On Thu, Mar 25, 2021 at 9:41 AM Rob Clark wrote:
>
> On Wed, Mar 24, 2021 at 9:15 PM Jason Ekstrand wrote:
> >
> > On March 24, 2021 22:25:10 Rob Clark wrote:
> >
> >> On Wed, Mar 24, 2021 at 3:52 PM Jordan Justen
> >> wrote:
> >>>
On March 24, 2021 22:25:10 Rob Clark wrote:
On Wed, Mar 24, 2021 at 3:52 PM Jordan Justen
wrote:
On 2021-03-23 09:38:59, Eric Anholt wrote:
On Tue, Mar 23, 2021 at 7:02 AM Jason Ekstrand wrote:
Trying to pick this discussion back up. Daniel Stone thinks it's a
half hour of API bashing
rb
On Thu, Feb 14, 2019 at 12:49 PM Lionel Landwerlin <
lionel.g.landwer...@intel.com> wrote:
> We were dropping negations/abs while trying to optimize.
>
> Signed-off-by: Lionel Landwerlin
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109601
> ---
>
On Wed, Dec 19, 2018 at 3:57 PM Dylan Baker wrote:
> Hi list,
>
> I've been off and on working on a new test runner for piglit written in
> C++14.
> Its been sitting on my gitlab publicly for a bit, and I've shared it with
> some
> people at Intel, but I thought I'd send out an announcement and
Wow. I think this may be the best use I've ever seen of uniform
initializers. :-)
Seems to do the thing in the bug.
Reviewed-by: Jason Ekstrand
> +
> +void main()
> +{
> +bool a = x == 37;// Must be true.
> +bool b = y == 46;// Must be true.
> +
> +/*
eturn result;
> -}
> -
> -void
> -piglit_init(int argc, char **argv)
> -{
> - static const char intel_id[] = "Intel Open Source Technology
> Center";
> - const char *vendor_str;
> - EGLDisplay egl_dpy = eglGetCurrentDi
commit dd553bc67f8ab1513fd196b6ffb7c4a76723adfd
Author: Jason Ekstrand
Date: Wed Oct 3 12:14:20 2018 -0500
nir/alu_to_scalar: Use ssa_for_alu_src in hand-rolled expansions
The ssa_for_alu_src helper will correctly handle swizzles and other
source modifiers for you. The expansions
---
.../fs_pack.shader_test.mako | 14 +++---
.../fs_unpack.shader_test.mako | 6 +++---
.../vs_pack.shader_test.mako | 14 +++---
.../vs_unpack.shader_test.mako | 6 +++---
4 files changed, 28
---
Makefile.am | 2 +
data/32x32-smile.ref.png | Bin 0 -> 315 bytes
src/tests/func/shader/constants.c | 245 ++
3 files changed, 247 insertions(+)
create mode 100644 data/32x32-smile.ref.png
create mode 100644
---
Makefile.am| 3 +
src/framework/test/t_dump.c| 2 +-
src/framework/test/t_result.c | 6 +-
src/tests/bug/104809.c | 160 +
src/tests/self/concurrent-output.c | 4 +-
5 files changed, 169 insertions(+), 6
We already had one, it just didn't have any load/store ops and it wasn't
exposed to the t_ API.
---
include/tapi/t_data.h| 2 ++
src/framework/test/t_data.c | 11
src/framework/test/t_phase_setup.c | 18
src/framework/test/test.h
We already had one, it just didn't have any load/store ops and it wasn't
exposed to the t_ API.
---
include/tapi/t_data.h| 2 ++
src/framework/test/t_data.c | 11
src/framework/test/t_phase_setup.c | 18
src/framework/test/test.h
---
tests/glx/glx-tfp.c | 102 +++-
1 file changed, 62 insertions(+), 40 deletions(-)
diff --git a/tests/glx/glx-tfp.c b/tests/glx/glx-tfp.c
index 26408741a..46e317adc 100644
--- a/tests/glx/glx-tfp.c
+++ b/tests/glx/glx-tfp.c
@@ -48,9 +48,12 @@ GLfloat
Thanks for writing a test!
On Wed, Jun 13, 2018 at 3:54 AM, wrote:
> From: Danylo Piliaiev
>
> Clipper can mess up provoking vertex.
