bump
On Wednesday, June 25, 2014 04:42:02 PM Dylan Baker wrote:
Python 2.6.6 was the final feature release of the 2.6 branch, it was
released in August 2010. Python 2.6.9 was the last security release, it
was released in October 2013. There is no support from upstream for 2.6,
either for bugs
On Tuesday, July 15, 2014 04:44:31 PM Jordan Justen wrote:
On Tue, Jul 15, 2014 at 7:02 AM, Dylan Baker baker.dyla...@gmail.com
wrote:
[snip]
+else:
+if __debug__:
+print('Warning: piglit.conf not found!\n'
+ '(searching
On Wednesday, July 16, 2014 10:31:45 AM Brian Paul wrote:
The GLSLParserTest, ShaderTest and GleanTest subclasses of the Test
class were prepending the TEST_BIN_DIR variable onto the executable
name (such as glslparsertest, shader_runner, and glean). But
the Test class itself also does that
())
os.chdir('..')
-run._get_config(None)
+core.get_config(None)
nt.ok_(core.PIGLIT_CONFIG.has_section('nose-test'),
msg='$PIGLIT_ROOT not found')
with the test fixed this is:
Reviewed-by: Dylan Baker baker.dyla...@gmail.com
signature.asc
ShaderTestParserException(No GL version set)
That looks good to me
Reviewed-by: Dylan Baker baker.dyla...@gmail.com
signature.asc
Description: This is a digitally signed message part.
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman
On Friday, July 11, 2014 08:14:13 AM Dylan Baker wrote:
This is an updated version of a series I sent out previously to make the
GLSLParserTest constructor more strict in what values it allows.
This version is considerably different than v1, largely because it does
a good deal of refactoring
On Friday, July 11, 2014 11:32:20 AM Dylan Baker wrote:
This series has a number of small cleanups for Test and it's derived
classes.
The only significant change in this series is the change to remove
implicitley setting the piglit platform, but always setting it
explicitely in the python
On Friday, July 11, 2014 11:32:20 AM Dylan Baker wrote:
This series has a number of small cleanups for Test and it's derived
classes.
The only significant change in this series is the change to remove
implicitley setting the piglit platform, but always setting it
explicitely in the python
On Friday, July 25, 2014 08:14:22 PM Ilia Mirkin wrote:
Something causes the tests to be generated in
generated_tests/ARB_gpu_shader_fp64 instead of arb_gpu_shader_fp64
like all the other tests (and like all.py expects... although it does
seem to find the tests anyways, weird).
It would be
This series implements improvements for piglit's test handling, mainly
in the python layer. Generally these are just cleanups, additions of
docstrings, formatting changes, renaming long functions, and removing
deprecated code.
The big change behavioraly in this series is that that it moves the
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/exectest.py | 48 +---
1 file changed, 37 insertions(+), 11 deletions(-)
diff --git a/framework/exectest.py b/framework/exectest.py
index dcace25..b4fb341 100644
--- a/framework
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/gleantest.py | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/framework/gleantest.py b/framework/gleantest.py
index 8f547e5..33cab5f 100644
--- a/framework/gleantest.py
+++ b/framework/gleantest.py
@@ -25,16
Test.run() is getting pretty long, splitting this out helps to reduce
that length and make the code more manageable.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/exectest.py | 4
1 file changed, 4 insertions(+)
diff --git a/framework/exectest.py b/framework/exectest.py
This replaces the camel case name with an all caps constant name.
