the single most time consuming
part of the travis process, to reduce the amount of turn around combine
the noaccel and accel profiles, so that a test is run per python version
rather than per profile.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.
This is a) good practice, b) fixes potential issues on windows, and c)
makes testing easier.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
.../gen_shader_image_load_store_tests.py | 1076 ++--
1 file changed, 541 insertions(+), 535 deletions(-)
diff
, and provides much better testing
of the output.
This also fixes a number of bugs found in piglit as a result of the
better testing.
Dylan Baker (49):
travis: combine noaccel and accel profiles
generators: use main function in gen_shader_image_load_store_tests
Generators
, which when set to True causes a FatalError to be
raised if that option falls through to the default.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
framework/test/deqp.py | 6 +-
unittests/deqp_tests.py | 8
2 files changed, 13 insertions(+), 1 deletion(-)
diff
e_2d_array", "300 es"),
> +("GL_OES_texture_storage_multisample_2d_array", "310 es"),
> ("GL_OES_blend_func_extended", "100"),
> ("GL_OES_shader_image_atomic", "310 es"),
> ("GL_OES_shader_io
There is no implementation of the code these tests exercise on windows,
so they shouldn't run, these were masked by the bad skip conditions
fixed in 701df09b7510aca2a010ad9eaff2e0258e4ea4bc.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/base_tests.py | 3 +++
1 file c
This fixes int64 failures.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/gen_constant_array_size_tests.py | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/generated_tests/gen_constant_array_size_tests.py
b/generated
Quoting Ian Romanick (2016-07-05 15:31:11)
> It looks like none of the compiler tests have GL_ARB_gpu_shader_int64 in
> the requirements section... so now I get ~400 spurious failures. :(
I have a patch that I think fixes this. I'm running it through Jenkins
now to be sure, I'll send it to the
# Find the config section
> for line in lines:
> --
> 2.8.1
>
> ___
> Piglit mailing list
> Piglit@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/pigl
Quoting Marek Olšák (2016-07-04 16:39:48)
> On Fri, May 27, 2016 at 7:53 PM, Dylan Baker <dy...@pnwbakers.com> wrote:
> > Quoting Marek Olšák (2016-04-16 15:16:34)
> >> Hi,
> >>
> >> This makes shader_runner very fast. The expected result is 40%
> &
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/junit_backends_tests.py | 1 -
1 file changed, 1 deletion(-)
diff --git a/unittests/junit_backends_tests.py
b/unittests/junit_backends_tests.py
index d5f753a..5b0bead 100644
--- a/unittests/junit_backends_tests.py
point_size", "310 es"),
> +("GL_EXT_geometry_point_size", "310 es"),
> ("GL_EXT_gpu_shader5", "310 es"),
> ("GL_OES_gpu_shader5", "310 es"),
> ("GL_EXT_texture_buffer", "310 es"),
> --
&
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/CMakeLists.txt | 2 ++
1 file changed, 2 insertions(+)
diff --git a/generated_tests/CMakeLists.txt b/generated_tests/CMakeLists.txt
index 7240572..c01ee72 100644
--- a/generated_tests/CMakeLists.txt
+++ b/generated
64s,
> + small_uints],
> + template='({0} << {1})',
> + extension="ARB_gpu_shader_int64")
> +f('op-rshift', 2, 150, _rshift, match_shift,
> + [int64s+uint64s,
> + small_uints],
> + template='({0} >> {1})',
> + extension=&quo
file mode 100644
> index 000..f6ea2e3
> --- /dev/null
> +++ b/generated_tests/templates/gen_conversion_int64/execution_base.mako
> @@ -0,0 +1,31 @@
> +## coding=utf-8
> +<%inherit file="base.mako"/>\
> +<%
> +(glsl_version, glsl_version_int)
Reviewed-by: Dylan Baker <dy...@pnwbakers.com>
Maybe a good future change would be to pull these "good security
practices, but not required by the spec" tests into a separate profile
and report fail?
Quoting Marek Olšák (2016-06-29 02:16:16)
> From: Marek Olšák &
.50
> > GL_ARB_tessellation_shader
> >
> ___
> Piglit mailing list
> Piglit@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/piglit
Reviewed-by: Dylan Baker <dy...@pnwbakers.com>
I've also gone ahead and pushed.
signature.asc
Description: signature
___
Piglit mailing list
Piglit@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/piglit
Quoting Matt Turner (2016-06-17 13:24:02)
> All four are
>
> Reviewed-by: Matt Turner
Thanks Matt.
signature.asc
Description: signature
___
Piglit mailing list
Piglit@lists.freedesktop.org
This spec actually wants 4.4, plus the other two ES compat extensions.
