platform/upstream/gstreamer.git
9 years agovalidate: launcher: Use a Queue to test for test completion
Ramiro Polla [Fri, 16 Jan 2015 18:03:07 +0000 (19:03 +0100)]
validate: launcher: Use a Queue to test for test completion

TestManager will use a Queue to track progress for all tests. This
commit implements a queue inside Test to simplify the transition.

Patch 3/4 to make TestManager handle waiting for processes instead of
expecting each Test to do it.

https://bugzilla.gnome.org/show_bug.cgi?id=743063

9 years agovalidate: launcher: Initialize Test start time outside of wait_process
Ramiro Polla [Fri, 16 Jan 2015 18:00:25 +0000 (19:00 +0100)]
validate: launcher: Initialize Test start time outside of wait_process

wait_process will be moved to TestManager, so the values used to track
process update must remain inside Test.

Patch 2/4 to make TestManager handle waiting for processes instead of
expecting each Test to do it.

9 years agovalidate: launcher: Split process_update() out of wait_process()
Ramiro Polla [Fri, 16 Jan 2015 17:57:06 +0000 (18:57 +0100)]
validate: launcher: Split process_update() out of wait_process()

Patch 1/4 to make TestManager handle waiting for processes instead of
expecting each Test to do it.

9 years agovalidate: launcher: Move logfile handling out of Reporter and into Test
Ramiro Polla [Fri, 16 Jan 2015 17:50:38 +0000 (18:50 +0100)]
validate: launcher: Move logfile handling out of Reporter and into Test

This makes each Test handle its own logfile, allowing the Reporter to
work on multiple tests at the same time.

Patch 5/5 to move logfile handling out of Reporter and into Test.

9 years agovalidate: launcher: Remove redundant check
Ramiro Polla [Fri, 16 Jan 2015 18:54:56 +0000 (19:54 +0100)]
validate: launcher: Remove redundant check

self.out is always available when _get_captured() is called.

Patch 4/5 to move logfile handling out of Reporter and into Test.

9 years agovalidate: launcher: Split test log file handling in Reporter
Ramiro Polla [Fri, 16 Jan 2015 17:45:52 +0000 (18:45 +0100)]
validate: launcher: Split test log file handling in Reporter

Patch 3/5 to move logfile handling out of Reporter and into Test.

9 years agovalidate: launcher: Separate Reporter from current Test
Ramiro Polla [Fri, 16 Jan 2015 17:42:19 +0000 (18:42 +0100)]
validate: launcher: Separate Reporter from current Test

Instead of saving the current Test in Reporter for every test, use
function parameters to achieve the same goal.

Patch 2/5 to move logfile handling out of Reporter and into Test.

9 years agovalidate: launcher: Initialize reporter timer before starting all tests
Ramiro Polla [Fri, 16 Jan 2015 17:25:56 +0000 (18:25 +0100)]
validate: launcher: Initialize reporter timer before starting all tests

Patch 1/5 to move logfile handling out of Reporter and into Test.

9 years agovalidate: launcher: Don't wait for processes longer than necessary
Ramiro Polla [Mon, 12 Jan 2015 12:09:33 +0000 (13:09 +0100)]
validate: launcher: Don't wait for processes longer than necessary

9 years agosocket interposer: Be even more platform restrictive.
Mathieu Duponchelle [Tue, 3 Feb 2015 15:48:49 +0000 (16:48 +0100)]
socket interposer: Be even more platform restrictive.

9 years agovalidate: do not compile for android.
Mathieu Duponchelle [Tue, 3 Feb 2015 14:41:01 +0000 (15:41 +0100)]
validate: do not compile for android.

9 years agovalidate: Implement a fault injection library.
Mathieu Duponchelle [Fri, 30 Jan 2015 17:52:57 +0000 (18:52 +0100)]
validate: Implement a fault injection library.

+ And implement a corrupt-socket-recv action
+ Only compile this on Linux, LD_PRELOAD won't work on Windows.