this patch generated by:
find . -name '*py' | xargs sed -i -e 's!globalParams!GLOBAL_PARAMS!g'
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/gleantest.py | 4 ++--
framework/tests
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/exectest.py | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/framework/exectest.py b/framework/exectest.py
index 27b98ef..9f5cf1c 100644
--- a/framework/exectest.py
+++ b/framework/exectest.py
@@ -165,7
This adds a note so that one knows why a specific test skipped.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/exectest.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/framework/exectest.py b/framework/exectest.py
index 52df05b..ca4ddf4 100644
This is shorter (and shorter is always better in a python), and it marks
the method as private (leading __) since it's only for use in a Test
method, not in it's children.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/exectest.py | 4 ++--
1 file changed, 2 insertions(+), 2
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/gleantest.py | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/framework/gleantest.py b/framework/gleantest.py
index 5229cf5..9932ab0 100644
--- a/framework/gleantest.py
+++ b/framework/gleantest.py
@@ -27,13
This allows more state in glean to be shared rather than storing the
same list in memory once for each glean test.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/exectest.py | 8 +---
framework/gleantest.py | 12 +++-
2 files changed, 12 insertions(+), 8 deletions
is not as descriptive as the previous name, but it is shorter and
includes fewer strange symbols.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/programs/run.py | 11 +++
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/framework/programs/run.py b/framework
an error. Third, it is very easy to see
which methods need to be overwritten, even with an editor that folds
code.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/exectest.py| 6 +-
framework/tests/exectest_test.py | 5 -
2 files changed, 5 insertions(+), 6
Gleantest skipping was implemented in PiglitTest, which is not an
ancestor of GleanTest, so it would never be triggered.
It also was set to skip only on 'gbm', however glean only runs on glx,
so if the platform isn't glx then it should be skipped.
Signed-off-by: Dylan Baker baker.dyla
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/gleantest.py | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/framework/gleantest.py b/framework/gleantest.py
index 33cab5f..703eb3f 100644
--- a/framework/gleantest.py
+++ b/framework/gleantest.py
@@ -1,4
behavior.
v2: - Drop xegl name. (Dylan)
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/programs/run.py | 3 ++-
tests/util/piglit-framework-gl/piglit_wfl_framework.c | 4 ++--
2 files changed, 4
From: Chad Versace chad.vers...@linux.intel.com
Because `strcmp(...) == 0` is crufty.
I generated this patch with the coccinelle patch below:
@@
expression E;
@@
-strcmp(env, E) == 0
+streq(env, E)
Signed-off-by: Chad Versace chad.vers...@linux.intel.com
---
.
This does not add skipped results to the new results file, since
presumably the driver has been changed since the last run.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/profile.py | 26 ++
framework/programs/run.py| 7 +++
framework
This had
// expected:pass
But a space is required after the :
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
tests/spec/glsl-1.30/preprocessor/if/if-arg-must-be-defined-01.frag | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
a/tests/spec/glsl-1.30/preprocessor
Previously the code read a StringIO, then immediately created a new one
to write the read text into. This is silly, by just setting the tell
back to position 0 the same effect can be had without the need to read
and write the contents.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
This restricts all keys to this set of values, it is a little lax in
some cases, but represents the lowest common denominator for all of the
keys. This should help to limit the number of errors from authors
creating tests.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework
This enforces a set of valid keys, any key that isn't valid will raise
an exception with a message explaining the problem.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/glsl_parser_test.py | 13 -
framework/tests/glsl_parser_test_tests.py | 31
This splits the creation of the command to be passed to super() into a
private method. This change results in more readable code.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/glsl_parser_test.py | 59 ++-
1 file changed, 36 insertions
These tests conform that the value of command is what is expected based
on the value of the config block.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/tests/glsl_parser_test_tests.py | 35 +++
1 file changed, 35 insertions(+)
diff --git
the constructor for GLSLParserTest is very long, and fairly dense. This
splits the parser section out of the constructor into a private (__
prefixed) method.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/glsl_parser_test.py | 105 --
1
Embarasingly enough v2 was very half-baked. There were a number of tests
that failed and were not fixed, and there were errors that were raised
incorrectly.
This series adds additional tests to catch more failures, and fixes the
various tests that are failing. In each case if tests will fail with
This removes a stray gl_version key, which was set to 3.1. It was never
parsed, but does work because required_gl_version() (in
piglit-util-gl.c) automatically matches glsl version 140 with glt 3.1
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
tests/spec/arb_explicit_attrib_location
On Wednesday, July 30, 2014 09:53:20 PM Ken Phillis Jr wrote:
Patches 16 to 18: these look reasonable.
Patches 1 to 18: did these get tested on windows ( and mac osx )
by any
chance?