> Ref: https://bugs.freedesktop.org/show_bug.cgi?id=103047
>
> Signed-off-by: Danylo Piliaiev
> ---
> tests/opengl.py| 1 +
This commit adds a new PIGLIT_TESTS_TIMEOUT environment variable which
overrides the pre-test timeout specified by the test runner.
---
This is almost certainly not the right solution but it gets the discussion
started. A couple of known issues:
- No documentation
- It overrides. Maybe it
---
tests/util/piglit-util-gl.c | 83 ++---
1 file changed, 3 insertions(+), 80 deletions(-)
diff --git a/tests/util/piglit-util-gl.c b/tests/util/piglit-util-gl.c
index 54fb3d0c4..a19eef5f0 100644
--- a/tests/util/piglit-util-gl.c
+++
Instead of doing 900 piglit_proble_pixel_rgba calls (one for each pixel)
which each call glReadPixels, use the new helper which only invokes
glReadPixels once.
---
tests/texturing/shaders/textureGather.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git
All the other helpers can only assert that a given rectangle is the
given solid color; this one takes an array of pixel data.
---
tests/util/piglit-util-gl.c | 8
tests/util/piglit-util-gl.h | 2 ++
2 files changed, 10 insertions(+)
diff --git a/tests/util/piglit-util-gl.c
---
tests/util/piglit-util-gl.c | 47 +
1 file changed, 47 insertions(+)
diff --git a/tests/util/piglit-util-gl.c b/tests/util/piglit-util-gl.c
index 39b6e58e9..54fb3d0c4 100644
--- a/tests/util/piglit-util-gl.c
+++ b/tests/util/piglit-util-gl.c
@@ -1296,6
---
tests/util/piglit-util-gl.c | 19 ++-
1 file changed, 14 insertions(+), 5 deletions(-)
diff --git a/tests/util/piglit-util-gl.c b/tests/util/piglit-util-gl.c
index 11e7d4b1d..39b6e58e9 100644
--- a/tests/util/piglit-util-gl.c
+++ b/tests/util/piglit-util-gl.c
@@ -1249,7
Reviewed-by: Jason Ekstrand
Thanks for writing this!
On Thu, Jun 7, 2018 at 11:58 AM, Lionel Landwerlin <
lionel.g.landwer...@intel.com> wrote:
> We've been running into crashes in i965 because of a number of issue
> throughout the driver stack. This is a really basic tes
On Thu, Jun 7, 2018 at 7:16 AM, Lionel Landwerlin <
lionel.g.landwer...@intel.com> wrote:
> We've been running into crashes in i965 because of a number of issue
> throughout the driver stack. This is a really basic tests that list
> the formats and their associated modifiers to just catch any
>
The extra indentation does not make markdown happy. Instead, use a
proper list and make the test script names bold.
---
README.md | 111 ++
1 file changed, 37 insertions(+), 74 deletions(-)
diff --git a/README.md b/README.md
index
Extra indentation results in code blocks
---
README.md | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/README.md b/README.md
index eee613095..f363211e2 100644
--- a/README.md
+++ b/README.md
@@ -225,12 +225,12 @@ To create some nice formatted test summaries, run
$
---
README.md | 19 +++
1 file changed, 7 insertions(+), 12 deletions(-)
diff --git a/README.md b/README.md
index 9aebbb47d..b60770338 100644
--- a/README.md
+++ b/README.md
@@ -237,19 +237,14 @@ Have a look at the results with a browser:
The summary shows the 'status' of a
---
README => README.md | 0
1 file changed, 0 insertions(+), 0 deletions(-)
rename README => README.md (100%)
diff --git a/README b/README.md
similarity index 100%
rename from README
rename to README.md
--
2.17.1
___
Piglit mailing list
---
README.md | 56 +--
1 file changed, 30 insertions(+), 26 deletions(-)
diff --git a/README.md b/README.md
index f363211e2..9aebbb47d 100644
--- a/README.md
+++ b/README.md
@@ -79,8 +79,7 @@ Now configure the build system:
This will start
---
README.md | 107 +-
1 file changed, 58 insertions(+), 49 deletions(-)
diff --git a/README.md b/README.md
index 9f50d2ec9..eee613095 100644
--- a/README.md
+++ b/README.md
@@ -74,34 +74,36 @@ For testing the python framework using "py.test
---
README.md | 50 +++---
1 file changed, 19 insertions(+), 31 deletions(-)
diff --git a/README.md b/README.md
index 66c6bca95..9f50d2ec9 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,7 @@
Piglit
---
+==
+
1. About
2. Setup
3. How to
---
README.md | 56 +++
1 file changed, 28 insertions(+), 28 deletions(-)
diff --git a/README.md b/README.md
index dac9ba26b..7259ba574 100644
--- a/README.md
+++ b/README.md
@@ -376,45 +376,45 @@ All new tests must be added to the appropriate
4, 2018 at 3:29 PM, Eric Anholt wrote:
> Jason Ekstrand writes:
>
> > All,
> >
> > Sorry for the mess of GitLab e-mails but there are a lot of questions to
> > ask as this process moves forward. Today, we're discussing piglit. I've
> > included both t
sa.