For now the registering of the action is done through
a call to socket_interposer_init, this will get better
when we refactor the action logic.

https://bugzilla.gnome.org/show_bug.cgi?id=743871

9 years agovalidate: tests more issues with caps.
Mathieu Duponchelle [Fri, 23 Jan 2015 01:04:47 +0000 (02:04 +0100)]
validate: tests more issues with caps.

https://bugzilla.gnome.org/show_bug.cgi?id=743387

9 years agovalidate: Add a test case for caps missing field.
Mathieu Duponchelle [Fri, 23 Jan 2015 00:40:59 +0000 (01:40 +0100)]
validate: Add a test case for caps missing field.

+ Make the fake decoder have video/x-raw caps.

https://bugzilla.gnome.org/show_bug.cgi?id=743387

9 years agovalidate: prepare tests for port to tracers backend.
Mathieu Duponchelle [Thu, 22 Jan 2015 21:29:10 +0000 (22:29 +0100)]
validate: prepare tests for port to tracers backend.

https://bugzilla.gnome.org/show_bug.cgi?id=743387

9 years agobin-monitor: add itself as gobject data.
Mathieu Duponchelle [Thu, 22 Jan 2015 21:07:37 +0000 (22:07 +0100)]
bin-monitor: add itself as gobject data.

9 years agovalidate: launcher: Fix test log header output
Ramiro Polla [Tue, 13 Jan 2015 01:32:16 +0000 (02:32 +0100)]
validate: launcher: Fix test log header output

Write log file header before running tests, instead of overwriting the
file afterwards.

https://bugzilla.gnome.org/show_bug.cgi?id=742966

9 years agovalidate: launcher: Always create log files
Ramiro Polla [Thu, 15 Jan 2015 14:32:12 +0000 (15:32 +0100)]
validate: launcher: Always create log files

Create log files even when stdout redirection is enabled.
This commit partially reverts 20c28de.

https://bugzilla.gnome.org/show_bug.cgi?id=742973

9 years agovalidate: launcher: Introduce new parameter for log file redirecting
Ramiro Polla [Thu, 15 Jan 2015 14:26:14 +0000 (15:26 +0100)]
validate: launcher: Introduce new parameter for log file redirecting

Allow log file redirection through the new --redirect-logs parameter.
Keep the old --logs-dir stdout/stderr parameter, but reset to the
default logs directory in that case, and set redirect_logs internally.
This also prevents the creation of an stdout/stderr directory for
writing xunit.xml.

https://bugzilla.gnome.org/show_bug.cgi?id=742973

9 years agopad-monitor: plug caps leak on iterator resync
Thiago Santos [Fri, 9 Jan 2015 17:04:16 +0000 (14:04 -0300)]
pad-monitor: plug caps leak on iterator resync

9 years agopad-monitor: use the same filter caps when querying downstream caps
Thiago Santos [Fri, 9 Jan 2015 15:36:31 +0000 (12:36 -0300)]
pad-monitor: use the same filter caps when querying downstream caps

To avoid comparing the real result that has been filtered against
a much larger caps that contains all possibilities.

9 years agovalidate: tests: disable g_log handler
Thibault Saunier [Tue, 9 Dec 2014 09:09:15 +0000 (10:09 +0100)]
validate: tests:  disable g_log handler

It messes up our own failures counter

And pass test-utils into gst-indent

9 years agotests: padmonitor: disable glog handling
Thiago Santos [Mon, 8 Dec 2014 21:53:55 +0000 (18:53 -0300)]
tests: padmonitor: disable glog handling

It messes up our own failures counter

9 years agopad-monitor: get correct caps to check for proxied fields in caps queries
Thiago Santos [Mon, 8 Dec 2014 20:27:52 +0000 (17:27 -0300)]
pad-monitor: get correct caps to check for proxied fields in caps queries

Elements should proxy the peer element's caps fields and not what they
have currently set on their pads when replying to a caps query

9 years agopad-monitor: Only add pending caps fields for source pads
Thiago Santos [Mon, 8 Dec 2014 20:17:08 +0000 (17:17 -0300)]
pad-monitor: Only add pending caps fields for source pads

As caps events are downstream, caps set travels from sinks to
sources. Adding pending setcaps values to sink pads makes no sense
as when a new caps is set on the sink it would compare with values
currently set on the source pad, causing a critical failure when
renegotiation happens.

