Changing the default value from http to https. At this point the playback links generated in the recording metadata.xml files should default to httpS protocol
When a shape is changed, the full shape objcect was being transmitted to the server again.
Do a diff to only send what changed (similarly as it was in tldraw v1) to save upload bw.
TODO:
Draw segments diffs (array) is still not working, so all the segments are still being sent every time.
The merge of 2.7 code into 3.0 dropped a change which was part of a fix
to prevent a recording processing hang when processing deskshare streams
which contain audio. Re-apply that change.
I found a failed recording where a deskshare video file could not be
decoded when seeking was used (oddly enough, it could decode without
seeking). Turns out that remuxing the file was enough to make things
work.
In the video processing code, after checking for other known video file
problems that are fixed by remuxing, also try decoding one frame of the
video with a seek time applied.
Since more videos can end up being remuxed now, I moved the output of
the remux to a new subdirectory. Since more file types than just 'flv'
can be remuxed, I switched the output to the 'nut' container instead.
(This is ffmpeg's generic container format, and can hold any audio/video
which ffmpeg can process, so it works well for temporary files.)
This is a workaround for #19178 - but it does not fix the issue. The
caption recording events with invalid empty `<locale/>` are simply
dropped with a warning message, to allow the recording and any valid
caption streams present to be processed.
Co-authored-by: Calvin Walton <calvin.walton@blindsidenetworks.com>
Both bbb-playback and the recording processing scripts fail if the version string is not in the format MAJOR.MINOR.PATCH (3.0.0 in this case instead of 3.0)
Recording archive may fail when remuxing invalid files from KMS or the
new recorder - eg when the raw files are 0-byte sized.
This commit handles the exception raised by EDL::encode so archive keeps
going, logs the issue as a warning and archives the problematic file anyways.
EDL::encode now removes the temporary file when the ffmpeg command execution
fails - this should avoid leaving any stale files around in case of failure.
No specific check for the nature of the error is done - the idea is that
subsequent phases will discard or fix the files if necessary according
to the processing scripts' necessities, making the behavior (in this
specific scenario) similar to what it was before the archive remuxing was
introduced.
Kurento may *rarely* generate WebM/MKV files with corrupt or absent
SeekHead sectors. bbb-webrtc-recorder also doesn't generate SeekHead or
even the Cues sectors by default.
While those are are *optional* fields by spec, files need to be seekable
for our recording processing scripts to work.
This commit adds a remuxing step for Kurento and bbb-webrtc-recorder raw
files that is executed during the archive phase. It should re-include
any of the missing fields that make files seekable and restore the Cues
sector in WebM files.
The code was skipping the check for cursor x or y position < 0 when the
tldraw whiteboard was in use. That condition is still needed on the
tldraw whiteboard to indicate that the cursor should be hidden.
Only the check for cursor x or y position >100 needs to be skipped when
the tldraw whiteboard is in use (since tldraw cursors are in the slide
coordinate space, they can go up to x=1440 or y=1080)
The Fcntl::F_SETPIPE_SZ constant was added in Ruby 3.0, but Ubuntu 20.04
still uses Ruby 2.7. Add some error handling so processing doesn't fail
if the constant is not found.
The encoding settings for intermediate files was using -preset veryfast
-crf 30, which resulted in very poor quality video.
After a bit of experimenting, I decided to change this to -preset
veryfast -crf 23. This results in files which are roughly twice the size
of before, but they look significantly better.
There's improvements possible at the same filesize by switching to a
slower encoding mode. But in the case of -preset medium for example,
when normalized to the same output file size, you end up using about 1.5
times as much cpu time to gain only a very small amount of video
quality.