Mesa chooses to ignore this, so set it to 4.3 (which allows most mesa
drivers to test it.)
At the very least though this should be bumped to 3.2, since this
extension is never valid in a GL < 3.2 context.
Signed-off-by: Dy
This stops the generator from creating bogus tests.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/gen_extensions_defined.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/generated_tests/gen_extensions_defined.py
b/generated
While this actually depends on 3.0 or ext_texture_array, this sets the
GLSL version requirement to 1.30, which is the GLSL version that
corresponds to GL 3.0.
This stops the generator from creating bogus tests.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated
This fixes the version requirement to be accurate, which prevents the
generator from producing bogus tests.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/gen_extensions_defined.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/generated
Quoting Kenneth Graunke (2016-06-16 16:04:18)
> On Thursday, May 5, 2016 5:48:40 PM PDT Dylan Baker wrote:
> > Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
> > ---
> > tests/all.py | 5 -
> > te
Quoting Andres Gomez (2016-06-13 09:32:17)
> On Fri, 2016-06-10 at 16:42 -0700, Dylan Baker wrote:
>
> [snip]
>
> >
> > I've had this plan to add additional helpers for generator writers
> > for
> > representing GLSL type information, and this seemed lik
Quoting Andres Gomez (2016-06-13 09:35:29)
> On Fri, 2016-06-10 at 16:42 -0700, Dylan Baker wrote:
>
> [snip]
>
> > +class Scalar(Container):
> > +"""A base class for Scalar types."""
> > +def __init__(self, name, contains):
&g
-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/modules/types.py | 199 +
unittests/generators/test_types.py | 79 +++
2 files changed, 278 insertions(+)
create mode 100644 generated_tests/modules/types.py
create mode
within mako and is part of the standard library, we can't
optimize it) and the regular.shader_test.mako. While I can spot some
micro optimizations that might get us a few tenths of a second here and
there, I don't see anything that's going to win another 10 or 15
seconds.
Signed-off-by: Dylan Baker
). I think that's a pretty impressive gain.
Andres, I saw your series that extends this generator. Do you think it's
better to consider landing this first and rebase your series ontop of
this, or land yours first and rebase this series?
Dylan Baker (2):
generators/modules: Add GLSL types
Quoting Dave Airlie (2016-06-09 22:27:40)
> On 10 June 2016 at 13:40, Dave Airlie wrote:
> >> No, you can't use int and long here, python 3 doesn't have a long type
> >> at all, and it's int will grow until it uses all of your memory if you
> >> ask it to, while python2 has an
Quoting Martin Peres (2016-06-10 03:24:51)
> On 09/06/16 19:50, Dylan Baker wrote:
> > Quoting Andres Gomez (2016-06-09 06:45:49)
> >> Nice optimizations, Dylan.
> >>
> >> I'm quite impressed/surprised on the gain through avoiding the
> >> inherit
Quoting Dave Airlie (2016-06-08 19:50:14)
> From: Dave Airlie
[snip]
> +# This source file defines a set of test vectors that can be used to
> +# test GLSL's built-in functions and operators. It is intended to be
> +# used by Python code that generates Piglit tests.
> +#
>
Quoting Andres Gomez (2016-06-09 06:45:49)
> Nice optimizations, Dylan.
>
> I'm quite impressed/surprised on the gain through avoiding the
> inheritance/include mechanisms in the mako templates.
>
> This series is:
>
> Reviewed-by: Andres Gomez
>
> Andres
I was definitely
bump
signature.asc
Description: signature
___
Piglit mailing list
Piglit@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/piglit
Don't build a list of filenames and print them after all of the tests
have run, act more like the other generators and print each name as the
test is generated. This both matches the behavior of the other
generators and also lowers runtime a bit
Signed-off-by: Dylan Baker <dylanx.c
think the new template is more readable than the original code,
and it is definitely faster, it does introduce more code duplication
that could be cleaned up in a follow on patch.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/CMakeLists.txt
Saves a little bit more runtime. It's not an extreme amount, but more
importantly is should save quite a bit of memory, which may help users
with memory-constraints.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/gen_vs_in_fp64.p
the new, faster templates later in this series
working.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
tests/shaders/shader_runner.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/tests/shaders/shader_runner.c b/tests/shaders/shader_runner.c
index 94c7826..5b1acb0
this is
saving us a little bit of time, it's not going to make a big difference.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
generated_tests/CMakeLists.txt | 1 +
generated_tests/gen_vs_in_fp64.py | 2 +-
.../gen_vs_i
for performance.