9 years agovalidate: launcher: Take the timeout as ref timeout to compute hard_timeout
Thibault Saunier [Mon, 8 Dec 2014 17:23:10 +0000 (18:23 +0100)]
validate: launcher: Take the timeout as ref timeout to compute hard_timeout

when it is provided.

9 years agovalidate: launcher: Set a hard timeout on GstValidate tests if we know the duration
Thibault Saunier [Mon, 8 Dec 2014 14:27:54 +0000 (15:27 +0100)]
validate: launcher: Set a hard timeout on GstValidate tests if we know the duration

9 years agovalidate: launcher: Force kill subprocess when done with them
Thibault Saunier [Mon, 8 Dec 2014 13:37:15 +0000 (14:37 +0100)]
validate: launcher: Force kill subprocess when done with them

Making sure that we do not end up having spurious subprocess around

9 years agolauncher: baseclass: add missing parameter
Thiago Santos [Mon, 8 Dec 2014 11:42:51 +0000 (08:42 -0300)]
launcher: baseclass: add missing parameter

Fixes "NameError: global name 'options' is not defined"

9 years agovalidate: launcher: Properly handle non default main dir
Thibault Saunier [Mon, 8 Dec 2014 09:09:57 +0000 (10:09 +0100)]
validate: launcher: Properly handle non default main dir

for the case of the new testsuite files

9 years agovalidate: launcher: Make the gstvalidate application a python module
Thibault Saunier [Sun, 7 Dec 2014 11:30:25 +0000 (12:30 +0100)]
validate: launcher: Make the gstvalidate application a python module

9 years agovalidate: Remove remaining reference to launcher/apps/validate
Thibault Saunier [Sat, 6 Dec 2014 09:53:37 +0000 (10:53 +0100)]
validate: Remove remaining reference to launcher/apps/validate

It has been removed now.

9 years agovalidate: Avoid assert removing an already removed signal handler
Thibault Saunier [Wed, 3 Dec 2014 10:28:28 +0000 (11:28 +0100)]
validate: Avoid assert removing an already removed signal handler

And, make sure that we set the return value != 0 when we receive
SIGINT

9 years agovalidate: Handle setting the HTTP server local path from testsuites
Thibault Saunier [Tue, 2 Dec 2014 16:32:18 +0000 (17:32 +0100)]
validate: Handle setting the HTTP server local path from testsuites

9 years agovalidate: Rename gst-qa-assets to gst-integration-testsuites
Thibault Saunier [Tue, 2 Dec 2014 14:41:17 +0000 (15:41 +0100)]
validate: Rename gst-qa-assets to gst-integration-testsuites

9 years agovalidate: print execution of set_subtitles actions
Thibault Saunier [Tue, 2 Dec 2014 14:39:09 +0000 (15:39 +0100)]
validate: print execution of set_subtitles actions

9 years agovalidate: Factor out an LauncherConfig class type to handle configurations
Thibault Saunier [Tue, 2 Dec 2014 09:02:09 +0000 (10:02 +0100)]
validate: Factor out an LauncherConfig class type to handle configurations

Allowing us to more simply define default value and expose an API on
top of it

9 years agovalidate: Remove file specific blacklisted tests
Thibault Saunier [Tue, 2 Dec 2014 09:00:42 +0000 (10:00 +0100)]
validate: Remove file specific blacklisted tests

9 years agovalidate: Let the user know when new tests are added, or tests are REMOVED
Thibault Saunier [Sat, 29 Nov 2014 12:43:06 +0000 (13:43 +0100)]
validate: Let the user know when new tests are added, or tests are REMOVED

9 years agovalidate: Add a cleaner API to setup tests in testsuite files
Thibault Saunier [Fri, 28 Nov 2014 23:03:04 +0000 (00:03 +0100)]
validate: Add a cleaner API to setup tests in testsuite files

With the testsuite format you will get a setup_tests(tests_manager,
options) function called for each TestManager.

The function will have the exact same role as with old config
file but with a clean API and not magic global variables.

