On Tue, 22 Jan 2019 at 15:32, Eric Engestrom wrote:
>
> Cc: Veluri Mithun
> Cc: Emil Velikov
> Cc: Rob Clark
> Cc: Nicolai Hähnle
> Signed-off-by: Eric Engestrom
> ---
> The extension is currently in development in this MR:
> https://gitlab.freedesktop.org/
On Tue, 22 Jan 2019 at 15:29, Eric Engestrom wrote:
> > I'd add a "TODO: add basic xml config validation/printing based on the
> > example in the spec."
>
> That would be good in theory, but:
> - we'd need to write a proper DTD, which we don't have right now
> - xml parsing is not trivial (we'd n
Hi Eric,
Thanks for writing this up.
On Tue, 22 Jan 2019 at 12:43, Eric Engestrom wrote:
>
> Cc: Veluri Mithun
> Cc: Emil Velikov
> Cc: Rob Clark
> Cc: Nicolai Hähnle
> Signed-off-by: Eric Engestrom
> ---
> The extension is currently in develop
On Fri, 11 Jan 2019 at 15:28, andrey simiklit wrote:
>
> Hello,
>
> Thanks for review.
> Could somebody help me with a push?
>
Done.
> @Emil:
> Yes, you are right about missing 'in-' in 'git send-email --in-reply-to='.
> So v3 wasn't added to the patchwork(
> Hope it isn't really mandatory and th
riable 'pos' which is always was 0
> - Implemented more effective number of levels calculator
> - Removed a stop when one of tests fails
> - Fixed two style issues
>( Emil Velikov )
Looks great, thank you.
Reviewed-by: Emil Velikov
-Emil
P.S.
Hi Andrii,
Pardon for joining so late. The patch looks ok, although a few bits I'm unsure
about - see the "Note" below.
There's a bunch of cosmetic suggestions as well - I won't read too much into
them though.
On Mon, 3 Dec 2018 at 15:07, wrote:
[snip]
> +#define TW 64
> +#define TH 64
#define
Hi Erik,
On Fri, 26 Oct 2018 at 14:03, Erik Faye-Lund
wrote:
> +#ifdef PIGLIT_USE_OPENGL
> +
> piglit_ortho_projection(piglit_width, piglit_height, GL_FALSE);
> +
> +#else // PIGLIT_USE_OPENGL_ES2
> +
> + tex_program = piglit_build_simple_program(vs_source, fs_source);
> + gl
On Fri, 26 Oct 2018 at 16:39, Erik Faye-Lund
wrote:
>
> On Fri, 2018-10-26 at 16:21 +0100, Emil Velikov wrote:
> > On Fri, 26 Oct 2018 at 14:03, Erik Faye-Lund
> > wrote:
> > >
> > > The EXT_texture_compression_s3tc spec lists that it requires
> > >
On Fri, 26 Oct 2018 at 12:35, Erik Faye-Lund
wrote:
>
> On Thu, 2018-10-25 at 14:39 +0100, Emil Velikov wrote:
> > On Thu, 25 Oct 2018 at 11:07, Erik Faye-Lund
> > wrote:
> > >
> > > On Wed, 2018-10-24 at 16:32 +0100, Emil Velikov wrote:
> > > > Hi
On Fri, 26 Oct 2018 at 14:03, Erik Faye-Lund
wrote:
>
> The EXT_texture_compression_s3tc spec lists that it requires
> OpenGL 1.1, not 1.0 like some of these list.
>
> In reality, this probably doesn't make a huge difference, as
> OpenGL 1.0 hardware/drivers are pretty much extinct, but let's
> ju
On Thu, 25 Oct 2018 at 11:07, Erik Faye-Lund
wrote:
>
> On Wed, 2018-10-24 at 16:32 +0100, Emil Velikov wrote:
> > Hi Erik,
> >
> > A bit of an open question, do we want this test on GLES? If yes, see
> > the GLES
> > notes below.
