Discoverer maintain a referernce on the discoverer object while
the async timeout callback is alive to prevent a potential crash
if the object is freed while the callback is pending.
But if g_main_context is released before calling the timeout callback,
the discoverer pointer which was weak referenced from GESUriClipAssetClass
will not be disposed because the discoverer object is not finalized.
And handle the fact that adding to a layer can fail.
Also plug some leaks in the dispose method (and use the dispose
vmethod instead of finalize as appropriate).
Basically if we do not emit a "duration" change of the clip being
splitted first when executing the 'reverse' operations would lead
to fully overallaping clips.
This is implemented on top of a Tree that represents the whole timeline.
SourceClips can not fully overlap anymore and the tests have been
updated to take that into account. Some new tests were added to verify
that behaviour in greater details
Each timeline element is in a layer (potentially spanning
over several), it is very often useful to retrieve an element
layer priority (from an app perspective more than the element
priority itself as that is a bit of an implementation detail
in the end).
Port tests to it
Auto transition when having 3 overlapping clips in a same point in the
timeline is not supported as we can't handle it in a nice way. Before we
to avoid creating 2 overlapping transitions (which is plain broken in
NLE) were completely disabling `auto-transition` and removing all
auto-transitions in the timeline but this is pretty weird for the end
user. This commit changes and now makes sure 2 transitions are not
created in the same place.
Also cleanup previous test case.
And start handling relocated assets.
Also expose the discoverer callback as a vmethod so that we can
overridde the discoverer when necessary (to handle discovering of
timeline through gesdemux for example)
This reverts commit e939cfebaf.
Class might not be initialized if they were already registered
when ges_init() was called, but were not created until ges_deinit() called.
To support ges_{init/deinit} multiple times in a process,
there should be a method for setting up internal object/table of
GESUriClipAssetClass. because *_class_init() will be called
only once in process lifecycle.
The GstFramePositioner might be finalized before the notify callback
Without this commit,
gst-editing-services / ges_basic / test_ges_timeline_remove_track
can reproduce the case.
Export GES library API in headers when we're building the
library itself, otherwise import the API from the headers.
This fixes linker warnings on Windows when building with MSVC.
Fix up some missing config.h includes when building the lib which
is needed to get the export api define from config.h
Fixes https://gitlab.freedesktop.org/gstreamer/gst-editing-services/issues/42
As we can have effects with several pads and the default ghosting
doesn't allow that.
This way we also filter the pads to ghost to match our track type.
When removing an effect from a clip, first the notify::priority signals
were being emitted for the remaining effects which changed priority, and only
at the end the child-removed signal. Now the child-removed signal is emitted
first.
Now subclassing a ghostpad with an alpha property so that
we can multiply the alpha of the frame positioning meta
and the alpha of that pad, setting it on the compositor pad.
https://bugzilla.gnome.org/show_bug.cgi?id=797169
g_clear_pointer() is now preserving the type of its arguments for the
free function.
ges-xml-formatter.c: In function ‘_dispose’:
ges-xml-formatter.c:1635:7: error: function called through a non-compatible type [-Werror]
(GDestroyNotify) g_hash_table_unref);
/usr/include/glib-2.0/glib/gmem.h:121:8: note: in definition of macro ‘g_clear_pointer’
(destroy) (_ptr); \
^~~~~~~
https://bugzilla.gnome.org/show_bug.cgi?id=797310
Marking them as children properties and properly allow serializing
clips children properties.
This doesn't handle several TrackElement of a same type with
different property values but this require more worked already
marked as fixme to allow specifying full path of elements in the
children properties API.
See https://gitlab.gnome.org/GNOME/pitivi/issues/1687
Until know we were doing it outside of the signal and subclasses didn't
have a chance to know that some assets was relocated.
This is required so that Pitivi can handle proxy delation and relocated
assets.
