for stage in const vs fs:
for d in pack unpack:
for type in Snorm Unorm Half:
generate ${stage}-${d}${type}2x16.shader_test
The tests are generated by a new Python script,
gen_builtin_packing_tests.py, and placed into directory
spec/glsl-es-3.00/execution/built-in-functions.
Signed-of
In order to generate tests for the GLSL packing functions (packSnorm2x16
and friends), we must be able to simulate them. This patch adds
a C executable that does exactly that.
I originally tried to write this tool in Python, but manipulating 16-bit
unsigned integers is not one of Python's strength
This series lives on my gles3-glsl-packing branch.
I tested it against my mesa:gles3-glsl-packing branch on Intel gen6 and gen7.
All tests passed. However, the Mesa implementation and tests share much common
code, so thorough review is welcome.
Chad Versace (2):
util: Add utility to simulate GL
In the newly added piglit_gen_ortho_uniform it introduced a
warning due to float *[4] being passed instead of the expected
const GLfloat * as the type of the 4th param. Change the array
of floats to const GLfloat [4][4] and provide the cast.
Signed-off-by: Tom Gall
---
tests/util/piglit-util-gl-
Works fine.
There is one more warning to squash that is the result of the recent
addition of piglit_gen_ortho_uniform.
Tested-by: Tom Gall
On Wed, Jan 9, 2013 at 6:53 PM, Tom Gall wrote:
> Also looks good. I'll test it out this evening, but I'm sure it's fine.
>
> Reviewed-by: Tom Gall
>
> On
Also looks good. I'll test it out this evening, but I'm sure it's fine.
Reviewed-by: Tom Gall
On Wed, Jan 9, 2013 at 2:31 PM, Chad Versace
wrote:
> The function macros in shader_runner_gles_workarounds.h cause many unused
> variable warnings. The function macros have form
> #define glMapBuffe
Looks good!
Reviewed-by: Tom Gall
On Wed, Jan 9, 2013 at 2:31 PM, Chad Versace
wrote:
> Removes the warning:
> shader_runner.c:1078:7: warning: assignment makes pointer from integer
> without a cast [enabled by default]
>
> CC: Tom Gall
> Signed-off-by: Chad Versace
Reviewed-by: Jose Fonseca
- Original Message -
> From: Roland Scheidegger
>
> The logic failed whenever the random number was 0 for formats which
> don't have
> an alpha channel. In this case all of expected/bias/value were 0 but
> the actual
> correct read back value is 1.0 (default va
On 01/09/2013 01:44 PM, srol...@vmware.com wrote:
From: Roland Scheidegger
The logic failed whenever the random number was 0 for formats which don't have
an alpha channel. In this case all of expected/bias/value were 0 but the actual
correct read back value is 1.0 (default value for alpha channe
From: Roland Scheidegger
The logic failed whenever the random number was 0 for formats which don't have
an alpha channel. In this case all of expected/bias/value were 0 but the actual
correct read back value is 1.0 (default value for alpha channel sampling for
integer textures is 1 just like it i
Removes the warning:
shader_runner.c:1078:7: warning: assignment makes pointer from integer
without a cast [enabled by default]
CC: Tom Gall
Signed-off-by: Chad Versace
---
tests/shaders/shader_runner_gles_workarounds.h | 2 +-
1 file changed, 1 insertion(+), 1 delet
The function macros in shader_runner_gles_workarounds.h cause many unused
variable warnings. The function macros have form
#define glMapBuffer(...) UNDEFINED_FUNCTION(glMapBuffer, 0)
and get expanded to the call
unsupported_function("glMapBuffer", 0)
Observe that the original arguments to glMap
Tom Gall writes:
> In this patch series:
>
> Add invalidate-es3-queries_gles2 to all.tests as well as make it
> follow naming conventions for the binary with _gles2 as the suffix.
>
> Add minmax for gles2. Changes based on Eric Anholt's feedback, removes
> a number of queries shouldn't be in the
Brian Paul writes:
> On 01/07/2013 10:17 PM, Matt Turner wrote:
>> On Tue, Dec 18, 2012 at 2:43 PM, Brian Paul wrote:
>>> create mode 100644 tests/spec/arb_es2_compatibility/ge7.frag
>>> create mode 100644 tests/spec/arb_es2_compatibility/ge7.vert
>> [snip]
>>> + * The vertex/fragment shader
On 6 January 2013 06:16, Marek Olšák wrote:
> Since we now have a GL 2.1 driver with MSAA support - r300g, I had to
> port most MSAA tests to GL 2.1, while not affecting the tests that
> *really* need GL 3.
>
> I have verified that the expected and reported precision of the
> accuracy tests wasn'
This test was added back in April (b859d0c) but was never added to the
build or all.tests.
---
Probably doesn't really need code review. Brian, let me know if this
works for you and I'll commit it.
tests/all.tests | 1 +
tests/spec/glsl-1.10/execution/CMakeList
On 9 January 2013 09:30, Brian Paul wrote:
> I happened to notice that the program tests/spec/glsl-1.10/**
> execution/glsl-render-after-**bad-attach.c which Eric added last April
> (b859d0c) isn't getting built.
>
> Adding a piglit_add_executable() line in the directory's CMakeLists.txt
> file d
On 12/21/2012 11:27 AM, Stuart Abercrombie wrote:
> Make GLSL reflect what the shader source specifies.
>
> Remove the explicit #version directives, as these will be inserted based on
> GLSL >=.
>
> This is part of the effort to get closer to GLES support.
> ---
> tests/shaders/glsl-array-compa
On 12/21/2012 11:27 AM, Stuart Abercrombie wrote:
> The only content is GLSL >= 1.10.
>
> This is part of the effort to get closer to GLES support.
> ---
> .../fs-bool-less-compare-false.shader_test |3 +++
> .../fs-bool-less-compare-true.shader_test |3 +++
> .../executi
On 01/09/2013 09:25 AM, Vinson Lee wrote:
From: Vinson Lee
Fixes dead assignment bugs reported by Clang Static Analyzer.
Signed-off-by: Vinson Lee
---
tests/general/stencil-twoside.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/tests/general/stencil-twos
I happened to notice that the program
tests/spec/glsl-1.10/execution/glsl-render-after-bad-attach.c which
Eric added last April (b859d0c) isn't getting built.
Adding a piglit_add_executable() line in the directory's
CMakeLists.txt file didn't work.
-Brian
From: Vinson Lee
Fixes dead assignment bugs reported by Clang Static Analyzer.
Signed-off-by: Vinson Lee
---
tests/general/stencil-twoside.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/tests/general/stencil-twoside.c b/tests/general/stencil-twoside.c
inde
22 matches
Mail list logo