> >
>
> Right. I
sion("GL_EXT_framebuffer_object");
GLES:
#ifdef PIGLIT_USE_OPENGL
piglit_require_extension("GL_EXT_framebuffer_object");
#else
piglit_require_extension("GL_OES_framebuffer_object");
// Or use GLES 2.0 a
On Wed, 3 Oct 2018 at 13:27, Emil Velikov wrote:
>
> From: Emil Velikov
>
> As pointed out by Mathias opening the card node can fail in some cases.
> In boils down to a) the node is owned by root:video b) by default, the
> user is not part of the video group c) logind dyna
On Fri, 5 Oct 2018 at 17:06, Dylan Baker wrote:
>
> Because wflinfo returns GL legacy (non-profile) for gles on glx.
>
> v2: - If the platform is pure GLX then warn that things aren't going to
> work.
> ---
> framework/wflinfo.py | 21 +++--
> 1 file changed, 19 insertions(+
On Tue, 9 Oct 2018 at 16:50, Dylan Baker wrote:
>
> Quoting Emil Velikov (2018-10-09 05:35:39)
> > On Fri, 5 Oct 2018 at 19:13, Dylan Baker wrote:
> > >
> > > In other cases like this we return skip instead of warn.
> > Can you be more specific about those &q
On Tue, 25 Sep 2018 at 11:17, Emil Velikov wrote:
>
> From: Emil Velikov
>
> We do not use wayland-egl anywhere ... so let's remove it.
>
> Signed-off-by: Emil Velikov
> ---
> CMakeLists.txt | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
&
On Fri, 5 Oct 2018 at 19:13, Dylan Baker wrote:
>
> In other cases like this we return skip instead of warn.
Can you be more specific about those "other cases like these"? It's
the core part of your argument so w/o a quote/reference it's
impossible to make a decision.
> I think that it should
> b
From: Emil Velikov
As pointed out by Mathias opening the card node can fail in some cases.
In boils down to a) the node is owned by root:video b) by default, the
user is not part of the video group c) logind dynamically allows any
user to open the node, once a user has logged-in.
Thus if we use
From: Emil Velikov
We do not use wayland-egl anywhere ... so let's remove it.
Signed-off-by: Emil Velikov
---
CMakeLists.txt | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 8f19457fc..698517733 100644
--- a/CMakeLists.txt
one more comment inline below:
>
> On Tuesday, 14 August 2018 15:44:09 CEST Emil Velikov wrote:
>> From: Emil Velikov
>>
>> Since the functionality is more or less identical to
>> EGL_MESA_platform_surfaceless, the test with a copy of it.
>>
With some commit message (feel free to reuse the above) patch is
Reviewed-by: Emil Velikov
HTH
Emil
___
Piglit mailing list
Piglit@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/piglit
From: Emil Velikov
Since the functionality is more or less identical to
EGL_MESA_platform_surfaceless, the test with a copy of it.
Changes, as listed in the test itself, include:
- s/MESA_platform_surfaceless/EXT_platform_device/g
- entrypoint handling - eglQueryDeviceStringEXT
On 8 August 2018 at 17:18, Dylan Baker wrote:
> Quoting Emil Velikov (2018-08-08 05:33:44)
>> On 7 August 2018 at 02:53, Rhys Kidd wrote:
>> > Signed-off-by: Rhys Kidd
>> > ---
>> > .travis.yml | 3 +++
>> > tox.ini | 9 +
>&g
On 8 August 2018 at 15:50, Eric Engestrom wrote:
> Cc: Emil Velikov
> Fixes: 7356bdd050b8475ed13c "dispatch: fallback to eglGetProcAddress for
> core functions"
> Signed-off-by: Eric Engestrom
Feel free to push this ASAP.