No, I don't have access to either windows or osx development
machines. I have compiled this branch with llvm,
On Thursday, August 07, 2014 04:16:39 PM Ben Widawsky wrote:
If users wish to clear their dmesg before a piglit run, the current code
will emit a warning. It is possible to query the information about the
running kernel, and most distros do package that, so default to that.
The live kernel
On Monday, July 28, 2014 03:35:09 PM Dylan Baker wrote:
This series implements improvements for piglit's test handling, mainly
in the python layer. Generally these are just cleanups, additions of
docstrings, formatting changes, renaming long functions, and removing
deprecated code.
The big
On Monday, July 28, 2014 05:34:10 PM Dylan Baker wrote:
Embarasingly enough v2 was very half-baked. There were a number of tests
that failed and were not fixed, and there were errors that were raised
incorrectly.
This series adds additional tests to catch more failures, and fixes the
various
On Wednesday, June 25, 2014 04:42:02 PM Dylan Baker wrote:
Python 2.6.6 was the final feature release of the 2.6 branch, it was
released in August 2010. Python 2.6.9 was the last security release, it
was released in October 2013. There is no support from upstream for
2.6,
either for bugs
Hopefully this is the last iteration of this series, since I actually
have use for it now and I'm not being pre-emptive anymore. I think I've
addressed Ilia's concerns and just went overboard with the locking in
patch 5.
I'd like to land this very soon.
From: Dylan Baker baker.dyla...@gmail.com
This replace _summary, _running, and _percent with _format*. This makes
the methods easier to understand, especially with names like __summary
also used in the class
v2: - rename _print_* to _format_* (Ilia)
Signed-off-by: Dylan Baker baker.dyla
From: Dylan Baker baker.dyla...@gmail.com
Using splat with a dictionary makes sense when the dictionary is already
available (like with locals()), but it isn't very clear.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/log.py | 16 +---
1 file changed, 9 insertions
From: Dylan Baker baker.dyla...@gmail.com
These methods were always called back to back, and required an
overlapping set of arguments. Initially I thought that having a second
method might be useful, but it doesn't really seem that useful at this
point.
Signed-off-by: Dylan Baker baker.dyla
From: Dylan Baker baker.dyla...@gmail.com
This adds a new logger option that produces no output. This is aimed at
headless setups or for systems like jenkins where the output data isn't
very useful.
Signed-off-by: Dylan Baker baker.dyla...@gmail.com
---
framework/log.py | 13
From: Dylan Baker baker.dyla...@gmail.com
This refactors the log module away from a single class that does
everything to having two classes. The first class LogFactory, is a state
manager that returns BaseLog derived objects when it's get() method is
called. However, it mainly acts as a shared
This warning annoys some people, and it isn't strictly necissary, so I
guess it should go.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/core.py | 7 +--
1 file changed, 1 insertion(+), 6 deletions(-)
diff --git a/framework/core.py b/framework/core.py
index d3922a9
On Tuesday, August 12, 2014 11:01:34 PM Emil Velikov wrote:
Hello list,
Recently I've been trying out piglit under Windows, and by default the
cmd
window is 80 columns in width. As such as soon as the status line
becomes
larger than 80 characters, a separate new line is printed for each test
On Wednesday, August 13, 2014 02:16:23 PM Emil Velikov wrote:
On 12/08/14 23:10, Dylan Baker wrote:
On Tuesday, August 12, 2014 11:01:34 PM Emil Velikov wrote:
Hello list,
Recently I've been trying out piglit under Windows, and by default the
cmd
window is 80 columns in width
On Monday, January 02, 2012 05:09:08 PM Matthew Atwood wrote:
From: Matt Atwood matthew.s.atw...@intel.com
Currently while running igt tests a kernel panic causes the results json
file to lose all data. This patch adds a command line option (-s,
--sync) that syncs the file descriptor to disk
to ensure that we
weren't unnecessarily syncing, we don't particularly care about a half done
test case being written, its either all or nothing in terms or whether or not
we care I think.
