| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Khronos removed GL_TEXTURE_BINDING, an enum introduced in
GL_ARB_direct_state_access that was ruled unimplementable.
|
|
arb_direct_state_access-compressedtextureimage should use a more mainstream
compression format in all.py. This patch changes it to
GL_COMPRESSED_RGBA_S3TC_DXT5_EXT.
|
|
|
|
|
|
|
|
|
|
|
|
v2: Added increased error checking
Review from Ian Romanick and Ilia Mirkin:
- Compat context 20 (and remove check for ARB_texture_cube_map)
- Change global_pass from GL_boolean to bool
- Clean up the texture in upload_subtest
- Restructure download commands to simplify use_pbo logic
- Use separate pack and unpack pbos
I have not applied the pre-emptive R-b's because this patch series adds a new
getteximage-targets test, and, because the tests are so similar, it makes
sense to have the whole thing reviewed and pushed as a big chunk.
|
|
|
|
v2: Added increased error checking
Review from Ian Romanick and Ilia Mirkin:
- Compat context 20 (and remove check for ARB_texture_cube_map)
- Change global_pass from GL_boolean to bool
- Clean up the texture in upload_subtest
- Restructure download commands to simplify use_pbo logic
- Use separate pack and unpack pbos
I have not applied the pre-emptive R-b's because this patch series adds a new
getteximage-targets test, and, because the tests are so similar, it makes
sense to have the whole thing reviewed and pushed as a big chunk.
|
|
v2: Added increased error checking
Review from Ian Romanick and Ilia Mirkin:
- Compat context 20 (and remove check for ARB_texture_cube_map)
- Change global_pass from GL_boolean to bool
- Clean up the texture in upload_subtest
- Restructure download commands to simplify use_pbo logic
- Use separate pack and unpack pbos
I have not applied the pre-emptive R-b's because this patch series adds a new
getteximage-targets test, and, because the tests are so similar, it makes
sense to have the whole thing reviewed and pushed as a big chunk.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
GetNamedBufferPointerv.
|
|
|
|
|
|
|
|
|
|
Reviewed-by: Martin Peres <martin.peres@linux.intel.com>
|
|
Reviewed-by: Martin Peres <martin.peres@linux.intel.com>
|
|
Testing NamedBufferStorage, NamedBufferData, CopyNamedBufferSubData, and
MapNamedBufferRange.
|
|
|
|
|
|
glClientWaitSync(fence, GL_SYNC_FLUSH_COMMANDS_BIT, GL_TIMEOUT_IGNORED)
often returns GL_TIMEOUT_EXPIRED, even though GL_TIMEOUT_IGNORED is a huge
integer (18446744073709551615 on Intel Sandybridge with Mesa). The third
argument to glClientWaitSync should represent the number of nanoseconds the
implementation waits for the signal from fence before giving up and returning
TIMEOUT_EXPIRED.
This test, adapted from bufferstorage-persistent, demonstrates that
glClientWaitSync(fence, GL_SYNC_FLUSH_COMMANDS_BIT, GL_TIMEOUT_IGNORED)
must be called multiple times before the signal is actually recorded and
either GL_CONDITION_SATISFIED or GL_ALREADY_SIGNALED is returned.
|
|
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
I made a somewhat silly assumption when writing the group_manager, that
all Test classes would use list arguments. While this assumption is true
for tests that require an entire command as their first argument it
isn't true of GleanTest, where only the name of the subtest is passed.
In this case it will be a string, and if a name isn't passed as well
then name will end up being " ".join()'ed, ie, "api" -> "a p i".
This patch adds support to the group_manager to support this feature.
I considered two other approaches here. The first was to say screw
correctness, and just add a duplicate name to the handful of tests that
use the group_manager. The second was to change GleanTest to take a 1
element list. Both of these had downsides that I couldn't justify.
This fixes a handful of glean tests that use the group_manager (fbo,
api2, etc)
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Tested-by: Jose Fonseca <jfonseca@vmware.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
This is somewhat simpler way to handle this, since check_output monitors
only stdout, and hides stderr, and it gives us a lovely exception if the
returncode isn't 0, which allows us to completely bypass the error
checking code if there is no error.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
This makes igt more like the rest of piglit which uses the __future__
version of print, division, and import
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
In some cases this is really just about formatting.
v3: - remove accidental whitespace insertion (Ian)
v4: - Update the module docstring to be accurate after commit 443abf5
(Thomas)
v5: - Fix small typo (Thomas)
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
Make constants all caps, per PEP8, make functions all lowercase with
underscores, and move some toplevel work into helper functions.
v2: - drop one pylint disable line that wasn't needed (Thomas)
v5: - Fix replacment of IGTTestProfile with TestProfile.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
v2: - Add this patch to the series
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
Currently there is a function that in some cases could return '', but
the output is assumed to be a list. This patch ensures that it always
returns a list.
v3: - further simplify the function by using two return statements
instead of a single return and assignment
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
v2: - Also remove stderr from subprocess.Popen
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Thomas Wood <thomas.wood@intel.com>
|
|
This test is redundant with:
glsl-1.20/compiler/structure-and-array-operations/array-selection.vert
Both contain the same shader code and the same spec citation - the only
difference is variable names and the exact portion of the text quoted.
The other one has a better name, so keep it and delete condition-07.
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
|
|
The
del profile.test_list[key]
statement was silently failing because the TestDict's key lowering magic
was not happening for deleted items.
Unfortunately this still is not enough to fix the glean exclusions:
somehow all glean test names are being munged with extra whitespace.
For example:
$ ./piglit-print-commands.py tests/llvmpipe.py | grep '^glean/p'
warning: test glean/pointAtten does not exist
warning: test glean/texCombine does not exist
[...]
glean/p o i n t a t t e n ::: bin/glean -o -v -v -v -t +pointAtten --quick
Reviewed-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
This fixes several tests on platforms that support
arb_transform_feedback3.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Mark Janes <mark.a.janes@intel.com>
|
|
This corrects a copy and paste error in all.py that matches the wrong
binary with a test name. This causes an artificial piglit regression on
mesa, at least for the Intel driver. This patch fixes that typo. Trivial
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This is a bug (https://bitbucket.org/logilab/pylint/issue/461/incorrect-hanging-indent-detection-with)
that misdetects the hanging indent of the context manager. This silences
roughly 300 warnings.
v2: - also turn this warning off in quick.py
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
The previous commit made True the default value. This removes ~400 LOC
from all.py (about 10%)
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This will flip the value of the flag for PiglitGLTest, ShaderTest, and
GLSLParserTest. It will not affect PiglitCLTest (That's OpenCL), because
of the previous patch, which causes PiglitCLTest to use a unique
concurrency setting method.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This will allow us to flip the default value of run_concurrent to True
in the next patch. Since the majority of tests are concurrent safe (and
honestly many of them that are marked not safe are), this change will
help to reduce the number of tests that are marked serial by default by
copy-n-paste.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