Reviewed-by
On 7 August 2018 at 02:53, Rhys Kidd wrote:
> Signed-off-by: Rhys Kidd
> ---
> .travis.yml | 3 +++
> tox.ini | 9 +
> 2 files changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/.travis.yml b/.travis.yml
> index b6e8136e8..45b23f009 100644
> --- a/.travis.yml
> +++ b/.travis.y
From: Emil Velikov
Plug the memory leaks in the PASS case. It might be worth doing the same
throughout, although that is too invasive for what I have time for.
Signed-off-by: Emil Velikov
---
.../egl_mesa_platform_surfaceless.c | 5 -
1 file changed, 4 insertions
From: Emil Velikov
The function returns EGL_BAD_DEVICE_EXT when an invalid device is used.
Signed-off-by: Emil Velikov
---
tests/egl/spec/egl_ext_device_query/egl_ext_device_query.c | 4
1 file changed, 4 insertions(+)
diff --git a/tests/egl/spec/egl_ext_device_query
platform_device
Any review and input will be appreciated.
Thanks
Emil
Emil Velikov (9):
egl_ext_device_query: check for eglQueryDeviceStringEXT
EGL_BAD_DEVICE_EXT
egl: Add basic EGL_MESA_device_software test
egl_ext_device_query: drop unused major/minor from eglInitialize
From: Emil Velikov
In the unlikely case that the function fails, we want some feedback
about the issue. It's not a failure per-se, but we don't want to
silently continue.
Signed-off-by: Emil Velikov
---
tests/egl/spec/egl_ext_device_query/egl_ext_device_query.c | 2 +-
1 file
From: Emil Velikov
Cover the new token and the interactions with EGL_EXT_platform_device.
Pay attention to Issue 1 of the spec, the fd ownership and ensure we
query back the same EGLDevice.
Signed-off-by: Emil Velikov
---
tests/egl/spec/CMakeLists.txt | 1
From: Emil Velikov
Since the functionality is more or less identical to
EGL_MESA_platform_surfaceless, the test with a copy of it.
Changes, as listed in the test itself, include:
- s/MESA_platform_surfaceless/EXT_platform_device/g
- eglQueryDevicesEXT and eglGetPlatformDisplayEXT entrypoing
From: Emil Velikov
Call eglTerminate, otherwise we'll end up with massive leaks reported in
Valgrind. Fairly useful when checking if the EGL implementation is
leak-free.
Signed-off-by: Emil Velikov
---
.../egl_ext_device_query.c| 20 +++
1 file change
From: Emil Velikov
The arguments are optional. Since we don't care about the version, use
NULL instead.
Signed-off-by: Emil Velikov
---
tests/egl/spec/egl_ext_device_query/egl_ext_device_query.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/tests/egl
From: Emil Velikov
The extension doesn't do anything useful but report a device ext.
string. Ensure that such a device does not attempt to expose DRM
specifics - nearly missed that during development.
Signed-off-by: Emil Velikov
---
tests/egl/spec/CMakeLists.txt
From: Emil Velikov
The first call gives us the number of devices, while the second
populates the user provided array. Check that the second call returns at
least one device.
Signed-off-by: Emil Velikov
---
.../egl_ext_device_enumeration/egl_ext_device_enumeration.c | 5 +
1 file changed
HI Illia,
Please don't top-post.
On 30 July 2018 at 10:51, Illia Iorin wrote:
> ping
>
> On Tue, Jul 10, 2018 at 2:25 PM Illia Iorin wrote:
>>
>> Hello,
>> Full explanation can be found in
Having a bug report is great, but the commit itself should be self sufficient.
>> https://bugs.freedeskto
did last year.
>
>
> Without negative feedback, I'm planning to push these soon.
>
To provide some clarity (in 3/3), might be worth coping the commit
message as inline comment.
Regardless, the series is
Reviewed-by: Emil Velikov
-Emil
___
Piglit mailing list
Piglit@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/piglit
Hi Dan,
Sharing a couple of small ideas.
On 5 June 2018 at 18:02, Daniel Stone wrote:
>>> drm-gralloc.git
Empty - please nuke, alongside bugzilla & other infra.