-Original Message-
From: Dylan Baker [mailto:baker.dyla...@gmail.com]
Sent: Wednesday
[snip]
Also using os.fsync will bypass any data that's still buffered and not
sent out to the underlying API. You probably want a self.file.flush()
in there.
I'm also surprised this works at all. os.fsync should be expecting an
int, so you'd have to use self.file.fileno(). Perhaps there's
From: Dylan Baker baker.dyla...@gmail.com
This adds a new logger option that produces no output. This is aimed at
headless setups or for systems like jenkins where the output data isn't
very useful.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/log.py | 16
From: Dylan Baker baker.dyla...@gmail.com
These methods were always called back to back, and required an
overlapping set of arguments. Initially I thought that having a second
method might be useful, but it doesn't really seem that useful at this
point.
Signed-off-by: Dylan Baker dylanx.c.ba
From: Dylan Baker baker.dyla...@gmail.com
This replace _summary, _running, and _percent with _format*. This makes
the methods easier to understand, especially with names like __summary
also used in the class
v2: - rename _print_* to _format_* (Ilia)
Signed-off-by: Dylan Baker dylanx.c.ba
Makes some locking changes in patch 4, and moves some unit tests to
patch 5 (They really belonged there to begin with)
This is available at my github: https://github.com/dcbaker/piglit
submit/log-refactor
___
Piglit mailing list
From: Dylan Baker baker.dyla...@gmail.com
This refactors the log module away from a single class that does
everything to having two classes. The first class LogFactory, is a state
manager that returns BaseLog derived objects when it's get() method is
called. However, it mainly acts as a shared
From: Dylan Baker baker.dyla...@gmail.com
Using splat with a dictionary makes sense when the dictionary is already
available (like with locals()), but it isn't very clear.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/log.py | 16 +---
1 file changed, 9
On Thursday, August 14, 2014 04:55:04 PM Atwood, Matthew S wrote:
It shouldn't be a big deal, fsync() only blocks while writing to disk,
if there's nothing to write, it wont block.
As to why write_dict_key is public, have a look at
JSONWriter.initialize_json() in the results.py module, I think
On Friday, August 15, 2014 04:12:59 PM Atwood, Matthew S wrote:
Ah, I was wondering what the proper way to do that was, sorry for the
confusion. My go to was to respond to previous version messages as
whether or not the feedback had been added. As far as default behavior
for igt tests, I
On Monday, August 18, 2014 10:44:41 AM Tom Stellard wrote:
---
framework/oclconform.py | 91
+
piglit.conf.example | 34 ++
2 files changed, 125 insertions(+)
create mode 100644 framework/oclconform.py
diff --git
@@ __all__ = ['PIGLIT_CONFIG',
'parse_listfile']
-PIGLIT_CONFIG = ConfigParser.SafeConfigParser()
+PIGLIT_CONFIG = ConfigParser.SafeConfigParser(allow_no_value=True)
def get_config(arg=None):
if arg:
Reviewed-by: Dylan Baker baker.dyla...@gmail.com
--
1.8.1.5
Hi Tom, I have a couple of comments, but this looks pretty good
On Tuesday, August 19, 2014 10:33:42 AM Tom Stellard wrote:
v2:
- Code cleanups
---
Hi Dylan,
Here is an updated patch addressing your comments. However, I was not able
to get the constructor to work using args, kwargs.
This class is useful in more than one module, so moving it to utils,
which is shared, is useful.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/programs_tests.py | 61 +++
framework/tests/utils.py | 51
This moves calls to __move_local and __unset_config into setup(), which
is called before each method. This reduces the code duplication even
further.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/programs_tests.py | 14 +++---
1 file changed, 3 insertions(+), 11
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/programs/run.py | 2 --
1 file changed, 2 deletions(-)
diff --git a/framework/programs/run.py b/framework/programs/run.py
index ff75afe..c4c9189 100644
--- a/framework/programs/run.py
+++ b/framework/programs/run.py
@@ -247,5
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/programs_tests.py | 22 ++
1 file changed, 10 insertions(+), 12 deletions(-)
diff --git a/framework/tests/programs_tests.py
b/framework/tests/programs_tests.py
index a395b76..2c114b4 100644
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/programs_tests.py | 1 -
1 file changed, 1 deletion(-)
diff --git a/framework/tests/programs_tests.py
b/framework/tests/programs_tests.py
index 3d61460..cd706c7 100644
--- a/framework/tests/programs_tests.py
+++ b/framework
This is groundwork to allow some additional code sharing between the
tests.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/programs_tests.py | 17 +
1 file changed, 5 insertions(+), 12 deletions(-)
diff --git a/framework/tests/programs_tests.py
b
This patch adds support to piglit.conf to set a default platform.
