Stopping to do it at the bin description level but properly
plugging them where they are needed and cleanly ghosting the pads
where it makes most sense.
This introduces support for GES to request pads on the most upstream
element in case no static pad can be ghosted.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/187>
ges_uri_clip_asset_get_duration does not tell us what the duration in
the timeline needs to be. Especially when we have time effects, or
effects with finite max-durations. So we should no longer expect the
duration to stay the same when replacing assets. Instead, we just check
that the new max-duration would be compatible with the current in-point
(which was not checked before), and the clip would not be totally
overlapped if its duration-limit changes.
This is based on the assumption that each source is replaced one-to-one
in its track. If a source is replaced with nothing in the same track,
this check may be a little too strong (but still mostly weaker than
before). However, problems could occur if track selection does
something unexpected, such as placing the new source in a track not
previously occupied.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/177>
The out-point, which is an internal time, is used instead of the
duration for determining the control binding value at the end of the
element.
Also, allow the user to switch off the auto-clamping of control sources
if they are not desired. And allow them to clamp specific control sources
individually.
Also, fix a lot of memory leaks related to control sources. In
particular, releasing the extra ref gained by source in
g_object_get (binding, "control-source", &source, NULL);
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/177>
If a layer priority sits between the priorities of two layers in the
timeline, i.e. it references a gap in the timeline's layers, then
ges_timeline_append_layer will never fill this gap and create the
desired layer, so the edit in timeline-tree would loop forever. So a
check was added to avoid this.
This would be a usage error, but a user can reasonably end up with a gap
in their layers if they remove a layer from the timeline.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/177>
Unlike ges_container_add, this lets you set the index and will check
that track selection did not fail. This is useful for time effects whose
addition would create an unsupported timeline configuration.
Also can use the clip add error in ges_timeline_add_clip to let the user
know when adding a clip to a layer that its in-point is set larger than
the max-duration of its core children.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/177>
Allow the user to register a child property of a base effect as a time
property. This can be used by GES to correctly calculate the
duration-limit of a clip when it has time effects on it. The existing
ges_effect_class_register_rate_property is now used to automatically
register such time effects for rate effects.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-editing-services/-/merge_requests/177>
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>
Otherwise there is a pretty rare race where we get the
_initialize_stack_func executed leading to the stack set up and
the source pushing buffers before the composition source pad is
activated, and a STREAM_ERROR is reported as we end up pushing a
buffer to a flushing pad.
Thanks rr chaos mode for showing that improbable race
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>