Mostly, this is a question of whether we consider piglit to be it's own
project on freedesktop or a sub-project of mesa. I don't know the answer
to that question.
Thoughts?
--Jason Ekstrand
___
Piglit mailing list
Piglit@lists.freedesktop
I just pushed an equivalent of this. Also, I pushed a patch to add Vulkan
headers to crucible so that it builds off the latest and not off whatever's
on your system.
On Wed, Feb 28, 2018 at 10:28 AM, Tapani Pälli
wrote:
>
>
> On 28.02.2018 20:26, Tapani Pälli wrote:
>
Looks good to me. All three are
Reviewed-by: Jason Ekstrand <ja...@jlekstrand.net>
On Tue, Feb 27, 2018 at 12:10 AM, Tapani Pälli <tapani.pa...@intel.com>
wrote:
> Patch adds a new test that is expected to fail, new option and
> some minor refactor done to check_memory
On Tue, Feb 27, 2018 at 12:10 AM, Tapani Pälli
wrote:
> Patch adds VK_EXT_global_priority support. This should not affect
> test results.
>
> Signed-off-by: Tapani Pälli
> ---
> src/tests/func/sync/semaphore-fd.c | 23 ---
> 1
nable all three external memory
extensions and all three external fence extensions. With that,
Reviewed-by: Jason Ekstrand <ja...@jlekstrand.net>
> + };
> +
> VkResult result = vkCreateDevice(t_physical_dev,
> &(VkDeviceCreateInfo) {
On Thu, Jan 18, 2018 at 11:45 AM, Fabian Bieler
wrote:
> Hello!
>
> The Crucible TODO.txt file contains:
> - Add a --diff option to crucible-run that dumps image diffs for failed
> tests.
> (Requires multiple command buffers).
>
> I do not understand how an optional
FYI, the new test fails on i965 with this:
Failed to compile vertex shader: 0:1(10): error: GLSL 1.40 is not
supported. Supported versions are: 1.10, 1.20, 1.30, 1.00 ES, and 3.00
ES
Is this expected? I'm guessing it is since we don't support compat.
On Mon, Oct 23, 2017 at 5:31 AM, Marek
The orange is not a 0/1 color. If the driver messes up and gives the
client sRGB decode on the texture, the test will now fail.
---
tests/glx/glx-tfp.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tests/glx/glx-tfp.c b/tests/glx/glx-tfp.c
index c27c89a..2640874 100644
Seriously, 1000 is not the only 64-bit integer. We should test
a few more of them.
These new integers trigger a bug in the GLSL IR int64 lowering code.
Cc: Ian Romanick
---
generated_tests/builtin_function.py | 24 ++--
1 file changed, 22
Tested-by: Jason Ekstrand <ja...@jlekstrand.net>
On Thu, Feb 23, 2017 at 4:29 PM, Dylan Baker <dy...@pnwbakers.com> wrote:
> The extension requires a GL_ prefix, which the template didn't have.
>
> v2: - Actually fix things
>
> Signed-off-by: Dylan Bake
On Nov 17, 2016 10:56 AM, "Plamena Manolova"
wrote:
>
> A basic test to check whether the values written to gl_SampleMaskIn
> are still correct after enabling the ARB_post_depth_coverage
> extension.
>
> Signed-off-by: Plamena Manolova
>
LGTM R-B
On Aug 11, 2016 12:30 PM, "Anuj Phogat" wrote:
> This is required following a change in 8X multisample positions
> in i965 driver.
>
> Signed-off-by: Anuj Phogat
> ---
> .../ext_framebuffer_multisample_blit_scaled/blit-scaled.cpp| 10
rule is that integer MSAA resolves are supposed to pick a single arbitrary
sample for each texel rather than trying to combine them in any way.
