Alicia Boya García [Fri, 15 Mar 2019 17:38:15 +0000 (17:38 +0000)]
validateflow: Fail when a pad is not attached
Previously validateflow tests did not fail when the pad was not
attached.
This was a limitation caused by how the Validate API worked. Before, the
`notify::validate-runner` signal was not emitted until a monitor was
attached to the override. This made impossible to listen for the
runner's `stopping` signal.
This patch fixes the problem by setting `validate-runner` for all
existing overrides when the runner is initialized and adding checks in
validateflow to error in the case no pad was attached.
Thibault Saunier [Sun, 10 Mar 2019 20:07:08 +0000 (17:07 -0300)]
validate:launcher: Move '_format_config_template' to the utilities
So it can be reused by GES tests
Thibault Saunier [Sun, 10 Mar 2019 20:06:13 +0000 (17:06 -0300)]
validate: Check that position <= duration from the pipeline monitor
We should not require a scenario for that check to happen
Thibault Saunier [Sun, 10 Mar 2019 20:05:15 +0000 (17:05 -0300)]
validate:scenario: Cleanup spurious error message
Thibault Saunier [Sun, 10 Mar 2019 20:03:09 +0000 (17:03 -0300)]
validate-flow: Add an ignored-event-fields configuration
It replaces `record-stream-id` as it is a more generic way of doing
the same thing.
Thibault Saunier [Thu, 7 Feb 2019 18:50:26 +0000 (15:50 -0300)]
docs: Document the validateflow plugin
Edward Hervey [Wed, 6 Mar 2019 09:15:21 +0000 (10:15 +0100)]
validate: More usage of g_strcmp0 instead of strcmp
Edward Hervey [Wed, 6 Mar 2019 09:11:02 +0000 (10:11 +0100)]
validateflow: Use glib variant of strcmp
safer, and avoids missing include
Edward Hervey [Mon, 4 Mar 2019 14:20:49 +0000 (15:20 +0100)]
validate: Include config.h before anything else
To avoid double-defines (such as GST_LEVEL_DEFAULT)
Tim-Philipp Müller [Mon, 4 Mar 2019 09:15:11 +0000 (09:15 +0000)]
Back to development
Tim-Philipp Müller [Tue, 26 Feb 2019 12:00:40 +0000 (12:00 +0000)]
Release 1.15.2
Tim-Philipp Müller [Tue, 26 Feb 2019 19:42:08 +0000 (19:42 +0000)]
scenarios: don't clobber scenarios in 'make install'
/usr/bin/install: will not overwrite just-created '_inst/share/gstreamer-1.0/validate/scenarios/default-seek-flags.scenario' with '../../../../data/scenarios/rtsp_overrides/includes/default-seek-flags.scenario'
https://gitlab.freedesktop.org/gstreamer/gst-devtools/merge_requests/44#note_123683
Tim-Philipp Müller [Tue, 26 Feb 2019 18:53:28 +0000 (18:53 +0000)]
validate: add new API to exports file
Thibault Saunier [Fri, 1 Feb 2019 13:02:22 +0000 (10:02 -0300)]
validate: Notice that issue summaries should be 'stable'
As they are used in the testsuite to define known issues
Alicia Boya García [Sun, 17 Feb 2019 14:38:53 +0000 (15:38 +0100)]
gst-validate-scenario: Make waits optional in appsrc-push
While in many cases it's desirable to wait for a buffer to be pushed
downstream when using appsrc-push, in some cases this is not possible as
such pushing action is dependent on following actions that would not be
executed if we wait.
An example for this is prerolling:
appsrc ! qtdemux ! video/x-h264 ! decodebin name=dec ! %(videosink)s
description, seek=false, handles-states=true
appsrc-push, target-element-name=appsrc0, file-name="raw_h264.0.mp4"
set-state, state=playing
appsrc-eos, target-element-name=appsrc0
In order for the preroll to occur, both the appsrc needs to push the
buffer and the state needs to reach PLAYING. But `set-state` cannot
finish if the buffer has not been pushed (the state transition does not
finish) and conversely pushing the buffer will not finish until the
state has reached.