>>> drm.git
Out of curiosity - this and others (say igt) projects are accessible
as mesa/$foo and drm/$foo.
I'd image the same approa
On 21 December 2017 at 09:26, Tapani Pälli wrote:
>
> On 12/18/2017 04:37 PM, Emil Velikov wrote:
>>
>> On 13 December 2017 at 12:54, Tapani Pälli wrote:
>>
>>>> The test doesn't care how well eglCopyBuffers itself works - aim it to
>>>> il
On 13 December 2017 at 12:54, Tapani Pälli wrote:
>> The test doesn't care how well eglCopyBuffers itself works - aim it to
>> illustrate the buggy validation in Mesa.
>> Hence the wait + pixmap readback are not really needed.
>>
>> Admittedly the test name is quite misleading as-is - I'm short o
Hi Fabian,
On 16 December 2017 at 19:21, Fabian Bieler wrote:
> ---
> tests/all.py| 2 +-
> tests/bugs/CMakeLists.gl.txt| 1 -
> tests/bugs/tex1d-2dborder.c | 126
>
> tests/spec/gl-1.4/CMakeLists.gl.txt | 1 +
>
Hi Tapani,
Thanks for having a look!
On 13 December 2017 at 06:10, Tapani Pälli wrote:
> Hi;
>
> On 11.12.2017 22:15, Emil Velikov wrote:
>> + /* Set the env. variable to force the platform 'detection' to use
>> +* different platform (than X11).
&
On 12 December 2017 at 09:08, Tapani Pälli wrote:
>
>
> On 12/12/2017 10:23 AM, Tapani Pälli wrote:
>>
>>
>> On 12/11/2017 10:15 PM, Emil Velikov wrote:
>>>
>>> From: Emil Velikov
>>>
>>> A couple of tests were calling eglBindAPI b
From: Emil Velikov
The latter is a legacy and somewhat of a hack.
Since we explicitly know that this requires the X11, simply use the
helper which will use the correct EGL_PLATFORM API.
Cc: Ian Romanick
Signed-off-by: Emil Velikov
---
tests/egl/egl-util.c | 4 ++--
1 file changed, 2
From: Emil Velikov
Allows us to consolidate native platform specifics in once place.
Will be reused in later commit(s).
Cc: Ian Romanick
Signed-off-by: Emil Velikov
---
tests/egl/egl-util.c | 12 +---
tests/egl/egl-util.h | 3 +++
2 files changed, 12 insertions(+), 3 deletions
From: Emil Velikov
A couple of tests were calling eglBindAPI before eglInitialize.
Some versions of Mesa had problem with such invocation order and the
seeming consensus was that apps should not do that.
This is likely a copy/paste mistake, that was also present in Xserver.
Yet the latter was
From: Emil Velikov
Current Mesa implementation was overly cautious, flagging an error
where it shouldn't - patch for Mesa is on the list.
See the test for details.
Cc: Ian Romanick
Signed-off-by: Emil Velikov
---
Nitpicks:
- suggestions for test name and category?
- should we bother a
From: Emil Velikov
Some places (like the next patch) want to pass non-default dpy.
Cc: Ian Romanick
Signed-off-by: Emil Velikov
---
tests/util/piglit-util-egl.c | 10 --
tests/util/piglit-util-egl.h | 11 ++-
2 files changed, 18 insertions(+), 3 deletions(-)
diff --git a
needs to be rewritten with proper English, not sure how. Maybe simply:
>
> "Mali 450 driver does not support eglGetPlatformDisplay, use
> piglit_egl_get_default_display wrapper instead."
>
>
Looks good, one could also reuse the text from 45095dc08b.
In either case, patch i
On 3 August 2017 at 19:24, Grigori Goronzy wrote:
>> I don't know if we need the no_error here. The test should check if
>> things work correctly when it's set.
>> Everything else falls outside the scope of the extension.