Signed-off-by: Jason Ekstrand <ja...@jlekstrand.net>
---
tests/spec/ext_framebuffer_multisample/formats.cpp | 196 -
1 file change
We used to always use GL_TEXTURE_RECTANGLE which we may not actually want.
Signed-off-by: Jason Ekstrand <ja...@jlekstrand.net>
---
tests/util/piglit-fbo.cpp | 14 +++---
tests/util/piglit-fbo.h | 8
2 files changed, 15 insertions(+), 7 deletions(-)
diff --git a/test
both are
Reviewed-by: Jason Ekstrand <ja...@jlekstrand.net>
On Wed, Apr 13, 2016 at 12:07 PM, Matt Turner <matts...@gmail.com> wrote:
> Unfortunately, the GLSL spec says that exponent values outside of [-126,
> 128] are undefined or may be flushed to zero.
> ---
&
I think the default mode we want here is more of a "continue after crash"
mode where, when it crashes, we don't rerun anything but instead just
continue starting from the next test. This is what Google does with their
"cherry" tool and it would prevent ensure that random crashes don't get
hidden
This is mostly a copy-and-paste of the original GLES3 dEQP runner with a
few added tidbits that makes running the Vulkan CTS easier. The two
force-skip cases were very useful in the early days when lots of tests were
still in development and may no longer be needed. However, I have a
feeling
it.
>
> v2: - Update tests to use actual deqp ouput.
> - Fix status loop, v1 had a bug that would cause the loop to not
> exit when it needed to, but it would pass the simplified tests.
Seems to work as advertised this time. Thanks!
Tested-by: Jason Ekstrand <jason.ekstr..
e_format_bgra/teximage.c
Let's call this api-errors.c or something like that. It really isn't
testing TexImage. We should probably eventually have a test that
tests that the texture actually gets on the screen. However, I'm not
too terribly worried about it since, as long as we don't error out
This catches a bug in the i965 backend compiler where we were running a
pass (to split vectors up) that didn't know about indirects before we
lowered the indirects away.
---
...fs-uniform-indexed-by-swizzled-vec4.shader_test | 24 ++
1 file changed, 24 insertions(+)
create
This isn't actually supported by the GL_ARB_copy_image spec.
---
tests/spec/arb_copy_image/formats.c | 17 ++---
1 file changed, 14 insertions(+), 3 deletions(-)
diff --git a/tests/spec/arb_copy_image/formats.c
b/tests/spec/arb_copy_image/formats.c
index 8a5faa7..aa9b4d3 100644
---
usable with copy_image, we should be ok.
Reviewed-by: Ilia Mirkin imir...@alum.mit.edu
Thanks!
On Fri, Mar 13, 2015 at 12:23 AM, Jason Ekstrand ja...@jlekstrand.net
wrote:
This isn't actually supported by the GL_ARB_copy_image spec.
---
tests/spec/arb_copy_image/formats.c | 17
was being created by shader_runner, which in turn meant that
the fp64 extension wasn't being exposed (since it's listed as
core-only).
Good enough for me.
On Thu, Feb 5, 2015 at 12:41 AM, Jason Ekstrand ja...@jlekstrand.net
wrote:
Does fp64 require 1.50? If not, then it seems
Does fp64 require 1.50? If not, then it seems as if the requirement of
having the gpu_shader_fp64 extension available is sufficient. Am I missing
something?
On Wed, Feb 4, 2015 at 7:34 PM, Ilia Mirkin imir...@alum.mit.edu wrote:
Mesa only exposes fp64 on core profiles.
Signed-off-by: Ilia
Fine with me
Reviewed-by: Jason Ekstrand jason.ekstr...@intel.com
On Mon, Jan 19, 2015 at 1:48 PM, Vinson Lee v...@freedesktop.org wrote:
Silence Coverity missing break in switch defects.
Signed-off-by: Vinson Lee v...@freedesktop.org
---
tests/spec/arb_copy_image/targets.c | 2 ++
1
The ARB_gpu_shader5 extension requires GLSL 1.50 and we were only requiring
GLSL 1.30
---
.../spec/arb_gpu_shader5/compiler/indirect-projGradOffset-shadow.frag | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
be used with
GL_RGB, GL_BGR, GL_RGBA, or GL_BGRA and do not mention GL_ABGR_EXT.