Making appsrc-push not wait for the buffer solves this problem. This
patch makes appsrc-push aware of this issue by only waiting for the
buffer to be pushed if the pipeline is in a state that allows buffers to
flow.
Alicia Boya García [Thu, 21 Feb 2019 21:01:24 +0000 (22:01 +0100)]
gst-validate-scenario: Fix (another) race condition in EOS handling
Since gst_validate_action_set_done() is asynchronous, the bus EOS
handler may already be running before the action is actually finished.
This patch ensures that is not a problem.
Charlie Turner [Tue, 19 Feb 2019 17:36:23 +0000 (17:36 +0000)]
validateflow: interpolate folder directories into pipeline descriptions.
Allow 'pipeline' fields in validateflow to be interpolated with
directory keys like $(medias)s.
Guillaume Desmottes [Thu, 7 Feb 2019 14:36:41 +0000 (15:36 +0100)]
validate: allow config to check for minimum buffer frequency on pads
This change allow tests to check performance of elements by checking the
frequency at which buffers are pushed on src pads.
I re-used most of the logic from fpsdisplaysink to compute the
frequency.
We can now uses something like:
GST_VALIDATE_CONFIG='core,min-buffer-frequency=60,target-element-factory-name=v4l2src'
The 'buffer-frequency-start' optional field can be used to ignore the
frequency during the start of the pipeline. This is useful when testing live
pipelines where configuring and setting up elements can take some time slowing
down the first buffers.
Guillaume Desmottes [Mon, 11 Feb 2019 15:07:28 +0000 (16:07 +0100)]
validate: factor out gst_validate_element_matches_target()
Charlie Turner [Mon, 18 Feb 2019 11:05:26 +0000 (11:05 +0000)]
validateflow: Fix double-free on stdout
Thibault Saunier [Mon, 4 Feb 2019 16:19:26 +0000 (13:19 -0300)]
validate: report: Fix the way we print 'repeat' values
Thibault Saunier [Mon, 4 Feb 2019 16:18:04 +0000 (13:18 -0300)]
validate: Force LC_NUMERIC to C as it is required by our expression parser
And... add some expression parser unit tests
Thibault Saunier [Sun, 3 Feb 2019 23:05:36 +0000 (20:05 -0300)]
validate:scenario: Enhance variable implementation
- Stop arbitrarily consider params as ClockTime based on their names
but add a convetion that the `.type` field of the ActionType should
end by `(GstClockTime)` when it is a clock time.
Thibault Saunier [Sun, 3 Feb 2019 23:03:40 +0000 (20:03 -0300)]
validate:scenario: Make gst_validate_action_new public
Mainly so it can be used in unit tests.
Thibault Saunier [Sat, 2 Feb 2019 16:43:35 +0000 (13:43 -0300)]
validate:scenario: Allow using set-vars from numeric expressions
And require them to follow the `$varname` (can't be $(varname) as
parenthesis have another meaning in those expressions).
Still accept "duration" and "position" as varname for backward compat
but update our scenarios anyway.
Thibault Saunier [Sat, 2 Feb 2019 16:32:50 +0000 (13:32 -0300)]
validate: scenario: Parse playback times as we execute the scenario
This way we will be able to use 'set-vars' for it
Thibault Saunier [Sat, 2 Feb 2019 01:54:13 +0000 (22:54 -0300)]
validate:scenario: Enforce a synthax $(varname) to reference variables
This way it is clear that you are using a variable reading the scenario
and we can verify that what the scenario writer intents is to use an
already set variable.
Guillaume Desmottes [Thu, 7 Feb 2019 16:34:56 +0000 (17:34 +0100)]
validate: don't override max-latency if config contains multiple structs
gst_validate_utils_get_clocktime() is resetting the value if it's not
present in the struct so we were overriding it on the next iterations.
Guillaume Desmottes [Fri, 8 Feb 2019 10:46:58 +0000 (11:46 +0100)]
validate: baseclasses: include env variable in logged command
We were missing the env variables in the command written to the log
file, making it impossible to re-run the test later from the logs.