This implies that we need default blacklist to be directly set
on the TestManager and not on options.blacklisted_test

9 years agovalidate: Add a way to sync all assets, including big ones
Thibault Saunier [Fri, 28 Nov 2014 21:58:09 +0000 (22:58 +0100)]
validate: Add a way to sync all assets, including big ones

9 years agovalidate: Remove the default testsuite implementation
Thibault Saunier [Fri, 28 Nov 2014 21:42:47 +0000 (22:42 +0100)]
validate: Remove the default testsuite implementation

The default testsuite implementation should belong to the default
asset repo where we have the corresponding knowledge.

We should style manage a sensible list of known blacklisted tests,
encoding profiles, and generators in GstValidate itself and allow testsuite
actual implementations to easily use them though the register_default_*
methods.

This allow us to be able to remove the ugly execfile() call.

9 years agovalidate: Make sure to at least listen to GStreamer and GLib g_logs
Thibault Saunier [Thu, 27 Nov 2014 11:11:43 +0000 (12:11 +0100)]
validate: Make sure to at least listen to GStreamer and GLib g_logs

If somewhere else someone is overriding the g_log default handler,
we would not get notified of anything.

9 years agovalidate: Disable coloration of GST_DEBUG logs when we have no-color
Thibault Saunier [Thu, 27 Nov 2014 12:48:17 +0000 (13:48 +0100)]
validate: Disable coloration of GST_DEBUG logs when we have no-color

Do that only when those logs are not saved to a file

9 years agovalidate: Factor out a function to print action types parametters
Thibault Saunier [Wed, 26 Nov 2014 16:50:11 +0000 (17:50 +0100)]
validate: Factor out a function to print action types parametters

+ Remove playback-type from the list and just print it

9 years agopad-monitor: do not enforce caps querying rules for converters
Thiago Santos [Wed, 19 Nov 2014 20:16:02 +0000 (17:16 -0300)]
pad-monitor: do not enforce caps querying rules for converters

Some encoders/decoders can also be converters, do not enforce
caps proxying rules for them

9 years agovalidate: fix typo in documentation
Vineeth T M [Fri, 28 Nov 2014 05:44:12 +0000 (11:14 +0530)]
validate: fix typo in documentation

There are some typing mistakes in gst-validate-launcher --help
Hence fixing the same.

https://bugzilla.gnome.org/show_bug.cgi?id=740833

9 years agovalidate: Handle unlimited tests duration
Thibault Saunier [Tue, 25 Nov 2014 14:35:09 +0000 (15:35 +0100)]
validate: Handle unlimited tests duration

Running full length scenario when the user asks

9 years agovalidate: Already having a monitor is no error
Thibault Saunier [Tue, 25 Nov 2014 14:32:31 +0000 (15:32 +0100)]
validate: Already having a monitor is no error

9 years agovalidate: pre commit hook: Do not try to run pep8 on non python files!
Thibault Saunier [Tue, 25 Nov 2014 14:30:42 +0000 (15:30 +0100)]
validate: pre commit hook: Do not try to run pep8 on non python files!

9 years agovalidate: Add scub_*_seeking_full scenarios
Thibault Saunier [Tue, 25 Nov 2014 14:29:29 +0000 (15:29 +0100)]
validate: Add scub_*_seeking_full scenarios

Which basically do the same thing as scrub_*_seeking but during
throughout the whole duration of the media

9 years agovalidate: Enhance documentation
Thibault Saunier [Fri, 21 Nov 2014 18:35:16 +0000 (19:35 +0100)]
validate: Enhance documentation

9 years agovalidate: Add an option to update all .media_info files
Thibault Saunier [Wed, 19 Nov 2014 16:58:23 +0000 (17:58 +0100)]
validate: Add an option to update all .media_info files

9 years agovalidate:scenario: Execute actions without playback time without a valid position
Thibault Saunier [Sun, 16 Nov 2014 22:05:45 +0000 (23:05 +0100)]
validate:scenario: Execute actions without playback time without a valid position

If the user did not specify any playback time we should be able to
execute actions even if the pipeline can't answer the position query

+ Make simpler to read the conditions of an action execution

9 years agovalidate:scenario: Properly handle ASYNC action execution in the API
Thibault Saunier [Sun, 9 Nov 2014 18:08:52 +0000 (19:08 +0100)]
validate:scenario: Properly handle ASYNC action execution in the API

The ->execute function now return a GstValidateExecuteActionReturn
which can be set as ASYNC in order to tell the scenario that the action
will be executed asynchronously, when the action is done, the caller is
responsible for calling gst_validate_action_set_done(); so that the
scenario keeps going on.