Dylan Baker (6):
generators/gen_vs_in_fp64: don't list-ize generators for no reason.
generators/gen_vs_in_fp64: don't store the filenames print them
immediately
generators/gen_vs_in_fp64: use more generators/iterators save more
runtime
shader_runner: Allow whitespace before
.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
framework/test/piglit_test.py | 20 +++-
1 file changed, 11 insertions(+), 9 deletions(-)
diff --git a/framework/test/piglit_test.py b/framework/test/piglit_test.py
index 32a991b..ddef66b 100644
--- a/framewor
---
unittests/backends_tests.py | 16 +--
unittests/base_tests.py | 35 +++---
unittests/compressed_backend_tests.py | 54 -
unittests/core_tests.py | 30 ++---
unittests/deqp_tests.py | 26 ++---
unittests/dmesg_tests.py
Quoting Brian Paul (2016-06-01 17:10:30)
> On 06/01/2016 05:50 PM, Dylan Baker wrote:
> > I'm tired of breaking piglit on windows. This series add appveyor (and
> > travis too, why not get free Linux testing while we're at it?) support
> > to piglit for the python framew
Which works just fine on Linux and doesn't on Windows.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/core_tests.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/unittests/core_tests.py b/unittests/core_tests.py
index 0e55ca3..7cb6745
Fixes running on windows.
cc: Brian Paul <bri...@vmware.com>
cc: Olivier Berthier <olivierx.berth...@linux.intel.com>
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
This is not the "right" solution, but it should get everyone up and
running again. Then we
This fixes a number of tests on windows, where being in a directory
prevents it from being deleted.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/core_tests.py | 18 ++
unittests/junit_backends_tests.py | 12 +---
unittests/utils/n
These fail because they assumes that an xz binary is available, which is
probably not the case on windows. This splits the binary and module
tests, and makes the skipping actually work in all of the cases it needs
to (or at least more of the cases).
Signed-off-by: Dylan Baker <dylanx.c
I prefer that skip is just a decorator so that the test itself doesn't
need to contain code to skip.
---
tox.ini | 1 +
unittests/base_tests.py | 31 +++-
unittests/compressed_backend_tests.py | 14 +---
unittests/core_tests.py |
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/compressed_backend_tests.py | 1 +
1 file changed, 1 insertion(+)
diff --git a/unittests/compressed_backend_tests.py
b/unittests/compressed_backend_tests.py
index 8ff2b0c..80957e7 100644
--- a/uni
it gets everybody up and
running again without breaking the monitoring users on POSIX systems.
This is available at my github:
https://github.com/dcbaker/piglit wip/appveyor-v2
Dylan Baker (20):
travis: Add travis.yml for python code
appveyor.yml: Add appveyor for python code
tox: split py27
This relies on and tests a linux (or maybe posix?) only feature, so skip
on windows.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/base_tests.py | 1 +
1 file changed, 1 insertion(+)
diff --git a/unittests/base_tests.py b/unittests/base_tests.py
index 2058703..2
The generator will otherwise hold the tempfile opened and cause an error
in the test.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/deqp_tests.py | 2 ++
1 file changed, 2 insertions(+)
diff --git a/unittests/deqp_tests.py b/unittests/deqp_tests.py
index c237b81..7
Some parts of the accel profile don't work on windows (or don't work
easily) such as lxml and subprocess32. Because of this create two
separate acceel profiles, one for windows and one for posix platforms.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
tox.ini | 15
the right thing.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/dmesg_tests.py | 27 +--
1 file changed, 9 insertions(+), 18 deletions(-)
diff --git a/unittests/dmesg_tests.py b/unittests/dmesg_tests.py
index 0b0490d..18fa361 100644
--- a/uni
Subprocess from Python 2 on windows is an interesting beast. The way
it's implemented is a little strange, and passing environment variables
as unicode breaks things. It works fine on POSIX, however. So for
windows convert to binary.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.
Tildes are not expanded as $HOME on windows, so don't run that test on
windows.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/core_tests.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/unittests/core_tests.py b/unittests/core_tests.py
index 9
---
unittests/compressed_backend_tests.py | 20
unittests/deqp_tests.py | 12 ++--
unittests/utils/nose.py | 35 ---
3 files changed, 18 insertions(+), 49 deletions(-)
diff --git
Mock solves this problem nicely and removes code.