Alicia Boya García [Sat, 9 Feb 2019 00:16:31 +0000 (01:16 +0100)]
gst-validate-scenario: Add GST_VALIDATE_SCENARIO_EOS_HANDLING_LOCK
There was a race in appsrc-push when the pushed buffer caused an EOS.
The EOS event could be handled by the main thread, finishing the test
while the action, executing in the streaming thread, has not finished
yet.
A mutex is now introduced to add mutual exclusion for the two threads so
that an EOS does not cause the termination of the test while the action
is still going.
Thibault Saunier [Thu, 7 Feb 2019 18:42:06 +0000 (15:42 -0300)]
validateflow: Print some indication that the flow checking is happening
Edward Hervey [Mon, 11 Feb 2019 10:09:21 +0000 (11:09 +0100)]
validate: Add autotools support for flow plugin
Thibault Saunier [Sun, 10 Feb 2019 04:23:50 +0000 (01:23 -0300)]
Revert "validate:launcher: Cache the result of meson introspect"
This reverts commit
05ce6d3b92e88341bd9743b021b33ce606c5d1bc.
We can't do that as it breaks meson logic to set envvars
Thibault Saunier [Sat, 9 Feb 2019 20:25:03 +0000 (17:25 -0300)]
validate:launcher: Do not sort tests all the time
Do it once only once it is fully populated
Thibault Saunier [Sat, 9 Feb 2019 20:24:10 +0000 (17:24 -0300)]
validate:launcher: Cache the result of meson introspect
Running it takes quite some time and we can easily cache it.
Thibault Saunier [Sat, 9 Feb 2019 20:23:28 +0000 (17:23 -0300)]
validate:launcher: Add a simple way to profile app
Guillaume Desmottes [Fri, 8 Feb 2019 13:36:56 +0000 (14:36 +0100)]
validate: scenario: not need to use an atomic to handle dropped count
It's all handled from the same thread.
Guillaume Desmottes [Fri, 8 Feb 2019 13:23:15 +0000 (14:23 +0100)]
validate: scenario: fix dropped checking when terminating scenario
We want to early return if either no max value has been set for the
scenario or if we didn't receive any QoS information.
Guillaume Desmottes [Fri, 8 Feb 2019 12:32:12 +0000 (13:32 +0100)]
validate: properly namespace config related checks
Alicia Boya García [Sun, 28 Oct 2018 17:27:22 +0000 (17:27 +0000)]
New validate plugin: validateflow
validateflow can be used to check the buffers and events flowing through
a custom pipeline match an expectation file. This can be used to test
non-regular-playback use cases like demuxers handling adaptive streaming
fragment pushing.
This patch includes also new actions used for these cases:
`appsrc-push`, `appsrc-eos` and `flush` (plus `checkpoint`, which is
only available with validateflow).
Guillaume Desmottes [Thu, 7 Feb 2019 16:04:52 +0000 (17:04 +0100)]
baseclass: add_validate_config: don't use self.proc_env
self.proc_env is created when starting the test but this API can be call
by generator when creating the test.
Guillaume Desmottes [Wed, 6 Feb 2019 17:24:19 +0000 (18:24 +0100)]
validate: allow scenarios to define a max nb of dropped buffers
The 'max-dropped' description field can now be used to specify the max
number of buffers than can be dropped by the QoS system.
Thibault Saunier [Wed, 6 Feb 2019 02:46:40 +0000 (23:46 -0300)]
validate:launcher: Error out in the check testsuite if rebuilding failed
Guillaume Desmottes [Mon, 4 Feb 2019 16:03:01 +0000 (17:03 +0100)]
validate: allow scenarios to define max pipeline latency
The 'max-latency' description field can now be used to specify the max
latency allowed for the running pipeline.
Mathieu Duponchelle [Sat, 2 Feb 2019 00:23:16 +0000 (01:23 +0100)]
Scenarios: override seek flags for RTSP tests
Our RTSP server is not accurate, it makes no sense to perform
accuracy checks on the client-side segments.
