This was complexifying the implementation for very little gain.
Each source type should ideally have its own API.
In that patch we make it so we do not have to subclass anything
but instead use GESAsset to pass information about how the pipeline
should look like.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/175>
While this is not correct, we can't predict from what thread a
GstElement will be disposed as it might still be referenced by
a GstMessage somewhere which is freed by, any thread.
In this specific case we can assume that GES user will already have
let go his timeline reference and we should not avoid assert in that
specific case as it should be safe to let the timeline be destroyed
at that point.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/175>
Each active non-core child must have a corresponding active core child
in the same track. Therefore, if we de-activate a core child, we also
need to de-activate all the non-core children in the same track.
Similarly, if we activate a non-core child, we need to activate the
corresponding core child as well.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
Make less assumptions about the priority of effects and core elements so
that the code would still work if the priority of an element was set
directly. In particular, the index of a top effect will always be its
position in the effect ordering.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
We first change the duration of the splitted clip, then we add the new
clip to the layer and assign the tracks for its children. Normally, when
a clip is added to a layer it will have its track elements created, if
needed, and then assigned to their tracks. This will fail if any sources
would fully or triple overlap existing sources in the same track.
However, here we were adding the clip to the layer *and* avoiding the
track assignment process and instead setting the tracks explicitly. In
particular, the order was:
+ add new clip to layer with no tracks assigned
+ shrink the split clip
+ assign the tracks for the new clip
This has been changed to:
+ shrink the split clip
+ add new clip to layer with no tracks assigned
+ assign the tracks for the new clip
Thus, the order of events for any users connecting to object signals
will be close to that of adding another clip to the layer.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
Any time a track element is added to a track, we need to check whether
we need to create a new corresponding auto-transition. This simply moves
the code from ges-clip.c to ges-timeline.c, where it is more appropriate.
Moreover, it technically opens the possibility for creating
auto-transitions for track elements in the timeline that have no
corresponding clip.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
Make sure that the priority of an appended layer is the lowest (highest
in value) when appending a layer to the timeline. This change is
important when appending a layer to a timeline, which can easily have a
gap in priorities if a layer has been removed.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
Since a group can only have its priority set whilst it is part of a
timeline, we can simply let the timeline-tree handle the move, which it
can already do, whilst checking that the move would be legal (not break
the timeline configuration). All the group has to do now if update its
priority value if the priority of any of its children changes. It
doesn't even need to keep track of the layer priority offsets.
Also, added a check to ensure added children belong to the same
timeline.
Also moved the sigids from the GObject data to a g_hash_table, which is
clearer.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
It should be sufficient to set the edit flag only on the toplevel, which
allows all of its children to know they are being edited and should not
move in response.
Also, removed some unnecessary setting/checking of this.
Also, supplied the ges_timeline_element_peak_toplevel, which unlike
ges_timeline_element_get_toplevel_parent, does not add a reference to
the toplevel. Some corresponding leaks in auto-transition have been
fixed by using this instead.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
Only emit snapping-ended if we have a valid snap time. Moreover, we
should emit a new snapping-started even if we are snapping at the same
location. This is because a new snap will always correspond to a new edit,
possibly involving different snapping elements, which a user would want
to know about.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
Editing has been simplified by breaking down each edit into a
combination of three basic single-element edits: MOVE, TRIM_START, and
TRIM_END.
Each edit follows these steps:
+ Determine which elements are to be edited and under which basic mode
+ Determine which track elements will move as a result
+ Snap the edit position to one of the edges of the main edited element,
(or the edge of one of its descendants, in the case of MOVE), avoiding
moving elements.
NOTE: in particular, we can *not* snap to the edge of a neighbouring
element in a roll edit. This was previously possible, even though the
neighbour was moving!
+ Determine the edit positions for clips (or track elements with no
parent) using the snapped value. In addition, we replace any edits of
a group with an edit of its descendant clips. If any value would be
out of bounds (e.g. negative start) we do not edit.
NOTE: this is now done *after* checking the snapping. This allows the
edit to succeed if snapping would cause it to go from being invalid to
valid!