---
framework/programs/run.py | 44 +++---
framework/tests/run_parser_tests.py | 117 +++-
piglit.conf.example | 8 +++
3 files changed, 148 insertions(+), 21
This adds a couple of tests for the run parser to test behavior, these
tests will be used to confirm stable behavior in the next patch of the
series.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/run_parser_tests.py | 58 +
1 file
This change is necessary to be able to regression test the parser, which
has grown quite large and complicated. The unit testing of the parser is
necessitated by changes coming later in this series.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/programs/run.py | 15
This is a stupid bug, the test doesn't actually call the function to be
tested. Adding the function to actually test fixes the issue.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/programs_tests.py | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
This constant is only used for these tests, and this will make the next
patch cleaner.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/programs_tests.py | 21 +++--
1 file changed, 11 insertions(+), 10 deletions(-)
diff --git a/framework/tests
This series adds support for setting the default platform used by waffle
in piglit.conf. It seems like a rather long series, but the majority of
the series is either refactoring code to allow changes, or working on
the unit tests, (In fact, 11 of the 18 patches in this series much with
unit
-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/programs/run.py | 25 -
1 file changed, 16 insertions(+), 9 deletions(-)
diff --git a/framework/programs/run.py b/framework/programs/run.py
index 78a905f..ced5783 100644
--- a/framework/programs/run.py
+++ b/framework
If a config file doesn't have a section for the external suites they
will raise an exception. This causes the tests to skip if there is a
NoSectionError
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/tests/integration_tests.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion
From: Dylan Baker baker.dyla...@gmail.com
Using splat with a dictionary makes sense when the dictionary is already
available (like with locals()), but it isn't very clear.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/log.py | 16 +---
1 file changed, 9
From: Dylan Baker baker.dyla...@gmail.com
This refactors the log module away from a single class that does
everything to having two classes. The first class LogFactory, is a state
manager that returns BaseLog derived objects when it's get() method is
called. However, it mainly acts as a shared
From: Dylan Baker baker.dyla...@gmail.com
This adds a new logger option that produces no output. This is aimed at
headless setups or for systems like jenkins where the output data isn't
very useful.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/log.py | 16
This version only modifies patch 4 from the previous series, with
courser locking.
I've forced pushed this to github:
https://github.com/dcbaker/piglit submit/log-refactor
___
Piglit mailing list
Piglit@lists.freedesktop.org
From: Dylan Baker baker.dyla...@gmail.com
This replace _summary, _running, and _percent with _format*. This makes
the methods easier to understand, especially with names like __summary
also used in the class
v2: - rename _print_* to _format_* (Ilia)
Signed-off-by: Dylan Baker dylanx.c.ba
.
If the generator is producing an empty requirement field we should fix
the generator to not do that.
On Mon, Jul 28, 2014 at 5:34 PM, Dylan Baker baker.dyla...@gmail.com wrote:
This replaces string searching with a single regex statement and the use
of a match object to find and write key value pairs
This looks good to me and has my review.
I'll push it as well, thanks.
On Sunday, January 08, 2012 03:11:07 PM Matthew Atwood wrote:
From: Matt Atwood matthew.s.atw...@intel.com
Currently while running igt tests a kernel panic causes the results json
file to lose all data. This patch adds a
Presumably you want to add this to all.py so it gets run.
