Commit graph

640 commits

Author SHA1 Message Date
Thiago Santos
e0d041d67b qa-monitor: implement intercept_report
It is used to iterate over overrides and modify the report level if
the overrides wants to do so.

Also adds a new mutex only for the overrides to avoid deadlocks when
reporting if we used the same lock for iterating the overrides
2013-07-29 12:01:02 -03:00
Thiago Santos
f7d7d093e3 qa-reporter: fix copy n paste left over 2013-07-29 11:35:20 -03:00
Thiago Santos
79e7392a9c qa-reporter: add function for intercepting reports
after report creation, this function is called and implementers can
modify the report to their liking before it is posted to the runner
2013-07-29 11:34:42 -03:00
Vincent Penquerc'h
6e19324584 qa-monitor: chain gst_qa_monitor_finalize to parent's finalize
It was chaining to the parent's dispose.
2013-07-29 10:46:29 -03:00
Thiago Santos
3a13d59a1d gst-qa-report: put the correct format to avoid compiler warnings 2013-07-29 10:06:48 -03:00
Vincent Penquerc'h
2bac08370c i18n: copy necessary files in-tree 2013-07-29 09:57:25 -03:00
Vincent Penquerc'h
1c053ffda7 GstIssueId: make this uintptr_t
As it's used a a placeholder pointer for g_hash_table use,
it needs to be converted back and forth to a pointer.
2013-07-29 09:56:45 -03:00
Thiago Santos
dbc61a88b6 gst-qa-override-registry: adding the override-registry
This registry should contain the list of GstQaOverride to
be used on the pipelines being monitored
2013-07-29 09:53:54 -03:00
Thiago Santos
794ea2652d qa-override: adds qa-override that can change the report level of issues
Useful for customizing the level of issues for particular elements/tests
when they are more relevant or have to be disabled
2013-07-26 12:16:29 -03:00
Thiago Santos
773d8adc5e qa-report: splitting a GstQaReport into a GstQaIssue and GstQaReport
Reports now point to Issues, that are uniquely identified and have
translatable descriptions. This way we are going to be able to uniquely
identify the issues and applications can enable/disable checks for
specific elements.
2013-07-26 12:16:29 -03:00
Thibault Saunier
1e1e0b922e qa: Make it possible to set a scenario from the command line in test apps 2013-07-25 16:58:15 -03:00
Thibault Saunier
382fcc9a01 qa: Properly set reporter's runner reference
That was never set, but it is needed for the reporter to properly add
reports to the runner.
We still keep a reference on the monitor to make things simpler
2013-07-25 16:58:14 -03:00
Thibault Saunier
d804ad475a reporter: Use Gst debugging log in the _report method directly
Using __VALIST__ was not properly working
+ Add a gstqareporter debug category
2013-07-25 16:58:14 -03:00
Thibault Saunier
4fb8f47770 scenario: Implement the GstQaReporter interface and make use of it
This way we can report issues from a scenario
Also add a Seek aread to the known areas list

We now need to pass the runner to the scenario instead of the
pipeline as the GstQaReporter interface needs it.
2013-07-25 16:58:13 -03:00
Thibault Saunier
4992249848 qa: Add a GstQaReporter interface that objects needing reporting can implement
Various type of object should be able to do some reporting, so we have
to make sure all the code to do that is in one place. Creating an interface
makes it simple to share information and it avoid to have a baseclass for
something that is not actually important enough to create a baseclass.

Conflicts:
	gst/qa/gst-qa-pad-monitor.c
2013-07-25 16:57:44 -03:00
Thibault Saunier
7508e35b7e qa: Add a GstQaScenario class making it possible to execute scenarios
A scenario correspond to a suite of action to execute on a pipeline,
for the time being, we only support seeking the pipeline, but in the
future we can imagine doing some queries, setting pipeline state, etc...

The scenario can be loaded thanks to the GST_QA_SCENARIO environment
variable, making it usable with any existant application, in case, the
application can be used interactively, the user should either, not load
any scenario or let the application run without interacting with it.
2013-07-25 16:56:40 -03:00
Thiago Santos
4a3f06885a pad-monitor: add lots of locking
When handling elements that spawn multiple threads (hardware
enc/decoders), the pad monitor has to protect its variables specially
because some checks involve iterating over internally linked pads to
add/get some data for comparison (expected events, timestamp ranges,
caps).

Aside from locking its own mutex, the pad monitor can also lock the
parent's mutex when it needs to use data from its internally linked
pads. The locking order should always be parent and then individual
pad-monitor mutexes. This should prevent deadlocks when multiple
pad-monitors from the same element start doing checks at the same time
from different threads.
2013-07-24 16:16:07 -03:00
Thiago Santos
48c49f5071 pad-monitor: remove already solved TODOs 2013-07-24 10:05:31 -03:00
Thiago Santos
887afecc22 pad-monitor: accept unexpected flow return if pad is eos
Track eos event and mark that pad as eos so that checking for the
flow return knows when 'unexpected' is acceptable
2013-07-24 10:05:23 -03:00
Thiago Santos
b5f31f0c8c pad-monitor: improve caps proxying check on getcaps
Only check if fields are proxied for sink getcaps as it is when
downstream restrictions should be proxied. Also improve the
fields comparison to handle single value x multi value
(list/array/range) contain relations.
2013-07-23 15:18:51 -03:00
Thiago Santos
6ebded71e2 pad-monitor: improve type conformance checking for caps
Replace the macro with a more powerful variadic function that can
check for more acceptable types for the same caps.