+ Determine whether the collection of edits would result in a valid
timeline-configuration which does not break the rules for sources
overlapping.
+ If all this succeeds, we emit snapping-started on the timeline.
+ We then perform all the edits. At this point they should all succeed.
The simplification/unification should make it easier to make other
changes.
Fixes https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/issues/97
Fixes https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/issues/98
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
Simplified keeping the start and the duration of a container/group up to
date with the earliest start of the children and the last end of the
children. The previous logic was spread between ges-group and
ges-container, now all the position handling is in ges-container.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
The duration-limit is the maximum duration that can be set for the clip
given its current children and their properties. If a change in the
children properties causes this to drop below the current duration, it
is automatically capped by this limit.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/169>
If `add_child` and `set_parent` succeed we want to return TRUE, even if
the added element is no longer a child by the end of the method. This is
because some users may call ges_container_remove during `child-added`.
This shouldn't be considered an error.
Made sure that adding a new track only uses select-tracks-for-object for
core children to determine whether a track elements should be added to the
new track or not, and *not* any other track. In particular, there should
be *no* change in the existing tracks of the timeline when adding another
track. Moreover, a new track should not invoke the creation of track
elements for other tracks.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/160>
The property had been deprecated and is unused.
This property is not needed. Any internal time effect that an nleoperation
wraps is itself responsible for converting seek/segment timestamps.
Previously, the ghostpads were performing a rate conversion after the
rate element had already done so, essentially doubling their effect on
seeks and segment times. This was always unnecessary, but went unnoticed
by the tempochange test because it was using an identity element rather
than an actual rate-changing element.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/160>
Avoid loosing (too much) precision when rescaling back and forth by
storing values in gdoubles.
Handle the fact that position values can be negative
Also fix debug category static variable
as it clashes with the instance variable name in a few methods.
Basically when using timeoverlay we where waiting for input-selector
to receive EOS on its active on the output-selector streaming thread
but... EOS was being sent from that same thread waiting for input-selector
to unblock to send EOS on its other pad.
In our specific use case we want EOS to be sent only on the active pad.
Fixes: https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/issues/103
Instead of focusing on the instances of the clips and their children,
we relax the check to allow moving track element clip between clips
that share a common asset. This makes it as correct conceptually but
more flexible, and the code becomes simpler.
Previously, the GESContainer ->paste method and GESGroup ->paste methods
were unnecessarily setting the timeline of groups, even though this is
handled by the GESGroup ->child_added method. This could result in the
group being added multiple times.
Previously, the code was not able to detect that an element overlaps on
its end, nor could it detect that an element overlaps two elements that
already overlap.
When the asset of a uri clip is reset, its core children are removed and
replaced by the new core children. When replacing, the `set_asset`
method attempts to copy children properties from the previous children
to the new children. However, the children were matched by track-type
only. This would not function as intended when a URI contains multiple
audio or video streams. Instead, we now match children by the tracks
themselves. This should work better, provided the user's
select-tracks-for-object is well behaved.
Also, fix a memory problem in `set_mute` for when a child is not in a
track.
Only copy the properties that can be both read and written, and are not
construct only. Similarly for child properties when a track-element is
deep copied.
Technically, an element can still be floating on the return from
`->paste` (e.g. a clip not in a layer). Since the return of the `_paste`
methods are (return full) a non-floating object is probably expected in
all cases.
By-pass the select-tracks-for-object signal for auto-transitions since
their track element must land in the same track as the elements it is
the auto-transition for.
The way a clip's track elements are added to tracks was re-handled. This
doesn't affect the normal usage of a simple audio-video timeline, where
the tracks are added before any clips, but usage for multi-track
timelines has improved. The main changes are:
+ We can now handle a track being selected for more than one track,
including a full copy of their children properties and bindings.
(Previously broken.)
+ When a clip is split, we copy the new elements directly into the same
track, avoiding select-tracks-for-object.
+ When a clip is grouped or ungrouped, we avoid moving the elements to
or from tracks.
+ Added API to allow users to copy the core elements of a clip directly
into a track, complementing select-tracks-for-object.