On Wednesday, August 20, 2014 09:55:33 PM Ilia Mirkin wrote:
Signed-off-by: Ilia Mirkin imir...@alum.mit.edu
---
tests/spec/arb_texture_view/CMakeLists.gl.txt | 1 +
.../sampling-2d-array-as-cubemap.c | 119
On Wednesday, August 20, 2014 12:13:11 PM Michel Dänzer wrote:
From: Michel Dänzer michel.daen...@amd.com
This reverts commit acb824ddc53c446124d88e37db610a4f8c59d56c.
This decreases the runtime of the gpu.py profile from around 15 minutes to
around 12 minutes on my machine, with no change
On Thursday, August 21, 2014 05:04:21 PM Anuj Phogat wrote:
On Tue, Aug 19, 2014 at 8:24 PM, Dylan Baker baker.dyla...@gmail.com wrote:
On Tuesday, August 19, 2014 03:22:02 PM Anuj Phogat wrote:
This patch causes no value assigned to 'require_extensions' field
in *.vert and *.frag files
.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/glsl_parser_test.py | 8 +++-
framework/tests/glsl_parser_test_tests.py | 22 ++
2 files changed, 21 insertions(+), 9 deletions(-)
diff --git a/framework/glsl_parser_test.py b/framework
(ext_framebuffer_object, 'fbo-cubemap')
--
2.1.0
For what it's worth:
Reviewed-by: Dylan Baker baker.dyla...@gmail.com
___
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit
signature.asc
Description
On Monday, August 25, 2014 06:20:00 PM Jordan Justen wrote:
On Tue, Aug 19, 2014 at 1:24 PM, Dylan Baker baker.dyla...@gmail.com wrote:
This series adds support for setting the default platform used by waffle
in piglit.conf. It seems like a rather long series, but the majority of
the series
Sorry it's taken me so long to get back to this (again), work has
dictated slightly different priorities lately.
On Tuesday, August 19, 2014 05:45:43 PM Ilia Mirkin wrote:
[snip]
No problem. Locking is probably one of the most complicated subjects
out there :) I'm again looking at the final
Just to be clear, with that one instance of pad= inside the lock you're
good to see this land? And did you want to put acked, reviewed, etc on
it?
- Dylan
On Tuesday, August 26, 2014 10:07:18 PM Ilia Mirkin wrote:
On Tue, Aug 26, 2014 at 8:44 PM, Dylan Baker baker.dyla...@gmail.com wrote
From: Dylan Baker baker.dyla...@gmail.com
This refactors the log module away from a single class that does
everything to having two classes. The first class LogFactory, is a state
manager that returns BaseLog derived objects when it's get() method is
called. However, it mainly acts as a shared
On Wednesday, August 27, 2014 09:21:54 PM Ilia Mirkin wrote:
No, not quite. But this works (at least in principle, will have to
have another look though).
What I was suggesting was simply removing all locking from the _print
function and asserting that the lock was held instead (since all
Since fsync is not relevant to all backends, changing it to file_fsync
makes it fit better into the Backend API
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/programs/run.py| 6 --
framework/results.py | 4 ++--
framework/tests/results_tests.py | 2
This changes initialize_json and close_json to be more like the
constructor and finalize methods of Backend.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/programs/run.py | 9 +
framework/results.py | 12
2 files changed, 13 insertions(+), 8
This series adds support to piglit for multiple results backends, and
implements a JUnit XML backend as the first consumer of this feature.
JUnit is a standard XML descriptive format used by jenkins for test
reporting. Currently piglit has a script that is able to convert
piglit's json output
as JSONWriter provides. This will make writing backends
simpler.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/results.py | 74
1 file changed, 74 insertions(+)
diff --git a/framework/results.py b/framework/results.py
index efc7029
This patch removes external calls to JSONWriter methods that are not
part of the Backend API (or closely equivalent), since those methods
will be changed from public to protected.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/programs/run.py | 14 +-
framework
Use an RLock directly rather than synchronized_self.
This is simple because only the three public methods need to lock,
rather than half a dozen methods.
Signed-off-by: Dylan Baker dylanx.c.ba...@intel.com
---
framework/results.py | 77 +---
1
901 - 1000 of 3305 matches
Mail list logo