It makes for a better user experience to simply adjust the pts of a late
item, but to preserve its duration: for instance a speech synthesis
element might use the duration as a hint for speeding up the audio.
Future late items may also be similarly offset anyway.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1930>
With standard voices, AWS polly supports passing a max-duration
attribute.
When the element gets raw text passed in, it can wrap it as SSML and set
the max duration attribute, this to make sure synthesized speech
doesn't overlap.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1930>
Otherwise when changing the target caps (e.g. for reducing quality)
there is a race condition between buffers between the converter elements
and renegotiation.
For example, videoconvertscale might've output a 1920x1080 buffer, then
the capsfilter is configured to 1280x720, the buffer arrives in
videorate, videorate notices that renegotiation is pending, tries to
renegotiate and ends up with EMPTY caps because it can only change the
framerate but not the resolution.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1949>
It is possible that in unprepare(), waiting for a task to complete while
holding the state lock, that task may be waiting to acquire the state lock and
result in a deadlock.
This is quick to reproduce when starting and stopping webrtcsink in very quick
succession.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1931>
When the transcriber is used in a live situation, it can be useful
to save a transcript for editing after the fact when producing a
VOD.
Each source pad now gets an "unsynced_" pendant. That unsynced pad
is pushed to from the context of the "live" source pad task. Flow
returns from the unsynced pads are ignored, we simply check the
last flow return before attempting to push the next transcript.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1915>
While at it, use PushSrc as base class. quinnquicsrc never supported
seeking and only ever operated in push mode. Length and offset for
create from BaseSrc was also never really honoured. Use PushSrc as
the base class which is more appropriate.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1634>
When the latency is configured to a value that is too low, items will be
pushed out with an adjusted timestamp, thus affecting synchronization.
It can be useful for the application to receive details about those
adjustments.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1793>
The GstWebRTC API web interface defaults to receiving an SDP offer and
generating an answer, but this can be overridden by entering "offer
options" before clicking to open the remote stream. The Python
webrtcsink-custom-signaller.py example failed in this mode as it was
coded to only generate an offer and receive an answer. Fix this by
implementing support for receiving an offer and sending an answer.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1883>
rustls allows the choice of ring or aws-lc-rs as the cryptographic
library implementation. This is enabled/selected via Cargo feature
flags. We have plugins directly or indirectly depending on rustls
like quinn, aws and spotify. In the presence of multiple plugins,
selecting different implementations as the default, rustls can
panic.
The safest way to avoid this is by using builder_with_provider
and selecting a provider explicitly.
See below issues for further discussion and clarifications.
https://github.com/rustls/rustls/issues/1877https://github.com/seanmonstar/reqwest/pull/2225
While at it, also specify features explicitly for quinn and rustls.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1878>