In this commit we make sure that the old API keeps working as
GST_VALIDATE_EXECUTE_ACTION_ERROR == FALSE and
GST_VALIDATE_EXECUTE_ACTION_OK == TRUE

Morevover GstValidateExecuteActionReturn is just a define

API:
    + gst_validate_action_set_done
    + GstValidateExecuteActionReturn

https://bugzilla.gnome.org/show_bug.cgi?id=739854

9 years agovalidate: Add a GstValidateActionTypeFlag flag
Thibault Saunier [Fri, 7 Nov 2014 22:19:59 +0000 (23:19 +0100)]
validate: Add a GstValidateActionTypeFlag flag

Allowing us to define action types more in detail.
Keep backward compatibility, at least with the C API

https://bugzilla.gnome.org/show_bug.cgi?id=739854

9 years agovalidate: Add a GstValidatePipelineMonitor subclass
Thibault Saunier [Fri, 21 Nov 2014 13:01:48 +0000 (14:01 +0100)]
validate: Add a GstValidatePipelineMonitor subclass

We had quite a bit of code dedicated to handled GstPipeline monitoring
inside GstValidateBinMonitor, cleanly split that code into a new object
type

https://bugzilla.gnome.org/show_bug.cgi?id=740704

9 years agovalidate:launcher: Force clock sync for some protocols
Thibault Saunier [Thu, 20 Nov 2014 10:55:45 +0000 (11:55 +0100)]
validate:launcher: Force clock sync for some protocols

In HLS for example, not having clock sync might lead to races and failures
do not test that for now

9 years agovalidate: Don't fail getting master report from a ghostpad without target
Thibault Saunier [Thu, 20 Nov 2014 10:53:34 +0000 (11:53 +0100)]
validate: Don't fail getting master report from a ghostpad without target

9 years agovalidate: Print current position even if we do not know the rate
Thibault Saunier [Wed, 19 Nov 2014 19:05:57 +0000 (20:05 +0100)]
validate: Print current position even if we do not know the rate

That could cause gst-validate-launcher to wrongly concider tests
as timeout

9 years agovalidate:scenario: Force clock sink for scenarios with a pause action
Thibault Saunier [Mon, 17 Nov 2014 10:39:12 +0000 (11:39 +0100)]
validate:scenario: Force clock sink for scenarios with a pause action

9 years agovalidate: Handle wrong paths when listing avalaible apps
Thibault Saunier [Sat, 15 Nov 2014 17:08:42 +0000 (18:08 +0100)]
validate: Handle wrong paths when listing avalaible apps

User can make mistake or we can have an empty path.

9 years agovalidate: Do not check if first buffer running time is 0
Thibault Saunier [Tue, 11 Nov 2014 19:56:04 +0000 (20:56 +0100)]
validate: Do not check if first buffer running time is 0

It can perfectly not be 0, so it makes no sense to check that.

https://bugzilla.gnome.org/show_bug.cgi?id=739965

9 years agovalidate: Add audiomixer test to the default testsuite
Thibault Saunier [Thu, 6 Nov 2014 22:43:47 +0000 (23:43 +0100)]
validate: Add audiomixer test to the default testsuite

9 years agopad-monitor: Give better details about segment mismatch issues
Thibault Saunier [Sat, 1 Nov 2014 08:24:15 +0000 (09:24 +0100)]
pad-monitor: Give better details about segment mismatch issues

9 years agovalidate: Be more precise in issue type for wrong seqnum
Thibault Saunier [Thu, 30 Oct 2014 13:10:33 +0000 (14:10 +0100)]
validate: Be more precise in issue type for wrong seqnum

Depending on the type of event where the bug occurs,
it is not the same issue type. That allows us to have
much precise reports, and better explain the user
where the issue stands.