Required for https://gitlab.gnome.org/GNOME/pitivi/issues/2203
In the rendering case we were getting random issues and often the
pipeline was not be able to preroll as some pad were not linked inside
encodebin.
https://bugzilla.gnome.org/show_bug.cgi?id=795422
Those are the elements he cares about and we should expose their APIs
as is, event if they are not classified as effects. For example if
the user want to use a capsfilter as effect, he should be able to set
its caps.
summary_:
This way the timeline can handle all priorities for the user
making the API simpler to use.
API:
+ ges_timeline_move_layer
reviewers_: Mathieu_Du
Differential Revision: https://phabricator.freedesktop.org/D232
Otherwise the changes won't be reflected in the NLE backend.
This makes speed changes working inside ges-launch-1.0
ges-launch-1.0 +clip /path/to/file i=10 d=5 +effect videorate set-rate 5.0
https://bugzilla.gnome.org/show_bug.cgi?id=794699
In the (now tested) scenario where we have a transition on the right
side of a clip we are splitting, auto transitions can't be created
because we resize the clip after adding the new one, meaning that
there are 3 elements in the "transition zone", we need to force
auto transition creation after the splitting.
Fixes https://gitlab.gnome.org/GNOME/pitivi/issues/2142
We need different export decorators for the different libs.
For now no actual change though, just rename before the release,
and add prelude headers to define the new decorator to GST_EXPORT.
The documentation states that it returns a (transfer full) list
of GESClip but it was returning a (transfer container) list. Make
sure to actually make it (transfer full).
https://bugzilla.gnome.org/show_bug.cgi?id=793874
When setting a new control binding on a track element, the old control
binding (if any) is going to be removed. Make sure the
"control-binding-removed" signal is emitted in this case.
Fixes https://phabricator.freedesktop.org/T7340#95666
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D1842
In playsink the default video queue max size is 3 buffers, which is
sometimes not enough for our use case.
Allow up to 2 seconds of buffered data, giving us more time to do
the transition between clips, and thus avoiding dropping frames in
the sink when bringing up new clip takes too much time.
Differential Revision: https://phabricator.freedesktop.org/D1854
We need to iterate over the previous element from trackelement_iter
to find the first element that is at the moving point. Several
elements can have the same start as the one initiating the move,
and we need to take all of them into account.
Fixes https://phabricator.freedesktop.org/T7819
With two exceptions:
* ges_clip_create_track_elements_func
* ges_uri_clip_set_uri
which were never declared in headers and should always have been static.
They are handled specially when moving the context and having them
part of the context can lead to weird behaviours.
Fixes https://phabricator.freedesktop.org/T7693
The following implementation details where exposed as public symbols:
- _ges_container_get_priority_offset
- _ges_container_set_height
- _ges_container_set_priority_offset
- _ges_uri_asset_cleanup
but it was not correct and that should never have been used outside
GES.
Moving those declarations to the internal header and marking as
internal.
GstDiscoverer objects were leaked by tests making the leaks detector
unusable.
Introduce ges_deinit(), similiar to gst_deinit(), doing some cleanup
before exiting the process.
https://bugzilla.gnome.org/show_bug.cgi?id=776805
We were using the actual mixer pad to release the smart mixer
pad, which seemed to be on purpose, but was not properly handle,
moreover, it is now forbiden to pass a pad not inside a GstElement
when releasing it.
Also properly remove ghost pads from Smart mixer, we were planly
failling at it.
It was making no sense to loose the information about the pspec itself
to retrieve the child associated to it and was failling when we were
forcing the AssociateType::prop synthax
The underlying integer type for enums are implementation defined and may
not be the same size as gint/guint. So implicitly casting from pointers-
to-enum-types to pointers-to-int-types is unsafe. MSVC warns on these.
https://bugzilla.gnome.org/show_bug.cgi?id=774641
In some cases when rippling clip we could get the algo lost because
a transition existed between two clips (for example at the end of c1
and at the begining of c2) but while rippling it would have required
a transition at the end of c2 and beginning of c1, and we were properly
not destroying the old one (as the two clips were in the moving context)
but we were still creating the other transition in the end...
