It is actually now recommendede not to use postal address at all
(http://www.gnu.org/licenses/gpl-howto.html).
Please, apply,
Matěj
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
Signed-off-by: Matěj Cepl mc...@redhat.com
---
tests/glslparsertest/glsl2/xonotic-vs-generic-diffuse.vert | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/tests/glslparsertest/glsl2/xonotic-vs-generic-diffuse.vert
I see requiring ARB_texture_storage is done inconsistently in
texture_view piglit tests that use TexStorage.
Thus, would either get a skip or a fail. A philosophical debate which
is better...I am okay either way.
Jon
On 03/25/2014 04:55 PM, Chris Forbes wrote:
On Wed, Mar 26, 2014 at 10:27
This function gets the default display for the given platform.
If the given platform platform is EGL_NONE, the the function wraps
eglGetDisplay(). Otherwise, it wraps eglGetPlatformDisplayEXT(). If
EGL does not support the platform extension for the given platform, then
it returns
Add new header piglit-log.h which exposes the following functions:
piglit_log_get_opt()
piglit_log_set_opt()
piglit_loge()
piglit_logi()
This new module should help to reduce redundant logging boilerplate
found in many Piglit files.
Signed-off-by: Chad Versace
Set PIGLIT_HAS_PTHREADS as a CMake variable and cpp feature macro if
CMake succeeds in compiling a small test file that uses pthread_self(),
Future EGL_KHR_fence_sync tests will be multi-threaded.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
CMakeLists.txt | 18
Without an explicit EGLDisplay parameter, it's not possible to call
piglit_require_egl_extension() without first binding the display to
a context with eglMakeCurrent(). Some EGL tests, though, will never need
to or want to create a context.
Also, this makes the signature of
Pre-patch, piglit-framework-gl.c:process_args() parsed subtest args
'-subtest' and '-list-subtests' with inline code. This patch moves that
code to a new function, piglit_parse_subtest_args().
This prepares for using Piglit's subtest magic in non-GL tests.
Currently, all the code for subtests
This prepares for using Piglit's subtest magic for non-GL tests.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
tests/util/piglit-framework-gl.c | 6 +++---
tests/util/piglit-framework-gl.h | 2 +-
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git
This prepares for using Piglit's subtest magic in non-GL tests.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
tests/spec/arb_compute_shader/api_errors.c | 2 +-
tests/util/piglit-framework-gl.c | 10 +-
tests/util/piglit-framework-gl.h | 12
Patches 1-6: The new EGL_KHR_fence_sync tests use Ian's new subtest mechanism.
These patches cleanup the subtest stuff so that EGL tests can use them.
Patches 7-9: Add EGL utility funcs to libpiglitutil for the new tests.
Patches 10-11: Add more utils for the new tests.
Patches 12-13: Finally,
Add the following subtests:
eglCreateSyncKHR_default_attributes
eglCreateSyncKHR_invalid_display
eglCreateSyncKHR_invalid_attrib_list
eglCreateSyncKHR_wrong_display_same_thread
eglCreateSyncKHR_with_display_bound_in_other_thread
eglCreateSyncKHR_invalid_sync_type
If the user specifies -subtest $NAME on the command line, but
the test defines no subtest with that name, then
piglit_parse_subtest_args() now reports PIGLIT_FAIL.
The test should fail early as possible on usage errors. Pre-patch, if
the user requested an invalid subtest name, the test did not
Please. Let's stop writing `!strcmp(...)` and `strcmp(...) == 0`. This
patch defines a little function streq() that reads much more nicely than
idioms using strcmp.
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
tests/util/piglit-util.c | 6 ++
tests/util/piglit-util.h | 6
I forgot to say... this series lives on a personal branch:
git://people.freedesktop.org/~chadversary/piglit EGL_KHR_fence_sync
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
On Wed, Mar 26, 2014 at 10:40 AM, Chad Versace
chad.vers...@linux.intel.com wrote:
Please. Let's stop writing `!strcmp(...)` and `strcmp(...) == 0`. This
patch defines a little function streq() that reads much more nicely than
idioms using strcmp.
Signed-off-by: Chad Versace
On Wed, Mar 26, 2014 at 11:06:01AM -0700, Matt Turner wrote:
It's a one line wrapper -- just define it in the header and let
inlining do its thing.
Makes sense. I'll make the change before comitting.
___
Piglit mailing list
---
tests/all.py |1 +
tests/spec/CMakeLists.txt |1 +
tests/spec/gl-1.4/CMakeLists.gl.txt | 14 +++
tests/spec/gl-1.4/CMakeLists.txt |1 +
tests/spec/gl-1.4/dlist-multidrawarrays.c | 169 +
5
18 matches
Mail list logo