It's probably noting somewhere that these tests will fail on CL
implementations where the maximum local workgroup size is < 16 (e.g.
CPU implementations that limit to 1 thread per core).
Also, I'm not sure if it really matters, but instead of:
index2 = index + 1;
index2 = index2 == 4 ? 0 : index2
Often, I make changes to Mesa which only affect tests that actually
execute on the GPU. For such changes, there's no real benefit to
running compiler-only/front-end tests, such as glslparsertest and
asmparsertest files. There are a lot of those, so cutting them can save
quite a bit of time when r
On 7 June 2013 11:12, Fabian Bieler wrote:
> Test that glProgramParameter() triggers the required errors.
>
> The geometry shader input primitive type is limited to a subset
> of OpenGL primitive types.
>
> >From the ARB_geometry_shader4 spec (section Errors):
> "The error INVALID_VALUE is genera
On 3 June 2013 05:42, Fabian Bieler wrote:
> This is usefull for core profile OpenGL contexts and geometry shaders.
>
s/usefull/useful/ here too.
Reviewed-by: Paul Berry
> ---
> tests/shaders/shader_runner.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/test
On 3 June 2013 05:39, Fabian Bieler wrote:
> This is usefull for core profile OpenGL contexts and geometry shaders.
>
s/usefull/useful/
With that fixed,
Reviewed-by: Paul Berry
> ---
> tests/util/piglit-util-gl.c | 22 +++---
> 1 file changed, 11 insertions(+), 11 deletions
On 10 June 2013 16:54, Eric Anholt wrote:
> Fabian Bieler writes:
>
> > Add primitives with adjacency and (for extreme future-proofness) patches
> to
> > recognized drawing modes.
> > ---
> > tests/shaders/shader_runner.c | 7 ++-
> > 1 file changed, 6 insertions(+), 1 deletion(-)
> >
> > d
On 10 June 2013 15:26, Fabian Bieler wrote:
> It used to be piglit_width which works, too, as the framebuffer is
> hardcoded
> to 250x250 pixels.
> ---
> tests/shaders/shader_runner.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
s/heigth/height/ in the subject.
>
> diff --git a/t
The old name existed simply as a holdover until the Summary name could
be cleared up. Since it is now available, use the Summary name.
Signed-off-by: Dylan Baker
---
framework/summary.py| 4 ++--
piglit-summary-html.py | 2 +-
piglit-summary-junit.py | 2 +-
piglit-summary.py | 2 +-
Since nothing now depends on the old summary backend, delete it. This
first eliminates dead code, but also ensures that no new functionality
becomes dependent on this code.
Signed-off-by: Dylan Baker
---
framework/summary.py | 232 +--
1 file chang
Since -s/--summary will always override -d/--diff don't allow them to be
called together. While this may seem a little confusing, the idea of
asking for --diffs and then not getting them is a bug in the original
implementation.
Signed-off-by: Dylan Baker
---
piglit-summary.py | 10 --
1
This changes piglit-summary.py to use the new text generation of the
summary backend, instead of the old.
Signed-off-by: Dylan Baker
---
piglit-summary.py | 64 +--
1 file changed, 10 insertions(+), 54 deletions(-)
diff --git a/piglit-summary.
These two new methods allow the new summary backend to generated an
equivalent output to the old summery backend output of
piglit-summary.py.
Signed-off-by: Dylan Baker
---
framework/summary.py | 42 ++
1 file changed, 42 insertions(+)
diff --git a/framew
Argparse requires less code, and is much easier to read and maintain.
Signed-off-by: Dylan Baker
---
piglit-summary.py | 94 +--
1 file changed, 29 insertions(+), 65 deletions(-)
diff --git a/piglit-summary.py b/piglit-summary.py
index 5169cc4
This field is required by the JUnit xsd, and was not provided. With this
change the JUnit xml being produced is valid.
xsd source: https://svn.jenkins-ci.org/trunk/hudson/dtkit/dtkit-format/
dtkit-junit-model/src/main/resources/com/thalesgroup/dtkit/junit/model/
xsd/junit-4.xsd
Signed-off-by: Dyl
Since this is no longer in use, remove it
Signed-off-by: Dylan Baker
---
framework/junit.py | 377 -
1 file changed, 377 deletions(-)
delete mode 100644 framework/junit.py
diff --git a/framework/junit.py b/framework/junit.py
deleted file mode
Stop using the old JUnit generation code, and start using the new code.
Signed-off-by: Dylan Baker
---
piglit-summary-junit.py | 84 +++--
1 file changed, 4 insertions(+), 80 deletions(-)
diff --git a/piglit-summary-junit.py b/piglit-summary-junit.py
This method is about 50% faster than the existing implementation which
uses junit.py.
Signed-off-by: Dylan Baker
---
framework/summary.py | 55
1 file changed, 55 insertions(+)
diff --git a/framework/summary.py b/framework/summary.py
index 40
Splits the code that finds problems, regressions, fixes, changes, and
skips out of the constructor. Instead of incurring the overhead of
calculating these every time whether they are needed or not the code can
now be smarter about it's path, only generating the ones it needs, if
any.
Signed-off-by
Using mako as an alternative to the junit.py file has a couple of
advantages: First there is a lot less code involved, mako just does most
of the work for us. Second, it works in almost the same way as the HTML
generation.
Signed-off-by: Dylan Baker
---
templates/junit.mako | 11 +++
1 f
There were whitespace errors as a result of the recent tabs to spaces
conversion. This patch corrects those
Signed-off-by: Dylan Baker
---
piglit-summary-junit.py | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/piglit-summary-junit.py b/piglit-summary-junit.py
This new page is able to convey the same information, but takes less
time to generate
Signed-off-by: Dylan Baker
---
framework/summary.py| 21 -
templates/empty_status.mako | 27 +++
2 files changed, 43 insertions(+), 5 deletions(-)
create mod
Replaces self.alltests, self.changes, self.problems, self.regressions,
self.fixes, and self.skipped with self.test, a dictionary containing all
of the same information. This cleans the namespace for the NewSummary
object, as well as allowing for some other code simplification
Signed-off-by: Dylan
It is absolutely silly to generate pages for fixes, changes, and
regressions when there is only one test result is provided. This still
generates empty pages when two results are provided, because some
developers have expressed a work flow that uses these empty pages.
Signed-off-by: Dylan Baker
-
23 matches
Mail list logo