+ Enforced the rule that a clip can only contain one core child in a
track, and all the non-core children must be added to tracks that
already contains a core child. This extends the previous condition
that two sources from the same clip should not be added to the same
track.
+ Made ges_track_add_element check that the newly added track element
does not break the configuration rules of the timeline.
+ When adding a track to a timeline, we only use
select-tracks-for-object to check whether track elements should be
added to the new track, not existing ones.
+ When removing a track from a timeline, we empty it of all the track
elements that are controlled by a clip. Thus, we ensure that a clip
only contains elements that are in the tracks of the same timeline, or
no track. Similarly, when removing a clip from a timeline.
+ We can now avoid unsupported timeline configurations when a layer is
added to a timeline, and already contains clips.
+ We can now avoid unsupported timeline configurations when a track is
added to a timeline, and the timeline already contains clips.
Fixes https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/issues/84
This was used to connect to the track element's notify::start signal in
order to update the duration of the timeline (it is not clear why the
notify::duration signal was not also connected to for the same reason).
However, this is already covered by the timeline_tree_move method, which
is always called to update the start of a track element, even if it is not
part of a clip (and similarly for timeline_tree_trim, which is called
when the duration is set).
Before the max-duration could be set arbitrarily when the clip was empty,
to indicate what the max-duration would be once the core children were
created. Now, we can also do this whilst the clip only contains non-core
children.
Rename the private "owners" to "creators" to avoid confusing this with
the owner of the track element's memory.
Also made the ungroup method for GESClip symmetric by making all the
children of the resulting clips share their creators, which allows them
to be added to any of the other ungrouped clips. Once the clips are
grouped back together, the tracks loose these extra creators.
Make sure we sink the child on adding, and keep it alive until the end
in case the method fails.
Also, since the child mappings hold a ref to the child, they should give
them up in their free method. This way, the ref will be given up on
disposing, even if ges_container_remove fails.
Also, reverse setting of the start of the container if adding fails.
gcc-10 defaults to -fno-common, which exposes a symbol conflict, so
use `static` correctly. Also we don't use `parent_extractable_iface`
in `ges-uri-clip.c`.
See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85678
GES_TESTING_ASSETS_DIRECTORY is prehistoric and since then
new mechanism for asset relocation have been added, it makes
no sense to keep that unused code path
Refactoring GESVideoSource so that #GESUriVideoSource can handle
still image in a simple way
MultiFileSource has been replaced with the new `imagesequencesrc`
element, this was totally broken anyway as `multifilesrc` can not seek
properly.
This means that we have all the information about the asset
when constructing the underlying GstElements.
This also allows to cleanup some code all around
And deprecate all GESTrackElement constructors, but the GESEffect one.
Those should **never** be created by users and should become internal
in the future.
Stop having docstring for the constructors that were internal.
APIs:
- ges_timeline_get_frame_time
- ges_timeline_get_frame_at
- ges_clip_asset_get_frame_time
- ges_clip_get_timeline_time_from_source_frame
Extracting ges_util_structure_get_clocktime to internal utilities adding
support for specifying timing values in frames with the special
f<frame-number> synthax.
This is often very useful to have a timeoverlay inside test sources.
We do not want to use it as an effect as segments are not the sames
in GES when it comes to nleoperations.
When the `child-added` signal emission was called, the
`GESContainer->child_added` vmethod was called (the signal is
`G_SIGNAL_RUN_FIRST`) so we need to call `GESContainer->child_removed`
ourself so subclasses know they do not control the child anymore.
The properties will only have their signal emitted when they change in
value, even when g_object_set, etc, methods are used.
The _set_start method already did this, but start was missing the
EXPLICIT_NOTIFY flag. There should be no need to check that the property
has changed in ->set_start or ->set_duration
The in-point of a clip is kept in sync with its core children, unless they
have no has-internal-source.
The max-duration is defined as the minimum max-duration amongst the
clip's core children. If it is set to a new value, this sets the
max-duration of its core children to the same value if they have
has-internal-source set as TRUE.
Non-core children (such as effects on a source clip) do not influence
these values.