Thibault Saunier [Fri, 1 Feb 2019 23:01:40 +0000 (20:01 -0300)]
validate:scenario: Set variables on all action fields
Thibault Saunier [Fri, 1 Feb 2019 22:30:44 +0000 (19:30 -0300)]
validate: Handle G_LOG_ERROR in our glog handler
Thibault Saunier [Fri, 1 Feb 2019 22:03:04 +0000 (19:03 -0300)]
validate:scenario: Rename action 'define-consts' to 'set-vars'
Those are not consts are they can be modified at runtime
Thibault Saunier [Fri, 1 Feb 2019 21:41:07 +0000 (18:41 -0300)]
validate: Cleanup flags/enum_from_string
Mathieu Duponchelle [Fri, 1 Feb 2019 22:08:16 +0000 (23:08 +0100)]
scenario: prioritize SCENARIOS_PATH when including
Thibault Saunier [Fri, 1 Feb 2019 04:24:19 +0000 (01:24 -0300)]
validate: Let people know the testsuite starts running
Now that we do not print infos about successful tests when redirecting.
Thibault Saunier [Fri, 1 Feb 2019 03:08:45 +0000 (00:08 -0300)]
validate: Keep trying to print stack traces when our log hanlder is removed
There are cases where a crash happens after the program ends
Thibault Saunier [Fri, 1 Feb 2019 02:43:28 +0000 (23:43 -0300)]
validate:scenario: Give pipeline position when failling on EOS
Thibault Saunier [Fri, 1 Feb 2019 02:35:50 +0000 (23:35 -0300)]
validate:launcher: Do not print passing tests if not running in a tty
Thibault Saunier [Fri, 1 Feb 2019 01:37:17 +0000 (22:37 -0300)]
validate:scenario: Explicitely mark generated stop() action as such
Thibault Saunier [Wed, 30 Jan 2019 18:57:13 +0000 (15:57 -0300)]
validate:launcher:punittest: Raise an exception if a testsuite can't be loaded
Edward Hervey [Wed, 25 Nov 2015 15:10:50 +0000 (16:10 +0100)]
pad-monitor: Remove unused pad getrange override
Edward Hervey [Wed, 25 Nov 2015 13:20:31 +0000 (14:20 +0100)]
validate: Use g_object_{get|set}_qdata where applicable
This provides a substantial speedup compared to using strings
Mathieu Duponchelle [Wed, 30 Jan 2019 00:24:16 +0000 (01:24 +0100)]
TestsManager: stop displaying blacklisted tests on stdout
It's basically spam, better suited to the debug logs
Thibault Saunier [Tue, 29 Jan 2019 18:59:44 +0000 (15:59 -0300)]
validate:launcher: Never print lines larger than the terminal
Thibault Saunier [Tue, 29 Jan 2019 15:54:01 +0000 (12:54 -0300)]
validate: scenario: Mark action as being executed right before calling ->execute
And make sure that we do not try to execute the following action
In the case the action type leads to the GMainContext to be ieterated
Jordan Petridis [Tue, 11 Dec 2018 09:42:25 +0000 (11:42 +0200)]
validate: Report the full test name in the xunit file
Looks like gitlab prefers this way of representing tests as it
displays only the name field in its junit reports.
Close #32
Thibault Saunier [Sat, 26 Jan 2019 13:27:47 +0000 (10:27 -0300)]
validate:launcher: Make baseclasses.py pep8 compliant
Thibault Saunier [Sat, 26 Jan 2019 01:27:07 +0000 (22:27 -0300)]
validate:launcher: Handle launching launching a sub launcher
If you use validate-launcher in a meson testsuite, those test now
gets integrated as one unique testsuite (with a pretty long namespace).
Thibault Saunier [Sat, 26 Jan 2019 01:13:28 +0000 (22:13 -0300)]
validate:launcher: Refactor the "main" function
- Move the parser code into a `LauncherConfig.create_parser()` method
- Remove the need to pass libsdir to the _TestsLauncher object
- Extract out a `setup_launcher_from_args` function
Thibault Saunier [Sat, 26 Jan 2019 01:09:30 +0000 (22:09 -0300)]
launcher: Move http serveur and xvfb server to the main test runner object
No good reason for it to be in the main function
Thibault Saunier [Sat, 26 Jan 2019 01:06:14 +0000 (22:06 -0300)]
launcher: Cleanup the way we find python test command line
By setting it before the test base class adds the current testsuite name in the classname
Thibault Saunier [Sat, 26 Jan 2019 01:03:57 +0000 (22:03 -0300)]
launcher: Fix test listing in meson VS gstcheck tests
You might select tests that match Meson but not gstcheck in which
case the 'meson only' variant is exposed but those should never exist.