>>
>
> I added testing with no-error==false because there were some issues
On 3 August 2017 at 01:10, Timothy Arceri wrote:
> One thing that is missing is a check that an error is generated (or the
> context is created successfully) when a shared context is created. Depending
> on whether the no error attributes match.
>
> I've got a partial Mesa implementation for the G
Hi Grigori,
On 19 July 2017 at 22:41, Grigori Goronzy wrote:
> This test verifies context creation with the
> EGL_KHR_create_context_no_error extension, which includes interaction
> with debug and robustness flags. The test also verifies that the
> KHR_no_error mode is successfully enabled with a
> expanded from macro 'NULL'
>^~
> 2 warnings generated.
>
> v2: change function pointer declaration (Emil)
>
> Signed-off-by: Juan A. Suarez Romero
Reviewed-by: Emil Velikov
Thanks
Emil
___
Piglit mailing list
Piglit@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/piglit
On 10 July 2017 at 11:20, Juan A. Suarez Romero wrote:
> On Fri, 2017-06-30 at 12:18 +0200, Juan A. Suarez Romero wrote:
>> Fixes
>>
>> /build/piglit/tests/egl/spec/egl_mesa_platform_surfaceless/egl_mesa_platform_surfaceless.c:94:30:
>> warning: incompatible pointer to integer conversion passing '
l-invalid-attr.c
> +
> +#include
> +#include
> +
Did not see these before - they should be dropped. Barring any
objections I'll address the above nitpicks and merge the patch in a
day or two.
If someone beats me to it:
Reviewed-by: Emil Velikov
-Emil
___
Piglit mailing list
Piglit@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/piglit
On 12 January 2017 at 05:43, Tapani Pälli wrote:
> On 01/11/2017 05:24 PM, Emil Velikov wrote:
>>
>> On 10 January 2017 at 10:51, Tapani Pälli wrote:
>>>
>>> This test does not work correctly for me (skips because
>>> EGL_KHR_image_base
>>>
On 10 January 2017 at 10:51, Tapani Pälli wrote:
> This test does not work correctly for me (skips because EGL_KHR_image_base
> is unsupported even though driver supports it). I believe you either would
> need to utilize egl-util framework (check the other tests in egl directory)
> or initialize E
Haven't looked at the actual implementation (2/2) but this looks quite
reasonable. We can even use libxkbcommon-x11 (for the X11/XCB backend)
at some later stage.
Reviewed-by: Emil Velikov
Emil
___
Piglit mailing list
Piglit@list
On 4 November 2016 at 21:09, Vinson Lee wrote:
> On Fri, Nov 4, 2016 at 8:45 AM, Brian Paul wrote:
>> On 11/03/2016 03:36 PM, Vinson Lee wrote:
>>>
>>> Fix MinGW build error.
>>>
>>> piglit-util.c: In function 'piglit_time_is_monotonic':
>>> piglit-util.c:583:18: error: storage size of 't' isn't
On 21 October 2016 at 07:51, Tapani Pälli wrote:
> Hi;
>
> I wanted to try running piglit tests with PIGLIT_PLATFORM=wayland. This
> makes Piglit to use piglit-framework-gl. Tests crashed always at end
> which I traced to abort removed in patch 2 but when building to debug
> mode I also hit couple
On 12 September 2016 at 14:55, Marek Olšák wrote:
> On Mon, Sep 12, 2016 at 3:33 PM, Emil Velikov
> wrote:
>> On 9 September 2016 at 14:10, Marek Olšák wrote:
>>> From: Marek Olšák
>>>
>>> Older versions don't work and there is no interest in fixing
t. System load no affect here, afaict.
What is your observation before/after the patch, please add some
numbers to the commit message.
As it helps at least some setups:
Acked-by: Emil Velikov
Thanks
Emil
___
Piglit mailing list
Piglit@lists.freedesktop.
On 9 September 2016 at 14:10, Marek Olšák wrote:
> From: Marek Olšák
>
> Older versions don't work and there is no interest in fixing the official
> gl45release CTS branch, so it's useless to add those forever-broken tests.