---
unittests/core_tests.py | 59 +
1 file changed, 6 insertions(+), 53 deletions(-)
diff --git a/unittests/core_tests.py b/unittests/core_tests.py
index 8a160e7..341663d 100644
---
-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
appveyor.yml | 55
ci/appveyor/compiler.cmd | 46
2 files changed, 101 insertions(+)
create mode 100644 appveyor.yml
create mode 100644 ci/ap
Using this context manager ensures that you change back to the original
directory when leaving the context manager
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
unittests/utils/nose.py | 16
1 file changed, 16 insertions(+)
diff --git a/unittests/utils/nos
For testing the framework with travis
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
.travis.yml | 16
1 file changed, 16 insertions(+)
create mode 100644 .travis.yml
diff --git a/.travis.yml b/.travis.yml
new file mode 100644
index 000..8532d47
--- /de
Quoting Brian Paul (2016-06-01 16:21:07)
> On 05/31/2016 02:45 PM, Dylan Baker wrote:
> > Quoting Olivier Berthier (2016-05-30 03:21:14)
> >> This patch disables the fnctl import on Windows.
> >> ---
> >
Quoting Olivier Berthier (2016-05-30 03:21:14)
[snip]
> +if sys.platform.startswith('linux'):
> +try:
> +with open(self._monitoring_source, 'r') as f:
> +lines = []
> +if self._is_locked:
> +#
char gs_source[] =
> "/* At least some versions of AMD's closed-source driver\n"
> --
> 2.8.3
>
> ___
> Piglit mailing list
> Piglit@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/list
Quoting Olivier Berthier (2016-05-30 03:21:14)
> This patch disables the fnctl import on Windows.
> ---
> framework/monitoring.py | 97
> ++---
> 1 file changed, 52 insertions(+), 45 deletions(-)
>
> diff --git a/framework/monitoring.py
Quoting Brian Paul (2016-05-27 17:10:45)
> The recent commit
>
> commit 21031482e7a9fc828622ea2015d87432704b1c7f
> Author: Olivier Berthier
> Date: Mon May 2 13:05:17 2016 +0200
>
> framework: Add aborting option when a monitored error is detected
>
>
tively.
>
> v2 (Dylan Baker):
> - Use cts_gles and cts_gl
> - Rename PIGLIT_CTS_foo to PIGLIT_CTS_GLES_foo
> ---
> piglit.conf.example | 18 ++
> tests/{cts.py => cts_gl.py} | 43
> ++-
> tests/{ct
Quoting Pohjolainen, Topi (2016-05-19 12:36:08)
> On Fri, May 06, 2016 at 01:40:15PM -0700, Dylan Baker wrote:
> > Quoting Pohjolainen, Topi (2016-05-04 10:25:31)
> > [sni[
> > > Okay, I don't how I mis-read the original tests so badly - CLAMP01 is the
> > > par
Quoting Emil Velikov (2016-05-19 02:52:35)
> On 19 May 2016 at 07:33, Juan A. Suarez Romero <jasua...@igalia.com> wrote:
> > On Wed, 2016-05-18 at 11:00 -0700, Dylan Baker wrote:
> >> I'm 50/50 on it. While I think it probably is nicer to change them,
> >>
Quoting Juan A. Suarez Romero (2016-05-18 09:27:34)
> On Wed, 2016-05-18 at 09:20 -0700, Dylan Baker wrote:
> > Quoting Juan A. Suarez Romero (2016-05-18 06:34:07)
> > > This adds support for running Khronos' deqp-based conformance suite
> > for
bump
signature.asc
Description: signature
___
Piglit mailing list
Piglit@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/piglit
clCreatProgramWithBinary expects a "const unsigned char **", but the code
passes an "unsighed char **". This is a particular kind of messiness
that is hard to fix, so just cast when passing into
clCreateProgramWithBinary.
Thanks to idr and Matt for help getting this right.
Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com>
---
v3 - Use OpenCL deprecation macros (EdB)
tests/util/piglit-util-cl-enum.h | 5 +
1 file changed, 5 insertions(+)
diff --git a/tests/util/piglit-util-cl-enum.h b/tests/util/piglit-util-cl-enum.h
index 51336ce..e08d3c3
as invariant."
> +#
> +[require]
> +GL ES >= 2.0
> +GLSL ES >= 1.00
> +
> +[vertex shader]
> +invariant gl_PointSize;
> +void main() {
> + gl_PointSize = 1.0;
> + gl_Position = vec4(0.0);
> +}
> +
> +[fragment shader]
> +invariant gl_Po
n specification:
> "In general, correct signedness of 0 is not required."
>
> To avoid this problem, remove the sign bit from both the results
> and expected values when the value is equal to (+/-) 0.0.
>
> v2: Use xs.dtype.type(0.0) (suggested by Dylan Baker)
801 - 900 of 3305 matches
Mail list logo