Reviewed-by: Alex Băluț <alexandru.balut@gmail.com>
Differential Revision: https://phabricator.freedesktop.org/D1362
We set TrackElement track type very early when creating effects
so it now uses that information to find TrackElement in clips
by track type.
Reviewed-by: Alex Băluț <alexandru.balut@gmail.com>
Differential Revision: https://phabricator.freedesktop.org/D1370
Computation was not taking into account the fact that the start of
the element being moved could be at the middle of a group and not
necessarily at the start!
Fixes T7544
Reviewed-by: Alex Băluț <alexandru.balut@gmail.com>
Differential Revision: https://phabricator.freedesktop.org/D1282
We were only concidering that we should let the group handle moving
transitions when changing transitions but in fact as soon as a
transition is happenning between two clips that are in a same group
the group properly handles moving the transition, so let the
group do its job.
Fixes T7543
Differential Revision: https://phabricator.freedesktop.org/D1281
GESLayer is now responsible for setting clips priorites. Also
GESClip top effects priorities are now set by the
ges_clip_set_top_effect_index method, the user should never call
ges_timeline_element_set_priority as it will anyway be overriden
by GES itself.
Differential Revision: https://phabricator.freedesktop.org/D1280
All operations should have higher priorites and sources should be
on top of those. We now first set the operations priorities in
a first pass and then stack sources on top of those.
Differential Revision: https://phabricator.freedesktop.org/D1279
In case effects have been added priorites might become wrong,
but until the timeline is not commited, it does not matter.
Make sure all priorities are correct before commiting compositions
Differential Revision: https://phabricator.freedesktop.org/D1277
Fix all tests as we now have 1 priority inside the layer
dedicated to transitions (basically no source clip will
ever have a priority of 0 inside a layer).
Differential Revision: https://phabricator.freedesktop.org/D1276
And simplify the way we start computing children priority
making min_priority already relative to the clip itself.
Differential Revision: https://phabricator.freedesktop.org/D1275
Had to separate timeline_emit_group_added from timeline_add_group
to avoid emitting group-added when the project is being loaded.
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D1302
Fixes a regression where we decode streams twice,
this was introduced when we started creating NLE
object at GESTrackElement construct time.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=769193
And reuse the same previously created element when adding the clip
back to a layer, avoiding losing all setting done on clip children
in that situation
This is a behaviour change but previous behaviour was actually totally
unexpected and people working around that weird behaviour will moste
probably not care about that change
Differential Revision: https://phabricator.freedesktop.org/D1094
It's what introspection.mak does as well. Should
fix spurious build failures on gnome-continuous
(caused by g-ir-scanner getting compiler details
via python which is broken in some environments
so passing the compiler details bypasses that).
To avoid compiler warning when using const string to create a new
GESUriClip as string is not modified and only passed to functions which
take a const string.
https://bugzilla.gnome.org/show_bug.cgi?id=766523
Otherwise if frame positionner is disposed after track element has been
finalized, it will raise a critical message because we will try to
disconnect a signal handler on a freed track element object.
https://bugzilla.gnome.org/show_bug.cgi?id=766525
Those are implemented with the exact same API at the GESTimelineElement
level now, and user of those APIs with high level languages will get the
exact same API.
This is formally an API break but I am sure no one ever used that and
we should make sure the method is removed as soon as possible because
it has no reason to be exposed.
This is the only header which shall be included by user. Otherwise some
language using gir to generate binding, e.g Vala, will includes all
headers files in alphabetical order which causes compilation errors due
to incomplete type.
https://bugzilla.gnome.org/show_bug.cgi?id=765856
Debugging must be configuring first (before any parsing), and then
the types are initialized at the end.