>
Surely you meant so say s/official gl45release CTS branch/official pre
On 8 September 2016 at 05:06, Adam Jackson wrote:
> On Thu, 2016-09-08 at 03:42 +0100, Emil Velikov wrote:
>> > On 7 September 2016 at 21:50, Adam Jackson wrote:
>> > index 075df97..6512b6b 100644
>> > --- a/tests/egl/egl-nok-swap-region.c
>> > +++ b/tests/
On 7 September 2016 at 21:50, Adam Jackson wrote:
> New eglext.h has the typedef corrected to end in PROC like everything
> else. Rename the old string to the new when building against old
> headers.
>
> Signed-off-by: Adam Jackson
> ---
> tests/egl/egl-nok-swap-region.c | 5 +++--
> 1 file chan
On Tuesday, 6 September 2016, Rob Clark wrote:
> On Tue, Sep 6, 2016 at 4:04 PM, Kristian Høgsberg > wrote:
>
> > If we have the gbm eglimage path, why do we keep the intel specific code
> around?
>
> only plausible reason I could come up with is that gbm path depends on
> a fairly recent vers
On 2 September 2016 at 07:15, Mathias Fröhlich
wrote:
>
> Great!
>
> One question that I cannot forsee from your branch:
>
> The EGL_EXT_device_enumeration spec says
>
>
>
> [...] All implementations must support
>
> at least one device.
>
> [...]
>
>
>
> Which means to me that once an applicatio
On 1 September 2016 at 15:44, Adam Jackson wrote:
> On Thu, 2016-09-01 at 15:36 +0100, Emil Velikov wrote:
>
>> Did you have the chance to respin the EGL device patches for mesa or
>> shall I ?
>
> I have a work-in-progress branch to port it to the drmDe
On 31 August 2016 at 19:39, Adam Jackson wrote:
> v2:
> - Cover more error paths, also look up QueryDeviceAttrib (Eric Anholt)
> - Fetch function pointers before EGL initialized (Mathias Fröhlich)
>
> v3:
> - Use piglit_check_egl_error and PFN* typedefs, fix NOT_INITIALIZED test
On 31 August 2016 at 08:53, Jussi Kukkonen wrote:
> On 30 August 2016 at 18:33, Emil Velikov wrote:
> ...
>>
>> Fwiw this and 1/2 are
>> Reviewed-by: Emil Velikov
>>
>> Do you have commit access or should I push these for you ?
>
> I don't, so
as a few bugs. The one that applies there is
that every 'secondary' library exposes it's own symbols alongside the
symbols of its dependencies. The latter of which pulls the
dependencies of the original library into the 'secondary' one.
But that and similar fun stuff is out
On 29 August 2016 at 09:39, Jussi Kukkonen wrote:
> Without this linking CXX executables (e.g. fbo-blit-stretch) fails
> with gold linker:
>
> | libpiglitutil_gl.so.0: error: undefined reference to 'dlsym'
> | libpiglitutil_gl.so.0: error: undefined reference to 'dlerror'
> | libpiglitutil_gl.so.0
Hi Daniel,
On 20 August 2016 at 04:20, Daniel Scharrer wrote:
> Signed-off-by: Daniel Scharrer
> ---
>
> This tests for the bug fixed by the Mesa patch at
> https://patchwork.freedesktop.org/patch/106411/
>
> The test passes on amdgpu-pro, llvmpipe (patched and unpatched)
> and radeonsi (patched
On 27 June 2016 at 17:57, Emil Velikov wrote:
> Hi all,
>
> As discussed over at mesa-dev@ some implementations' (such as mesa and
> GLVND) libGLESv2.so.1 do not provide 'core' symbols via dlsym().