Either way,
Reviewed-by: Jason Ekstrand jason.ekstr...@intel.com
+ if (format == GL_ABGR_EXT IsPackedType(datatype))
+ return false;
+
if (format == GL_RG !haveRG)
return
On Tue, Jan 13, 2015 at 2:24 AM, Samuel Iglesias Gonsálvez
sigles...@igalia.com wrote:
On Monday, January 12, 2015 01:14:23 PM Jason Ekstrand wrote:
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
---
tests/spec/arb_gpu_shader5/compiler/indirect-projGradOffset-shadow.frag | 1 +
1 file changed, 1 insertion(+)
diff --git
a/tests/spec/arb_gpu_shader5/compiler/indirect-projGradOffset-shadow.frag
b/tests/spec/arb_gpu_shader5/compiler/indirect-projGradOffset-shadow.frag
index
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
allowed by the spec. NVIDIA allows it but AMD doesn't and our driver
hasn't
On Sat, Jan 10, 2015 at 2:22 AM, Chris Forbes chr...@ijw.co.nz wrote:
On Sat, Jan 10, 2015 at 7:27 PM, Jason Ekstrand ja...@jlekstrand.net
wrote:
This adds a test that uses an indirect offset into an array of 2D shadow
samplers with a textureProjGradOffset. The primary purpse
This adds a test that uses an indirect offset into an array of 2D shadow
samplers with a textureProjGradOffset. The primary purpse of this is to
use a lot of arguments with an indirect offset. This hits a bug in the
current NIR code where we originally assumed a max of 4 texture arguments.
---
---
tests/all.py | 3 +++
tests/texturing/teximage-colors.c | 54 ---
2 files changed, 42 insertions(+), 15 deletions(-)
diff --git a/tests/all.py b/tests/all.py
index 87ff8e9..50d4b39 100644
--- a/tests/all.py
+++ b/tests/all.py
@@
This commit does a couple of things. First, we increase the size of the
texture uploaded while benchmarking to 1024 square to reduce the affect of
set-up overhead. Second, we do a render and then call glFinish at the end
of the upload loop to ensure that things actually get flushed through the
On Fri, Jan 9, 2015 at 2:58 PM, Dylan Baker baker.dyla...@gmail.com wrote:
On Friday, January 09, 2015 02:53:25 PM Jason Ekstrand wrote:
---
tests/all.py | 3 +++
tests/texturing/teximage-colors.c | 54
---
2 files changed, 42
On Fri, Nov 21, 2014 at 11:52 AM, Emil Velikov emil.l.veli...@gmail.com
wrote:
This will allow us to use waffle with its upcoming WGL support for
Windows. With that done, the final step to removing glut is to convert
piglit to use waffle for MacOS.
Current implementation does not have input
That's better
On Dec 13, 2014 10:58 PM, Vinson Lee v...@freedesktop.org wrote:
Silences Coverity report about missing break in switch.
Signed-off-by: Vinson Lee v...@freedesktop.org
---
tests/texturing/teximage-colors.c | 1 +
1 file changed, 1 insertion(+)
diff --git
---
tests/shaders/ssa/fs-swap-problem.shader_test | 19 +++
1 file changed, 19 insertions(+)
create mode 100644 tests/shaders/ssa/fs-swap-problem.shader_test
diff --git a/tests/shaders/ssa/fs-swap-problem.shader_test
b/tests/shaders/ssa/fs-swap-problem.shader_test
new file mode
In particular, we want to test the case
a = ...
if (foo) {
a = ...
} else {
break;
}
... = a;
We had a bug in one of the into-ssa passes that caused the use of a at the
end to use the first definition of a instead of the second.
---
tests/shaders/ssa/fs-if-def-else-break.shader_test |
way by control flow that happens all the time in piglit.
Jason Ekstrand (3):
ssa: Add a test for the classic swap problem in out-of-SSA
translation
ssa: Add a test for the classic lost copy problem in out-of-SSA
translation
ssa: Add a test for an edge case in into-SSA translation
What did you do (besides verifying that the tests pass) to ensure that the
generated result is equivalent to the previous generator? Have you
verified that they generate basically the same code?