9 years agovalidate: Fix a few annotation issues
Thibault Saunier [Thu, 16 Oct 2014 15:32:56 +0000 (17:32 +0200)]
validate: Fix a few annotation issues

9 years agovalidate: Add the notion of WAIT_MULTIPLIER for the wait action
Thibault Saunier [Wed, 15 Oct 2014 15:03:48 +0000 (17:03 +0200)]
validate: Add the notion of WAIT_MULTIPLIER for the wait action

Allowing the user to decide to wait more, or less, or even not wait
for the wait action to execute when running scenarios.

9 years agovalidate:scenario: Report an EXECUTION_ERROR on action execution failure
Thibault Saunier [Sun, 28 Sep 2014 20:37:01 +0000 (22:37 +0200)]
validate:scenario: Report an EXECUTION_ERROR on action execution failure

If the action type handles a better error report type, it should just
return TRUE, and report its issue itself.

9 years agovalidate: Do not exit when we can not discover a result file
Thibault Saunier [Fri, 19 Sep 2014 07:13:13 +0000 (09:13 +0200)]
validate: Do not exit when we can not discover a result file

Loggable.error actually exit the process, it is not what we want!

+ Avoid a backtrace

9 years agoscenario: Allow set-property action to work much earlier
Edward Hervey [Mon, 3 Nov 2014 10:50:54 +0000 (11:50 +0100)]
scenario: Allow set-property action to work much earlier

By default an action has no playback-time, this makes it actionable
immediatly.

When no playback-time is set on a set-property action, it will
be activated the moment the element is added in the pipeline.

10 years agovalidate-bin-monitor: Initialize local variable
Edward Hervey [Fri, 31 Oct 2014 15:01:52 +0000 (16:01 +0100)]
validate-bin-monitor: Initialize local variable

Avoids segfaults when freeing them if they didn't get filled in

10 years agovalidate: update pre-commit hook.
Mathieu Duponchelle [Sun, 26 Oct 2014 13:47:12 +0000 (14:47 +0100)]
validate: update pre-commit hook.

+ Allows to run multiple pre-commit hooks.
+ Always relink the hooks on autogen.
+ Run pep8 on commited python files.

https://bugzilla.gnome.org/show_bug.cgi?id=739208

10 years agoapps: gstvalidate.py: fix various pyflakes / uncaught pep8 issues.
Mathieu Duponchelle [Sat, 25 Oct 2014 12:59:49 +0000 (14:59 +0200)]
apps: gstvalidate.py: fix various pyflakes / uncaught pep8 issues.

https://bugzilla.gnome.org/show_bug.cgi?id=739208

10 years agovalidate-launcher: utils: fix various pyflakes / uncaught pep8 issues.
Mathieu Duponchelle [Sat, 25 Oct 2014 12:50:54 +0000 (14:50 +0200)]
validate-launcher: utils: fix various pyflakes / uncaught pep8 issues.

https://bugzilla.gnome.org/show_bug.cgi?id=739208

10 years agovalidate-launcher: main: fix various pyflakes / uncaught pep8 issues.
Mathieu Duponchelle [Sat, 25 Oct 2014 12:49:26 +0000 (14:49 +0200)]
validate-launcher: main: fix various pyflakes / uncaught pep8 issues.

https://bugzilla.gnome.org/show_bug.cgi?id=739208

10 years agovalidate-launcher: loggable: fix various pyflakes / uncaught pep8 issues.
Mathieu Duponchelle [Sat, 25 Oct 2014 12:46:26 +0000 (14:46 +0200)]
validate-launcher: loggable: fix various pyflakes / uncaught pep8 issues.

https://bugzilla.gnome.org/show_bug.cgi?id=739208

10 years agovalidate-launcher: baseclasses: fix various pyflakes / uncaught pep8 issues.
Mathieu Duponchelle [Fri, 24 Oct 2014 12:38:00 +0000 (14:38 +0200)]
validate-launcher: baseclasses: fix various pyflakes / uncaught pep8 issues.

https://bugzilla.gnome.org/show_bug.cgi?id=739208

10 years agovalidate-launcher: pep8ify sources.
Mathieu Duponchelle [Fri, 24 Oct 2014 12:23:52 +0000 (14:23 +0200)]
validate-launcher: pep8ify sources.

https://bugzilla.gnome.org/show_bug.cgi?id=739208

10 years agovalidate-utils: downgrade ERROR to DEBUG.
Mathieu Duponchelle [Thu, 23 Oct 2014 19:43:45 +0000 (21:43 +0200)]
validate-utils: downgrade ERROR to DEBUG.

