In certain cases, roundedcorners would negotiate to I420 even when user
supplied a non-zero border radius.
For example, the below pipeline leads to I420 being negotiated even
though a non-zero border radius was given. Ideally, this pipeline
should have failed at the negotiation stage.
```bash
gst-launch-1.0 -v \
videotestsrc num-buffers=1000 pattern=white ! \
video/x-raw,width=320,height=180 ! \
roundedcorners border-radius-px=10 ! videobox border-alpha=0 top=-10 left=-10 right=-10 bottom=-10 fill=yellow ! \
compositor name=comp sink_0::xpos=960 sink_0::ypos=0 sink_0::width=320 sink_0::height=180 sink_0::alpha=1.0 sink_1::xpos=960 sink_1::ypos=180 sink_1::width=320 sink_1::height=180 sink_1::alpha=1.0 \
sink_2::xpos=960 sink_2::ypos=360 sink_2::width=320 sink_2::height=180 sink_2::alpha=1.0 sink_3::xpos=0 sink_3::ypos=0 sink_3::width=960 sink_3::height=720 sink_3::alpha=1.0 ! \
video/x-raw,width=1280,height=720! x264enc ! mp4mux ! filesink location=test.mp4 \
videotestsrc num-buffers=1000 pattern=red ! \
video/x-raw,width=320,height=180 ! roundedcorners border-radius-px=10 ! comp. \
videotestsrc num-buffers=1000 pattern=blue ! \
video/x-raw,width=320,height=180 ! roundedcorners border-radius-px=10 ! comp. \
videotestsrc num-buffers=1000 pattern=green ! \
video/x-raw,width=960,height=720 ! roundedcorners border-radius-px=10 ! comp.
```
If border radius is non-zero, we should not really allow negotiation
to select I420. Fix this by returning only A420 for border-radius > 0
in `transform_caps` instead of returning both like earlier.
Another example of a simpler pipeline like below which would earlier work
```bash
gst-launch-1.0 videotestsrc pattern=red ! videoconvert ! video/x-raw,width=1923,height=1087,format=I420 ! roundedcorners border-radius-px=40 ! video/x-raw,format=I420 ! videoconvert ! gtksink
```
now fails with
```bash
WARNING: erroneous pipeline: could not link roundedcorners0 to videoconvert1, roundedcorners0 can't handle caps video/x-raw, format=(string)I420
```
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1211>
After calculating the earliest pts, the fragment_start_pts and
chunk_start_pts in State are updated. However, when checking if the
stream is filled, the previous start_pts (set to None) is used instead.
This means that chunk_filled and fragment_filled will be false the first
time aggregate() is called, assuming timeout is false, all_eos is false,
and the sinkpad is not EOS. This requires aggregate() having to be
called a second time before the first fragment is sent.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1207>
In some rare cases, the webrtc-test entered a deadlock while executing
`WebRTCSink::unprepare`. Attaching gdb to a blocked instance showed:
* `gstrswebrtc::signaller:👿:Signaller::stop()` parked, waiting for a
`Condvar` in `Signaller::stop()`. This was most likely awaiting for the
receive task to complete while it was locked in `element.end_session()`.
This code path is triggered from `unprepare` with the `State` `Mutex` locked.
* `webrtcsink:👿:WebRtcSink::process_stats` waiting for a contended `Mutex`,
which is also the `State` `Mutex`. This prevented completion of the signal
`gst_webrtc_bin_get_stats`.
This commit aborts the task in charge of periodically collecting stats and
ensures any remaining iteration completes before requesting the Signaller to
stop.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1202>
Prior to this commit, we were sending over words concatenated together
with no separators, for instance "Idon'twanttobeanemperor".
The translation service seems clever enough to translate the contents
anyway, but there is no reason to make its task harder than necessary,
and it didn't re-add separators when the target language was the same as
the source language, which resulted in less than ideal output.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1171>
Session ending is bidirectional: the signaller can tell the sink that a
session was ended, and the sink can tell the signaller to end a session.
As such, two signals are needed, before this patch the second case was
not working as in essence the sink was telling itself that a session was
ended, and obviously failing to even find it when trying to end it again.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1167>