>
> This series resolves this, plus one could even follow the app
On 22 July 2016 at 16:17, Adam Jackson wrote:
> + EGLBoolean (*queryDisplayAttrib)(EGLDisplay dpy, EGLint name,
> +EGLAttrib *value);
> + const char *(*queryDeviceString)(EGLDisplay dpy, EGLint name);
> + EGLBoolean (*queryDeviceAttrib)(EG
On 30 June 2016 at 10:19, Andres Gomez wrote:
> On Wed, 2016-06-29 at 17:46 +0100, Emil Velikov wrote:
>> On 14 June 2016 at 22:36, Andres Gomez wrote:
>> > For some cases we want to have shaders where we load an exact bit
>> > pattern into a signed int.
>>
On 14 June 2016 at 22:36, Andres Gomez wrote:
> For some cases we want to have shaders where we load an exact bit
> pattern into a signed int.
>
> This fixes the errno-based range validation that was broken when the
> integer vbo attribute parsing on 32-bit systems was recently fixed.
>
Is this be
On 28 June 2016 at 16:10, Ilia Mirkin wrote:
> On Tue, Jun 28, 2016 at 10:45 AM, Emil Velikov
> wrote:
>> On 27 June 2016 at 18:04, Ilia Mirkin wrote:
>>> Don't we build against local GL headers? So we pretty much control
>>> what's available and
On 27 June 2016 at 18:04, Ilia Mirkin wrote:
> Don't we build against local GL headers? So we pretty much control
> what's available and what's not. i.e. do we need those ifdefs?
>
No we don't have/build against local GL headers. We use the system ones.
-Emil
_
From: Emil Velikov
The definition and/or interpretation what is core function may vary
between vendors. Thus let's opt for the more robust option and use
eglGetProcAddress in case dlsym() fails.
Signed-off-by: Emil Velikov
---
tests/util/piglit-dispatch-init.c
Hi all,
As discussed over at mesa-dev@ some implementations' (such as mesa and
GLVND) libGLESv2.so.1 do not provide 'core' symbols via dlsym().
This series resolves this, plus one could even follow the approach used
and greatly simplify the dispatch. Let me know if you're ok with the
direction
From: Emil Velikov
The symbol is already the default, thus there's no need to specify it.
Signed-off-by: Emil Velikov
---
tests/util/piglit-dispatch-init.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/util/piglit-dispatch-init.c
b/tests/util/piglit-dis
From: Emil Velikov
Analogous to previous commit.
Signed-off-by: Emil Velikov
---
tests/spec/arb_direct_state_access/gettextureimage-luminance.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/tests/spec/arb_direct_state_access/gettextureimage-luminance.c
b/tests/spec
From: Emil Velikov
piglit_use_vertex_program/piglit_use_fragment_program are used _after_
piglit_init() at which point the winsys layer is already up and the
dispatch is already setup.
Signed-off-by: Emil Velikov
---
tests/util/piglit-util-gl.c | 2 --
1 file changed, 2 deletions(-)
diff
From: Emil Velikov
In order to do that move the function further up, and while where doing
that use tabs to indent, as the rest of the file.
Signed-off-by: Emil Velikov
---
tests/util/piglit-dispatch-init.c | 56 +--
1 file changed, 24 insertions(+), 32
From: Emil Velikov
Signed-off-by: Emil Velikov
---
tests/util/piglit-dispatch-init.c | 14 +-
1 file changed, 5 insertions(+), 9 deletions(-)
diff --git a/tests/util/piglit-dispatch-init.c
b/tests/util/piglit-dispatch-init.c
index 7ae512b..56b7a3d 100644
--- a/tests/util/piglit
From: Emil Velikov
Signed-off-by: Emil Velikov
---
tests/glx/glx-swap-event.c | 9 +
1 file changed, 1 insertion(+), 8 deletions(-)
diff --git a/tests/glx/glx-swap-event.c b/tests/glx/glx-swap-event.c
index 7519a51..a9f4136 100644
--- a/tests/glx/glx-swap-event.c
+++ b/tests/glx/glx
From: Emil Velikov
Analogous to previous commit(s).