As part of this, we no longer track in-point in GESContainer. Unlike start
and duration, the in-point of a timeline element does not refer to its
extent in the timeline. As such, it has little meaning for most
collections of timeline-elements, in particular GESGroups. As such, there
is no generic way to relate the in-point of a container to its children.
As such, they only emit a signal if their value changes, either through
their _set_inpoint or _set_max_duration methods, or through
g_object_set, etc.
Also, we now require the ->set_max_duration method to be implemented.
This was added to GESGroup, which will only allow the max-duration to be
set to GST_CLOCK_TIME_NONE.
Unless this property is set to TRUE, the in-point must be 0 and the
max-duration must be GST_CLOCK_TIME_NONE.
Also added EXPLICIT_NOTIFY flags to the active and track-type
properties such that their notifies are emitted only if the property
changes, even when the g_object_set, etc, methods are used.
Also added a missing notify signal to the set_active method.
Previously the control bindings were not properly copied into the pasted
clip. Also changed the order so that elements are added to the clip
before the clip is added to the timeline.
Add the child-property-added and child-property-removed signals to
GESTimelineElement.
GESContainer is able to use this to keep their child properties in sync
with their children: if they are added or removed from the child, they
are also added or removed from the container.
Hold the notify signals for the container and the children until after
the child has been fully added or removed.
After the previous commit, this was used to ensure that the
notify::priority signal was sent for children of a clip *after* the
child-removed signal. This stopped being the case when the code in
->child_removed was moved to ->remove_child (the latter is called before
the child-removed signal is emitted, whilst the former is called
afterwards). Rather than undo this move of code, which was necessary to
ensure that ->add_child was always reversed, the notify::priority signal
is now simply delayed until after removing the child has completed. This
was done for all notify signals, as well as in the add method, to ensure
consistency.
This allows the test_clips.py test_signal_order_when_removing_effect to
pass.
Also make subclasses take a copy of the list of the children before
setting the start and duration, since this can potentially re-order the
children (if they have the SET_SIMPLE flag set).
Previously, we relied on ->child_removed to reverse the priority changes
that occured in ->add_child. However, ->child_removed is not always
called (the signal child-removed is not always emitted) when a
->add_child needs to be removed. However, ->remove_child is always
called to reverse ->add_child, so the code was moved here. Otherwise, we
risk that the priorities of the clip will contain gaps, which will cause
problems when another child is added to the clip.
Handle the child priorities in a way that keeps the container children
list sorted by priority at all times. Also, no longer rely on the
control_mode of the container, since we have less control over its value,
compared to private variables.
Also fixed the changing of priorities in set_top_effect_index:
previously *all* children whose priority was above or below the new
priority were shifted, when we should have been only shifting priorities
for the children whose priority lied *between* the old and the new
priority of the effect. E.g.
effect: A B C D E F
index: 0 1 2 3 4 5
After moving effect E to index 1, previously, we would get
effect: A B C D E F
index: 0 2 3 4 1 6
(this would have also shifted the priority for the core children as
well!). Whereas now, we have the correct:
effect: A B C D E F
index: 0 2 3 4 1 5
Only allow elements that were created by ges_clip_create_track_elements
(or copied from such an element) to be added to a clip. This prevents
users from adding arbitrary elements to a clip.
As an exception, a user can add GESBaseEffects to clips whose class
supports it, i.e. to a GESSourceClip and a GESBaseEffectClip.
This change also introduces a distinction between the core elements of a
clip (created by ges_clip_create_track_elements) and non-core elements
(currently, only GESBaseEffects, for some classes). In particular,
GESBaseEffectClip will now distinguish between its core elements and
effects added by the user. This means that the core elements will always
have the lowest priority, and will not be listed as top effects. This is
desirable because it brings the behaviour of GESBaseEffectClip in line
with other clip types.
We were implementing the logic for moving/trimming elements specific
to SourceClip but this was not correct ass the new timeline tree allows
us to handle that for all element types in a generic and nice way.
This make us need to have groups trimming properly implemented in the
timeline tree, leading to some fixes in the group tests.
This adds tests for the various cases known to not be handled properly
by the previous code.