On Wed, Nov 12, 2014 at 3:45 PM, Dylan Baker baker.dyla...@gmail.com
wrote:
This rather lengthy
Sure, fine by me
Reviewed-by: Jason Ekstrand jason.ekstr...@intel.com
On Fri, Nov 7, 2014 at 3:11 PM, Vinson Lee v...@freedesktop.org wrote:
Fixes GCC maybe-uninitialized warning.
formats.c:560:11: warning: 'gtype' may be used uninitialized in this
function [-Wmaybe-uninitialized
Reviewed-by: Jason Ekstrand jason.ekstr...@intel.com
I also pushed it.
On Wed, Oct 29, 2014 at 12:50 PM, Mark Janes mark.a.ja...@intel.com wrote:
b8d8cfadcc83fd09c1367821337d473d80bb6607 inadvertenty disabled the
--junit_suffix functionality, which is needed to differentiate results
from
With the formatting fixed:
Reviewed-by: Jason Ekstrand jason.ekstr...@intel.com
On Mon, Oct 13, 2014 at 8:59 AM, Brian Paul bri...@vmware.com wrote:
On 10/08/2014 06:00 PM, srol...@vmware.com wrote:
From: Roland Scheidegger srol...@vmware.com
Some formats had some implied precision which
On Mon, Sep 29, 2014 at 8:14 PM, Jordan Justen jljus...@gmail.com wrote:
On Mon, Sep 29, 2014 at 7:39 PM, Matt Turner matts...@gmail.com wrote:
On Mon, Sep 29, 2014 at 6:45 PM, Jordan Justen jljus...@gmail.com
wrote:
On Mon, Sep 29, 2014 at 5:01 PM, Matt Turner matts...@gmail.com
wrote:
Yeah, I'm OK with this too. At some point it would be good to put a lite
effort into making both formulas valid so that piglit works on both legacy
and post 3.2 contexts. But that's more of a to do list item than a comment
on this patch.
--Jason
On Sep 25, 2014 3:21 PM, Brian Paul
Reviewed-by: Jason Ekstrand jason.ekstr...@intel.com
On Fri, Sep 19, 2014 at 2:45 PM, Chris Forbes chr...@ijw.co.nz wrote:
Signed-off-by: Chris Forbes chr...@ijw.co.nz
---
tests/spec/arb_copy_image/formats.c | 8
1 file changed, 8 insertions(+)
diff --git a/tests/spec
Forgot to reply-all
On Sep 12, 2014 9:05 AM, Jason Ekstrand ja...@jlekstrand.net wrote:
The teximage-colors test that I pushed to piglit a week or two ago takes a
--benchmark flag that bumps the texture size and does the upload 1000 times
and gives you the average time to upload.
--Jason
Since we can't give a 12-bit input, trying to do an exact test doesn't
really make sense.
---
tests/texturing/teximage-colors.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/tests/texturing/teximage-colors.c
b/tests/texturing/teximage-colors.c
index 815ef4d..712e11b
Yeah, on second thought, not quite working... I'll poke at it a bit.
On Thu, Sep 4, 2014 at 10:11 PM, Jason Ekstrand ja...@jlekstrand.net
wrote:
Thanks a bunch Dylan. It's working great for me.
--Jason
On Thu, Sep 4, 2014 at 5:11 PM, Dylan Baker baker.dyla...@gmail.com
wrote
Dylan,
There's one more on line 196 that you're missing. Wrapping that one in
escape_pathname fixes it up nicely. With that change,
Reviewed-by: Jason Ekstrand jason.ekstr...@intel.com
--Jason
On Fri, Sep 5, 2014 at 11:35 AM, Jason Ekstrand ja...@jlekstrand.net
wrote:
Yeah, on second
Signed-off-by: Jason Ekstrand jason.ekstr...@intel.com
Cc: Ilia Mirkin imir...@alum.mit.edu
---
tests/spec/arb_texture_view/sampling-2d-array-as-cubemap.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/spec/arb_texture_view/sampling-2d-array-as-cubemap.c
b/tests/spec
Previously, we only ever worked on level 0. Now we work on levels 1 and 3
so that cross-level copying gets tested.
---
tests/spec/arb_copy_image/formats.c | 92 -
1 file changed, 61 insertions(+), 31 deletions(-)
diff --git
On Sat, Aug 30, 2014 at 4:20 PM, Dylan Baker baker.dyla...@gmail.com
wrote:
This causes problems in the html summary generator, since os.mkdir
doesn't understand that it's a string not a full path (and the
underlying file system probably can't understand that either). Instead
of raising an
This way test names can contain / characters. I to name tests after git
branches and I tend to have git branches named wip/whatever. This prevents
piglit from crashing.
