Summary:
Handle the fact that some new features can be added and that means
generated files will not be fully understandable by older versions of
the formatter.
Make sure that we set the format version to 0.2 when we serialize the
GstEncodingProfile.enabled property.
Add some tests around that.
+ Fix a minor bug in the test-utils
+ Add a meta on the projects to tell in what format version a project
has been serialized/parsed back
API:
GES_META_FORMAT_VERSION
Depends on D178
Reviewers: Mathieu_Du
Differential Revision: http://phabricator.freedesktop.org/D184
Summary: It must only be done when the object is commited.
We can do that in constructed though, as the changes will
anyway be commited when the object is added to a composition.
Also update the tests, as we set properties spearately then
check the stop, we can commit the source at its creation without
removing meaning from the tests.
Reviewers: thiblahute
Differential Revision: http://phabricator.freedesktop.org/D84
I got some trouble with
arc land
and I wanted to push the 3 commit coming after this revert as 3
different commits but they ended up being all squash into one single
commit, which is clearly not cool for later bisecting and blaming.
Reverting that commit and re pushing those 3 commits as they were
supposed to be.
This reverts commit 9fe15ef435.
The strategy here is to seek at the new end of the composition. And in
GES we always add a 1ns long gap at the end of the tracks so that all
track have the exact same duration, and we have black frames when the
timeline is empty
They were not serialized until now.
That implies several changes:
* Override GESTimelineElement [start, inpoint, duration] properties in
GESGroup to ensure that those properties are not serialized as they
should not be.
* Rename GESBaseXmlContainer->clips field to
GESBaseXmlContainer->containers as the hashtable now contains Groups
https://bugzilla.gnome.org/show_bug.cgi?id=709148
For example if the size has been serialized in a file, but the user has
not personalized the size, we want that whenever the restriction caps
change the size, the video should take the size of the track
restriction caps.
We know need to keep track of the current positionner.size even if
setting through caps size changes.
https://bugzilla.gnome.org/show_bug.cgi?id=739527
We now seek on ready and thus do not need to do magic trying to seek
the source as soon as possible as we now do it even sooner than soon.
Co-Authored by: Thibault Saunier <tsaunier@gnome.org>
We can not expect a seek event anymore as we are seeking in READY the elements
themselves
+remove actual sinks
Co-Authored by: Thibault Saunier <tsaunier@gnome.org>
Starting from now we will not claim that we support GnlObject that have
several source pads as this is
1- Not true at all;
2- the design of priorities in the GnlComposition tree does not allow that;
3- Not very useful in most of the cases and it complexifies quite a lot the code
in the composition.
Conflicts:
configure.ac
tests/check/Makefile.am
We can't rely on that, in particular now that it does not actually
add its children all the time but only when it is needed (and that
it has an internal bin where actual things happen).
When users (can be formatters) set timeline element names in the
default 'namespace' we need to update our counter to avoid setting
twice the same name on TimelineElements so afterward there is no
problem adding them in the GESTimeline
+ add a testcase to check that new code and fix leaks on the
existing testcases.
+ Sensibly enhance debugs
The priority handling of clip is now handled by GESLayer itself, and
handling clip as a ordered list should be implemented in GESLayer itself
too, this way the user can decide to switch mode at any time instead of
And fix all the tests as we need to wait for the project to be loaded
to check the reference count of the timeline (as we keep a ref on the
timeline in project to later emit "loaded" on idle).
If you want a custom clip then you have to subclass GESClip,
This class was pre historicall and only used for testing purposes, we
have GESTestClip for that.
https://bugzilla.gnome.org/show_bug.cgi?id=706855
The idea is that we should first play until the time we reach the first
position, at that point we PAUSE the pipeline, then, afterward do the
seeks as asked.
If we get the position before the ASYNC DONE, just accept it.
Add a new test that creates a given number of layers. Each layer has the same
assets / clips shifted by a different amount in the timeline. Alpha and volume
properties are different for each layer. This test is similar to the mixer
example in:
http://gist.github.com/MathieuDuponchelle/5736992#file-mixit-py
We should be able to add more clips to each layer, but this example test only
tests mixing 1 clip across 4 layers.
Conflicts:
tests/check/ges/integration.c
This second set of seeking tests performs the seeks in a PAUSED
pipeline. After all seeks are successful, the pipeline is resumed so that the
test does not timeout.
Conflicts:
tests/check/ges/integration.c
- Use g_list_remove_link so that ordering of seeks is not mandatory
- use g_slice allocator for SeekInfo structs
- Fix leak in freeing seek list
- Check for NULL seeks at end of test, otherwise fail and free failed seeks
A Seekinfo structure consists of 2 fields:
- position: the position to seek to
- seeking_position: the position to perform the seek from
Seeks can be appended to a global list e.g. from code:
seeks = g_list_append (seeks, new_seek_info (0.2 * GST_SECOND, 0.6 * GST_SECOND));
seeks = g_list_append (seeks, new_seek_info (1.0 * GST_SECOND, 1.2 * GST_SECOND));
seeks = g_list_append (seeks, new_seek_info (1.5 * GST_SECOND, 1.8 * GST_SECOND));
The get_position callback checks the current position and attempts to perform
the corresponding seek with gst_element_seek_simple
Those are test with real media files, they are run separetely from other
unit tests using the make check-integration command (can be done from
the toplevel directory)
Currently we can end up overflowing the start of others child of our
parent, avoid that making sure we can set our start to what was
requested by the user before actually doing it
+ Add a test
Check if an object rthat has already been freed has been destroyed is not safe.
Add a helper function that uses weak reference to check that objects that are expected
to be destroyed when unrefing an object are actually destroyed.
This exact same method will be needed in GESGroup, so we should have the method
in the common parent class.
API:
- ges_clip_edit
+ ges_container_edit
+ GESContainer->edit vmethod
The GNL API changed to go from a model where user could
enable/disable updates in the composition, which leaded to races
in many places, to a model where any positioning change in the
composition is not directly done but 'cached' and then the user
has to commit those changes so they become effective in the media
processing stack.
The new API in GES is pretty similare and is basically copy
pasting this new design.
We still need to see if in some context it would make sense to add
a mode where we would commit any changes ourself at the end of our
operation for basic use cases.
Removed APIs:
ges_timeline_enable_update
ges_timeline_is_updating
ges_track_enable_update
ges_track_is_updating
New APIs:
ges_track_commit
ges_timeline_commit
+ Fix tests (starting using GESTestClip instead of GESCustomClip)
Now the height is not only growing, but can also go down, as the value
is just simply computed
API:
GESContainer::compute_height virtual method
+ Fix tests (we still need a round of modernisation, making use of
assets where it makes sense)
There is no reason to use those method outside of GES, so remove them,
cleaning the API and making it easier for users.
Removed APIs:
-----------
* ges_clip_create_track_element
* ges_clip_create_track_elements
This way we let the possibility to the user to actually do it, but we avoid him to do it
without knowing it is absolutely not recommanded to.
API:
+ ges_uri_clip_asset_request_sync
... Not the other way round.
+ Add and enhance debugging info on the way
The user should not be responsible for removing the GESTrackElements from
GESTracks, instead, removing it from a GESClip should imply removing
it from any GESTrack it is in.
This patch changes sensibly the behaviour when we remove a
GESTrackElement from a GESTrack, not remoing it from the GESClip it is
in. *But*, users should never remove a GESTrackElement from a GESTrack
anyway. The testsuite has been updated to that new behaviour.