This function is called in places where it is legit for it
to return NULL.

10 years agolauncher: add a way to specify an application directory.
Mathieu Duponchelle [Thu, 23 Oct 2014 19:36:03 +0000 (21:36 +0200)]
launcher: add a way to specify an application directory.

https://bugzilla.gnome.org/show_bug.cgi?id=739091

10 years agolauncher: Don't implement product-specific TestManagers.
Mathieu Duponchelle [Thu, 23 Oct 2014 19:34:27 +0000 (21:34 +0200)]
launcher: Don't implement product-specific TestManagers.

This manager will be moved in GES.

https://bugzilla.gnome.org/show_bug.cgi?id=739091

10 years agovalidate-launcher: restructure filesystem
Mathieu Duponchelle [Thu, 23 Oct 2014 13:21:14 +0000 (15:21 +0200)]
validate-launcher: restructure filesystem

https://bugzilla.gnome.org/show_bug.cgi?id=739091

10 years agovalidate: Fix compiler warning about implicit enum type conversion
Ramprakash Jelari [Fri, 24 Oct 2014 13:11:30 +0000 (18:41 +0530)]
validate: Fix compiler warning about implicit enum type conversion

gst-validate-reporter.c:119:39: error: implicit conversion from enumeration type
      'GstValidateReportingDetails' to different enumeration type
      'GstValidateInterceptionReturn' [-Werror,-Wenum-conversion]
  GstValidateInterceptionReturn ret = GST_VALIDATE_SHOW_UNKNOWN;
                                ~~~   ^~~~~~~~~~~~~~~~~~~~~~~~~
gst-validate-reporter.c:124:11: error: implicit conversion from enumeration type
      'GstValidateReportingDetails' to different enumeration type
      'GstValidateInterceptionReturn' [-Werror,-Wenum-conversion]
    ret = iface->get_reporting_level (reporter);
        ~ ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
gst-validate-reporter.c:127:10: error: implicit conversion from enumeration type
      'GstValidateInterceptionReturn' to different enumeration type
      'GstValidateReportingDetails' [-Werror,-Wenum-conversion]
  return ret;
  ~~~~~~ ^~~

10 years agobuild: We install all headers system wide for now.
Mathieu Duponchelle [Wed, 22 Oct 2014 12:16:45 +0000 (14:16 +0200)]
build: We install all headers system wide for now.

Will be fixed when the API is deemed stable enough

10 years agovalidate: rename GstValidateReportingLevel.
Mathieu Duponchelle [Tue, 21 Oct 2014 21:31:37 +0000 (23:31 +0200)]
validate: rename GstValidateReportingLevel.

Removes the confusion with GstValidateReportLevel.

Modeled on GstDebugGraphDetails.

10 years agovalidate: Verify that elements always send a segment before pushing EOS
Thibault Saunier [Sat, 18 Oct 2014 16:55:59 +0000 (18:55 +0200)]
validate: Verify that elements always send a segment before pushing EOS

EOS is some kind of data flow and thus a segment event should always be
pushed before the EOS is sent

10 years agovalidate: Properly check that the seqnum of the EOS is always properly set
Thibault Saunier [Sat, 18 Oct 2014 16:53:03 +0000 (18:53 +0200)]
validate: Properly check that the seqnum of the EOS is always properly set

In the pipeline, an EOS should always have the same seqnum of the
previous SEGMENT event that was received. If the segment is the result
of a seek, it should always be the same as the seek seqnum too.

+ (Mathieu Duponchelle): fix reporting and concatenation tests.

10 years agovalidate: Add support for text based override files
Thibault Saunier [Fri, 3 Oct 2014 16:51:17 +0000 (18:51 +0200)]
validate: Add support for text based override files

Allowing user to easily determine the severity of issue
types in a config file