Thibault Saunier [Sat, 26 Jan 2019 12:19:35 +0000 (09:19 -0300)]
validate:launcher: Add `%(config_path)s` in the pipeline desc vars
When defining pipelines_descriptions to run test on in a `.json` file, you might
need to point to paths in the testsuite directory (for media files URIs
for example), you can now do
`"pipeline": "filesrc location="$(config_path)s/../medias/some/file.mkv...`
Tim-Philipp Müller [Thu, 17 Jan 2019 10:01:50 +0000 (10:01 +0000)]
Release 1.15.1
Tim-Philipp Müller [Thu, 17 Jan 2019 09:58:47 +0000 (09:58 +0000)]
Fix distcheck
Work around broken disthook check in release.mak so we don't
have to update the common submodules for that (applies only
to this module because the version number is in the top-level
meson.build but the package/dist directory is a subdir). This
only became a problem now because the common submodule hadn't
been updated for the last few years.
Tim-Philipp Müller [Thu, 17 Jan 2019 09:38:13 +0000 (09:38 +0000)]
win32: update .def file for new API
Fixes distcheck
Thibault Saunier [Tue, 15 Jan 2019 19:52:24 +0000 (16:52 -0300)]
validate:launcher: Display unix nickname of signals leading to test failure
Thibault Saunier [Tue, 15 Jan 2019 19:05:41 +0000 (16:05 -0300)]
validate:launcher: Fix error message about 'crashed' test
Alicia Boya García [Mon, 14 Jan 2019 21:55:35 +0000 (22:55 +0100)]
pre-commit-python: Allow line breaks between binary operators
pre-commit-python overrides the list of ignored Python style errors.
Unfortunately, before this patch the list did not exclude W503 and
W504 (which are otherwise ignored by default).
The consequence of having those two warnings enabled at the same time is
that it's not possible to break lines on binary operators, which is an
unreasonable unintentional restriction:
'validateflow': "validateflow, expectations-dir=\"" +
expectations_dir + "\", actual-results-dir=\"" +
actual_results_dir + "\"",
W504 line break after binary operator
'validateflow': "validateflow, expectations-dir=\""
+ expectations_dir + "\", actual-results-dir=\""
+ actual_results_dir + "\"",
W503 line break before binary operator
This patch excludes W503 so that there is a valid style for breaking
lines on binary operators.
Thibault Saunier [Sat, 12 Jan 2019 18:25:53 +0000 (15:25 -0300)]
validate: Plug newly introduced leak
Thibault Saunier [Sat, 12 Jan 2019 12:59:12 +0000 (09:59 -0300)]
validate: Use filename instead of full path in dotfiles names
Thibault Saunier [Sun, 6 Jan 2019 23:06:30 +0000 (00:06 +0100)]
validate:tests: Fix race in `validate_padmonitor.buffer_outside_segment`
We were using a fakesrc which data flow was potentially breaking the
test.
Thibault Saunier [Sun, 6 Jan 2019 23:05:50 +0000 (00:05 +0100)]
validate: Lower a ERROR message to INFO
This behaviour is totally valid when running unit tests
Thibault Saunier [Thu, 3 Jan 2019 13:15:16 +0000 (14:15 +0100)]
validate:tests: s/discount_buffer/discont_buffer/
Thibault Saunier [Thu, 3 Jan 2019 11:09:09 +0000 (12:09 +0100)]
validate:tests: Fix race in `validate_padmonitor.buffer_before_segment`
We were using a fakesrc which data flow was potentially breaking the
test. Stop using it and remove dead code.
Fixes #34
Guillaume Desmottes [Mon, 17 Dec 2018 09:34:43 +0000 (10:34 +0100)]
validate: fix crash if timeout when media_descriptor is None
Some tests may not have any media_descriptor. If those were failing to
shutdown after EOS we were calling get_protocol() on None.