Fixes issues with debugging categories not being available at the
start
Before this patch, NLE and GES did not support NleOperations (respectively
GESEffects) that changed the speed/tempo/rate at which the source plays. For
example, the 'pitch' element can make audio play faster or slower. In GES 1.5.90
and before, an NleOperation containing the pitch element to change the rate (or
tempo) would cause a pipeline state change to PAUSED after that stack; that has
been fixed in 1.5.91 (see #755012 [0]). But even then, in 1.5.91 and later,
NleComposition would send segment events to its NleSources assuming that one
source second is equal to one pipeline second. The resulting early EOS event
(in the case of a source rate higher than 1.0) would cause it to switch stacks
too early, causing confusion in the timeline and spectacularly messed up
output.
This patch fixes that by searching for rate-changing elements in
GESTrackElements such as GESEffects. If such rate-changing elements are found,
their final effect on the playing rate is stored in the corresponding NleObject
as the 'media duration factor', named like this because the 'media duration',
or source duration, of an NleObject can be computed by multiplying the duration
with the media duration factor of that object and its parents (this is called
the 'recursive media duration factor'). For example, a 4-second NleSource with
an NleOperation with a media duration factor of 2.0 will have an 8-second media
duration, which means that for playing 4 seconds in the pipeline, the seek
event sent to it must span 8 seconds of media. (So, the 'duration' of an
NleObject or GES object always refers to its duration in the timeline, not the
media duration.)
To summarize:
* Rate-changing elements are registered in the GESEffectClass (pitch::tempo and
pitch::rate are registered by default);
* GESTimelineElement is responsible for detecting rate-changing elements and
computing the media_duration_factor;
* GESTrackElement is responsible for storing the media_duration_factor in
NleObject;
* NleComposition is responsible for the recursive_media_duration_factor;
* The latter property finally fixes media time computations in NleObject.
NLE and GES tests are included.
[0] https://bugzilla.gnome.org/show_bug.cgi?id=755012
Differential Revision: https://phabricator.freedesktop.org/D276
Read only properties will throw a GLib warning like this
when accessed with "set_child_property":
Warning: g_object_set_property: property 'text-x' of object class 'GstTextOverlay' is not writable
Avoiding all the pending_xx dance and making the code simpler.
This is now possible thanks to the various recent refactoring.
Thanks to that the user is able to set_child_property on objects
that are not in GESTrack yet, as expected.
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D739
Making it possible to create the nleobject right at the creation
of the element.
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D738
In get_property we should return the default values if
we have not created any GESTitleSource yet
(instead of segfaulting).
And fix GESTitleSource default values!
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D737
This way we have informations about the content of the
children as soon as possible.
Most code paths where already ready to handle that as we use it for
copying clips.
Fix framepositionner to properly handle that (it would have broke
with copied clips before).
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D736
Allowing pasting groups paste exactly what had been copied
And not the new version of the contained objects
This technically breaks the C API but this is a new API and I believe
and hope nobody is using it right now.
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D616
it should always have been private
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D617
This commit was causing issue where we were reporting the toplevel
element as an element but that element was actually in another
not serialized group. That is very tricky to handle for end users
as they are not guaranteed the toplevel clips were actually not
contained in another element.
This reverts commit ceb82ba302.
Reviewed-by: Thibault Saunier <thibault.saunier@collabora.com>
Differential Revision: https://phabricator.freedesktop.org/D627
ges-validate.c:237:22: error: implicit conversion from enumeration type
'GESEdge' to different enumeration type 'GESEditMode'
[-Werror,-Wenum-conversion]
GESEditMode edge = GES_EDGE_NONE;
~~~~ ^~~~~~~~~~~~~
ges-validate.c:277:41: error: implicit conversion from enumeration type
'GESEditMode' to different enumeration type 'GESEdge'
[-Werror,-Wenum-conversion]
new_layer_priority, mode, edge, position))) {
^~~~
https://bugzilla.gnome.org/show_bug.cgi?id=759758
Allowing application to force the asset system to recheck if an
asset has been "fixed" and can be used again
API:
+ ges_asset_needs_reload
Differential Revision: https://phabricator.freedesktop.org/D584