https://bugzilla.gnome.org/show_bug.cgi?id=737852

10 years agovalidate: Remove unused method
Thibault Saunier [Fri, 3 Oct 2014 16:53:42 +0000 (18:53 +0200)]
validate: Remove unused method

gst_media_descriptor_add_frame is not used anywhere

https://bugzilla.gnome.org/show_bug.cgi?id=737852

10 years agovalidate: Move GstStructure file parsing into utils
Thibault Saunier [Fri, 3 Oct 2014 16:42:04 +0000 (18:42 +0200)]
validate: Move GstStructure file parsing into utils

So it can be reused, at least in GstValidate.

10 years agovalidate: report: Simplify the issue ID registering using GQuarks
Thibault Saunier [Thu, 2 Oct 2014 13:34:28 +0000 (15:34 +0200)]
validate: report: Simplify the issue ID registering using GQuarks

+ Remove unused issue types

https://bugzilla.gnome.org/show_bug.cgi?id=737790

10 years agovalidate: Check all buffers when we have the info from MediaDescriptor
Thibault Saunier [Mon, 15 Sep 2014 15:27:54 +0000 (17:27 +0200)]
validate: Check all buffers when we have the info from MediaDescriptor

We now check that each buffer is the expected one for each buffer that
come into the decoder.

+ Fix some minor leaks in test-utils

https://bugzilla.gnome.org/show_bug.cgi?id=736138

10 years agovalidate:media-descriptor-parser: Add a way to create from a string
Thibault Saunier [Thu, 2 Oct 2014 09:27:30 +0000 (11:27 +0200)]
validate:media-descriptor-parser: Add a way to create from a string

So it is simple to make use of it from the testsuite

https://bugzilla.gnome.org/show_bug.cgi?id=736138

10 years agoreport: g_critical are CRITICAL issues!
Thibault Saunier [Wed, 1 Oct 2014 14:24:58 +0000 (16:24 +0200)]
report: g_critical are CRITICAL issues!

https://bugzilla.gnome.org/show_bug.cgi?id=736138

10 years agovalidate: launcher: Fix printing of errors in final report
Thibault Saunier [Wed, 17 Sep 2014 15:32:52 +0000 (17:32 +0200)]
validate: launcher: Fix printing of errors in final report

https://bugzilla.gnome.org/show_bug.cgi?id=736138

10 years agovalidate: Add the possibility to generate media infos with frame descs
Thibault Saunier [Mon, 15 Sep 2014 17:14:27 +0000 (19:14 +0200)]
validate: Add the possibility to generate media infos with frame descs

+ Fix a little issue when the generation fails.

https://bugzilla.gnome.org/show_bug.cgi?id=736138

10 years agovalidate: MediaDescriptors: Add md5sum to buffer informations
Thibault Saunier [Mon, 15 Sep 2014 15:26:23 +0000 (17:26 +0200)]
validate: MediaDescriptors: Add md5sum to buffer informations

In the media descriptor files, we now have the md5sum of the actual
content of encoded buffers so that we can check that the buffer content is
perfectly what is was supposed to be.

+ Fix the check of whether a frame is a keyframe in the string
  comparison (g_ascii_strcasecmp  return 0 if string matches)

https://bugzilla.gnome.org/show_bug.cgi?id=736138

10 years agovalidate: Move some method between GstMediaDescriptorParser and GstMediaDescriptor
Thibault Saunier [Mon, 15 Sep 2014 15:25:14 +0000 (17:25 +0200)]
validate: Move some method between GstMediaDescriptorParser and GstMediaDescriptor

So that method land where they actually belong.

https://bugzilla.gnome.org/show_bug.cgi?id=736138

10 years agovalidate: Add a way to pass a MediaDescriptor around monitors
Thibault Saunier [Mon, 15 Sep 2014 15:22:52 +0000 (17:22 +0200)]
validate: Add a way to pass a MediaDescriptor around monitors

And add an option in gst-validate so that the user can define what
media descriptor file to use.

https://bugzilla.gnome.org/show_bug.cgi?id=736138