---
framework/summary.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/framework/summary.py
Sorry, I should have been more clear by test name. I mean the name for
the entire run; i.e., piglit run -n 'wip/whatever' quick.py
results_folder. We should do something intelligent here rather than crash
when generating the summary. If we don't like directories, we could mangle
names instead
it's impossible to add a
duplicate name and have piglit run at all.
--Jason Ekstrand
This patch implements a), it introduces 9 regressions on Intel OCL SDK
all
of which are previously hidden fails. No changes on r600.
I don't really use python, so I went for least invasive changes
shell script before passing the name
to piglit but then they're munged in the table labels in the HTML.
--Jason
But, either way you need to put makedirs in a try/except block
try:
os.makedirs(...)
except OSError:
pass
Yup
On Saturday, August 30, 2014 11:49:51 AM Jason
I've had this one kicking around my personal repo for a while. It's nicer
than the equivalent glean test (but doesn't cover quite as much) and is
also useful for benchmarking texture uploads.
On Fri, Aug 22, 2014 at 1:00 PM, Jason Ekstrand ja...@jlekstrand.net
wrote:
---
tests/all.py
on top of the
non-concurrent test. In this case the concurrent test's window would
destroy the front buffer of the nom-concurrent test.
This will only work if we can guarantee that the non-concurrent test
doesn't pop up any windows.
--Jason Ekstrand
If in the future there are tests which really
of just pass/fail
v5: Test copying between different portions of the same image
Signed-off-by: Jason Ekstrand jason.ekstr...@intel.com
---
tests/all.py| 5 +
tests/spec/arb_copy_image/CMakeLists.gl.txt | 1 +
tests/spec/arb_copy_image/formats.c | 889
path
Signed-off-by: Jason Ekstrand jason.ekstr...@intel.com
---
tests/all.py| 6 +
tests/spec/CMakeLists.txt | 1 +
tests/spec/arb_copy_image/CMakeLists.gl.txt | 13 ++
tests/spec/arb_copy_image/CMakeLists.txt| 1 +
tests/spec
of just pass/fail
Signed-off-by: Jason Ekstrand jason.ekstr...@intel.com
---
tests/all.py| 5 +
tests/spec/arb_copy_image/CMakeLists.gl.txt | 1 +
tests/spec/arb_copy_image/formats.c | 851
3 files changed, 857 insertions
---
tests/spec/arb_copy_image/CMakeLists.gl.txt | 1 +
tests/spec/arb_copy_image/api_errors.c | 264
2 files changed, 265 insertions(+)
create mode 100644 tests/spec/arb_copy_image/api_errors.c
diff --git a/tests/spec/arb_copy_image/CMakeLists.gl.txt
Paul and Dylan Baker
Signed-off-by: Jason Ekstrand jason.ekstr...@intel.com
---
tests/all.py| 72 +
tests/spec/arb_copy_image/CMakeLists.gl.txt | 1 +
tests/spec/arb_copy_image/targets.c | 416
3 files changed, 489
One of the greatest difficulties in properly testing multisample textures
is actually generating and working with reference images. This adds a nice
little util function that allows you to upload a multisampled image as if
you were using glTexImage3D.
Signed-off-by: Jason Ekstrand jason.ekstr
of just pass/fail
Signed-off-by: Jason Ekstrand jason.ekstr...@intel.com
---
tests/all.py| 5 +
tests/spec/arb_copy_image/CMakeLists.gl.txt | 1 +
tests/spec/arb_copy_image/formats.c | 851
3 files changed, 857 insertions
One of the greatest difficulties in properly testing multisample textures
is actually generating and working with reference images. This adds a nice
little util function that allows you to upload a multisampled image as if
you were using glTexImage3D.
Signed-off-by: Jason Ekstrand jason.ekstr
This is a simple sanity test for the ARB_copy_image extension. This tests
tests the different combinations of textures and renderbuffers for intputs
and outputs to glCopyTexSubImage.
v2: Incorperate some comments from Brian Paul
Signed-off-by: Jason Ekstrand jason.ekstr...@intel.com
---
tests
1 - 100 of 111 matches
Mail list logo