Alexandru Băluț [Fri, 14 Dec 2018 11:00:18 +0000 (12:00 +0100)]
validate:launcher: Fix error message
Alexandru Băluț [Fri, 14 Dec 2018 10:17:41 +0000 (11:17 +0100)]
validate:launcher: Fix discovery of commands
Matthew Waters [Mon, 10 Dec 2018 02:25:58 +0000 (13:25 +1100)]
gst: allow building static libraries for e.g. Android/iOS
Nirbheek Chauhan [Sun, 9 Dec 2018 12:25:37 +0000 (17:55 +0530)]
meson: Fix build on {cross-,}win{32,64}
We use visual studio module definitions for the list of symbols to
export when targetting Windows. Fixes CI failure:
../validate/tools/gst-validate.c:460: undefined reference to `gst_validate_spin_on_fault_signals'
Thibault Saunier [Fri, 7 Dec 2018 12:05:09 +0000 (09:05 -0300)]
validate:launcher: Do not CK_FORK on our test
Otherwise the process can't cleanly quit on assertion because of the way libcheck runner is implemented
Thibault Saunier [Fri, 7 Dec 2018 12:03:24 +0000 (09:03 -0300)]
validate: launcher: Add a way to retrieve trace without coredumpctl
Simply spnning on segfaults (like gst-launch) and catch that in
the launcher to transform the timeout into a segfault and grab a gdb
backtrace
Thibault Saunier [Fri, 7 Dec 2018 02:17:29 +0000 (23:17 -0300)]
validate:launcher: Do not take CK_DEFAULT_TIMEOUT into account to set the timeout
The timeout is what is set in the meson build definition otherwise we will not behave as 'meson test' which we should avoid
Thibault Saunier [Thu, 6 Dec 2018 18:35:18 +0000 (15:35 -0300)]
validate:launcher: Make failure as <failure> in our xunit reporter
I think it was a mistake to call them <error> as the two notions are
different (we marked failed test as "failures" in the <testuite> node).
Should make gitlab happy with our file!
Sebastian Dröge [Thu, 6 Dec 2018 09:53:10 +0000 (11:53 +0200)]
Automatic update of common submodule
From eb6a86e to 59cb678
Thibault Saunier [Fri, 30 Nov 2018 13:59:51 +0000 (10:59 -0300)]
validate:launcher: Add a GST_VALIDATE_LAUNCHER_MAIN_DIR env variable
This is generally usefull so we do not have to pass -M every time we launch the launcher
And it adds support for nesting launcher calls always respecting the provided main directory
+ Fix some new pep8 errors
Thibault Saunier [Wed, 28 Nov 2018 13:14:35 +0000 (10:14 -0300)]
validate:launcher: Don't about unexisting tests when filtering tests
It was wrong
Thibault Saunier [Wed, 28 Nov 2018 13:11:00 +0000 (10:11 -0300)]
validate:launcher: Fix setting meson tests as "parallel"
Thibault Saunier [Mon, 26 Nov 2018 13:13:22 +0000 (10:13 -0300)]
validate:launcher: Avoid using not yet set variables
And make the file pep8 compliant with latest pep8 checker.
Thibault Saunier [Sun, 25 Nov 2018 14:36:06 +0000 (11:36 -0300)]
validate: launcher: Add a way to load pipeline tests from a scenario
Thibault Saunier [Fri, 23 Nov 2018 00:06:36 +0000 (21:06 -0300)]
validate: Update default testsuite git repository
Thibault Saunier [Sat, 17 Nov 2018 12:48:41 +0000 (09:48 -0300)]
validate: Enhance printing actions with fields
If you have maby field, printed actions where unreadable, clean that
up by adding new lines.
Thibault Saunier [Sat, 17 Nov 2018 12:09:34 +0000 (09:09 -0300)]
validate:scenario: Add an action to 'include' another scenario
This is particularly useful for scenario that define constants
that are used to check video frame checksum for example, we can
now have one single 'scenario' file that defines consts for the
checksum of the frames, and those can be reused everywhere.
Thibault Saunier [Sun, 28 Oct 2018 15:21:38 +0000 (15:21 +0000)]
validate:override: Notify override object when they get attached