With this commit no GL test in piglit (barring glean) relies on this
broken behaviour. Checked with the following:
$ git grep "defined.*\
---
.../arb_es2_compatibility/arb_es2_compatibility-depthrangef.c
Hi all,
As mentioned a while back, piglit should _not_ optionally compile the
test(s) depending on the presence of symbols in the GL headers.
Doing so, is a bad idea since the test(s) will skip regardless if the
API is available at runtime.
The lot is dead trivial, so hopefully we'll get a bra
From: Emil Velikov
Signed-off-by: Emil Velikov
---
tests/spec/arb_copy_image/formats.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/tests/spec/arb_copy_image/formats.c
b/tests/spec/arb_copy_image/formats.c
index f109e99..3af55a7 100644
--- a/tests/spec/arb_copy_image/formats.c
+++ b
From: Emil Velikov
Analogous to previous commits.
Signed-off-by: Emil Velikov
---
tests/egl/egl-nok-swap-region.c | 14 --
1 file changed, 14 deletions(-)
diff --git a/tests/egl/egl-nok-swap-region.c b/tests/egl/egl-nok-swap-region.c
index 4d2734d..075df97 100644
--- a/tests/egl
From: Emil Velikov
Analogous to previous commit(s).
Signed-off-by: Emil Velikov
---
tests/spec/arb_copy_image/formats.c | 29 +
1 file changed, 5 insertions(+), 24 deletions(-)
diff --git a/tests/spec/arb_copy_image/formats.c
b/tests/spec/arb_copy_image/formats.c
From: Emil Velikov
Analogous to previous commit - we don't want the test to differ between
builds. This could lead to the test silently skipping when built against
old headers, _regardless_ if at runtime the extension is supported or
not.
With this commit no GLX test in piglit (barring
From: Emil Velikov
Those symbols are part of the ABI/API and cannot change. Just define
them locally when needed.
Signed-off-by: Emil Velikov
---
tests/util/piglit-util-gl.c | 18 +++---
1 file changed, 11 insertions(+), 7 deletions(-)
diff --git a/tests/util/piglit-util-gl.c b
From: Emil Velikov
Analogous to previous commits.
With this commit no EGL test in piglit relies on this broken behaviour.
Checked with the following:
$ git grep "defined.*\
---
tests/egl/egl-nok-texture-from-pixmap.c | 14 --
1 file changed, 14 deletions(-)
diff --git a/test
From: Emil Velikov
We want the test to be complete/consistent regardless if the headers
we're building against are new or old. In case of the the latter we
could add the definitions locally, if needed.
Signed-off-by: Emil Velikov
---
tests/glx/glx-swap-event.c | 14 --
1
On 20 June 2016 at 15:46, Brian Paul wrote:
> On 06/17/2016 07:59 PM, Emil Velikov wrote:
>>
>> On 17 June 2016 at 16:53, Brian Paul wrote:
>>>
>>>
>>> I spent a few hours yesterday pulling out my hair trying to understand
>>> why
>>&g
On 17 June 2016 at 16:53, Brian Paul wrote:
>
> I spent a few hours yesterday pulling out my hair trying to understand why
> the piglit fbo-mipmap-copypix test was failing on Windows. But it was only
> failing when I ran it directly. It passed when I ran it via piglit-run.py
>
> The key differen
On 19 May 2016 at 07:33, Juan A. Suarez Romero 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,
>> it
>> also creates more churn for people using the functionality.
>
>
> Then, I'll keep as it is. We can always ch
On 31 March 2016 at 12:21, Jose Fonseca wrote:
> On 31/03/16 00:44, Emil Velikov wrote:
>>
>> On 30 March 2016 at 21:54, Jose Fonseca wrote:
>>>
>>> On 30/03/16 11:14, Emil Velikov wrote:
>>>>
>>>>
>>>> On 29 March 2016
1 - 100 of 329 matches
Mail list logo