This removes a few more false positives
2013-07-23 15:16:35 -03:00
Thiago Santos
c9955c14e4 pad-monitor: only expect a new segment if pad is running on push mode
For pull mode, it should just provide the buffers, regardless of getting
a new segment or not
2013-07-23 12:52:22 -03:00
Edward Hervey
d7b2977657 gst-qa: show help and exit when no arguments are provided
Instead of attempting to create empty pipelines and weird things
happening :)
2013-07-23 12:14:26 -03:00
Thiago Santos
04b3517e41 pad-monitor: fix flushes checking
flush events shouldn't fail, so we don't need to rollback when it
returns false from downstream (this is common when downstream is still
not-linked) and it would cause gst-qa to spit false positives.

Also refactor the common event handling for both sink and src event
functions into a common place. Currently we handle flushes the same
for both pad's directions
2013-07-23 12:11:08 -03:00
Thiago Santos
d8fa81eeff pad-monitor: only merge caps if they exist
Downtream can not be linked, so we shouldn't try to merge
NULL caps
2013-07-23 11:51:07 -03:00
Thiago Santos
b58a7e533d configure: add nano version to enable Werror 2013-07-22 20:50:02 -03:00
Thiago Santos
11cc2e6b3b pad-monitor: fix expected setcaps fields comparison
Use the correct structure when getting the GValues and print different
messages for missing and different fields on the setcaps caps
2013-07-22 20:09:35 -03:00
Thiago Santos
363de0e555 pad-monitor: improve timestamp ranges comparison message a little
Show the buffer range that is being compared.
2013-07-22 20:09:07 -03:00
Thiago Santos
e94427e639 pad-monitor: add check for serialized events order
Store expected serialized events and their 'timestamps' to check if
they are pushed in the same order/time as they were received
2013-07-22 15:05:04 -03:00
Thiago Santos
c650034fb4 pad-monitor: add check for setcaps passing audio/video fields
Checks that the common audio/video fields are correctly passed
downstream after a setcaps
2013-07-22 09:50:23 -03:00
Thiago Santos
0aaf23cb0f pad-monitor: add check for getcaps proxying audio/video fields
Checks that the common audio/video fields are correctly proxied by
the elements after a getcaps
2013-07-19 16:52:45 -03:00
Thiago Santos
694935fdea element-monitor: add is_encoder flag
Easy access to knowing if the monitored element is an encoder
2013-07-19 16:52:11 -03:00
Thibault Saunier
ab98bd242e qa-transcoding: Add a binary program to easily test transcoding 2013-07-19 11:16:47 -03:00
Thibault Saunier
799b7a8c3e monitor(s): Avoid trying to disconnect handlers on instances that do not exist anymore 2013-07-19 11:15:54 -03:00
Thibault Saunier
0127706e09 qa-report: Avoid reporting tons of times the exact same issue to users
Some of the issue can be reported once and for all. We are here avoiding to flood the
user with the same information repeated infinitely.
2013-07-19 11:15:53 -03:00
Thibault Saunier
affddfe626 qa-pad-monitor: Do not use gst_private.h 2013-07-19 11:15:52 -03:00
Thiago Santos
3db8c081a9 docs: add design and usage docs 2013-07-19 11:14:39 -03:00
Thiago Santos
a7353e988b pad-monitor: improve timestamp ranges check
Keep the full range stored by the element in the monitor and check
if outgoing timestamps are within that range. It is simple and
should generally work.
2013-07-19 09:57:07 -03:00
Thiago Santos
a11cc64f90 pad-monitor: fix combined flow checks
We can only check if we found a downstream monitor
2013-07-18 14:49:23 -03:00
Thiago Santos
16b1d750c8 pad-monitor: fix caps field type checks
The type is GstValueList and not GArray
2013-07-18 14:49:01 -03:00
Thiago Santos
9d7f9cfa54 pad-monitor: improve report messages with arguments 2013-07-18 14:48:46 -03:00
Thibault Saunier
eebb3522bf qa-report: Pass the whole monitor when creating a report
So we have the proper source name already avalaible and in the future
we might need some more informations about the monitor itself.
2013-07-18 13:43:29 -03:00
Thibault Saunier
dd672acf8b qa-pad-monitor: Properly set target_name with as much info as possible 2013-07-18 13:43:28 -03:00
Thibault Saunier
cb4792a8e5 qa-monitor: Add a target name field that can be used even when the target is freed 2013-07-18 13:43:25 -03:00
Thibault Saunier
d1508ce651 qa-monitor: Make the reference to the target a weak reference 2013-07-18 13:43:23 -03:00
Thiago Santos
d0272be23b qa-report: add debug flags for criticals
Allows the user to enable program abort if a report
is created with a certain level.

Use:
GST_QA=fatal_criticals,fatal_warnings,fatal_issues
2013-07-18 12:22:43 -03:00
Thibault Saunier
4b808aeb81 qa-report: Do not keep a ref to the source but keep its name instead
We currently do not need to access the object source after its creation
but we need to be able to have a usefull for debugging name.
2013-07-18 12:15:24 -03:00
Thibault Saunier
7faac43afb qa-runner: Add a 'report-added' signal
So it is possible to plug into the runner to get information about
what is going from outside of it.
2013-07-18 12:15:24 -03:00
Thibault Saunier
5f62fd2e2a qa-report: Make it a boxed type
And make it refcounted, in 1.0 it should become a GstMiniObject, for
now, it is enough that way.

The goal is to be able to use it in signals
2013-07-18 12:15:24 -03:00
Thibault Saunier
440d20ed00 qa-monitor: Allow detaill message to be in printf format
So we can give proper informations about what is wrong to users
2013-07-18 12:15:24 -03:00