Fixes https://gitlab.freedesktop.org/gstreamer/gst-editing-services/issues/92
Unless the pipeline is in certain modes, we do not want to try and link
every track. The previous debug message implied this, but the method did
not actually end early.
Also, we always end early if we receive a track that is neither video
nor audio.
Previous usage of the property proxy-target seemed to alternate between
the two definitions:
+ The asset we are the default proxy of
+ The asset we are in the proxy list of
Now, the latter definition is used, which seems more useful to a user
since knowing the latter can easily allow you to find out the former.
The previous behaviour of ges_asset_set_proxy (asset, NULL) was not very
clear. It is now defined so that it clears all the proxies for 'asset'.
This means that after this call, the GESAsset:proxy property will indeed
be NULL.
Also fixed:
+ We can call ges_asset_set_proxy (asset, proxy) when 'proxy' is already
in the proxy list of 'asset'.
+ Handling of removing the default proxy in ges_asset_unproxy. This was
sending out the wrong notifies.
+ Prohibiting circular proxying. Before we could only prevent one case,
we should now be able to prevent all cases. This will prevent a hang
in ges_asset_request.
Fix the ownership in ges_asset_request. This should be transfer-full,
but for proxies it would fail to add a reference. Also,
ges_asset_cache_put was leaking memory if the asset already existed.
We should not be accepting ges_asset_set_proxy (NULL, proxy) as part of
the API! This behaviour was used internally in combination with
ges_asset_try_proxy, which is called on a still loading asset, so it was
moved to ges_asset_finish_proxy.
Previously, we tested that the given priority was `>0`, when it seems
that `>=0` was intended. A priority of `-1` means leave the priority
unchanged, whilst a priority of 0, or more, means move to this layer
priority.
Stop overwriting the ->list_children_properties virtual method in
subclasses because the timeline element class handles everything itself
anyway.
Note that containers already automatically add the children properties of
their child elements in ges_container_add.
Previously, we were setting the inpoint_offset using the start offset in
the duration callback!
Also added a notify for when the duration is changed in the child start
callback.
Previously, either the track or track_type arguments had to be specified
in order to find **any** track elements. Now, you can specify neither,
which will match any track element, of the specified type.
If a clip is already part of a layer, then adding it to another layer
should fail. Previously, in this case, `ges_layer_add_clip` was adding a
reference to the clip instead, without subsequently giving up ownership.
This meant that the clip would be left with an unowned reference.
This has now been corrected by also calling `unref` after the
`ref_sink`.
Note that, since `clip` is already part of `current_layer`, it should
already be non-floating, so the `ref_sink`-`unref` should do nothing
overall. But we keep both to make the ownership (transfer floating/none)
explicit.
The method steals ownership of `copied_from`, so should be responsible
for unreffing it. Also make sure we fail when `layer != -1`, since this
functionality is not supported.
We try to do our best to have the video frames scaled the best way
to fill most space on the final frames, keeping aspect ratio. The user
can later on rescale or move the sources as usual but it makes the
default behaviour a better and more natural especially now that we
set default restriction caps to the video tracks.
And fix the unit test to take that change into account
This way the user can easily know how the clip would look like
if no scaling was applied to the clip, this is useful to be able
to properly position the clips with the framepositionner element.
We used to always call the `->set_child_property` virtual method
of the object that `ges_timeline_element_set_child_property` was called
from, but that means that, in the case of referencing GESContainer
children properties from its children, the children wouldn't know
what child property have been set, and the children override wouldn't
be takent into account, in turns, it means that the behaviour could be
different in the setter depending on parent the method was called,
which is totally unexpected.
We now make sure that the vmethod from the element that introduced the
child property is called whatever parent method is called, making the
behaviour more uniform.
Fix the python override to make sure that new behaviour is respected.
By default, video track output full HD@30fps, this makes the behaviour
of clip position much more understandable and guarantess that we
always have a framerate.
The user can modify the values whenever he wants
Basically we were advertising that the source size would be the
size of the track if it hadn't been defined by end user, but since
we started to let scaling happen in the compositor, this is not true
as the source size is now the natural size of the underlying video
stream.
Remove the unit test and reimplemented using a validate scenario which
make the test much simpler to read :=)
