Repository navigation
fix/flush and rtp timeline follow ups - #28
Merged
dangusev merged 7 commits intoOct 13, 2026
Merged
Conversation
dangusev
added this pull request to stack #31
October 13, 2026 14:22
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit d3c5964. Configure here.
dangusev
force-pushed
the
fix/flush-and-rtp-timeline-follow-ups
branch
2 times, most recently
from
October 13, 2026 14:54
d3c5964 to
72907b8
Compare
flush took the resampler lock with a blocking lock, and a write holds that lock for the full resample: about 130 ms for a 30 s write at 24 kHz. Also, a flush that came before a write took the lock did not remove the audio of that write. A flush counter now marks the writes that started before a flush: their audio is not queued, and the write resets the filter. flush no longer waits for the resampler.
Each write path kept its own counters: the packetizer for write_pcm and write_sample, VideoEncoder.seq and VideoClock.rtp_ts for write_i420, and fwd_seq with the source timestamp for write_rtp. continue_rtp_from did not carry the forward state over. When a forwarding track took over a stopped sender, or a track took over the sender of a forwarding track, the sequence numbers jumped or went back on the same SSRC, and the receiver dropped the packets. TrackCore now holds one RtpTimeline (last sequence number, next timestamp, forward offset) for all write paths. Packetized writes take their sequence numbers and timestamps from it. Forwarded packets keep the spacing of the source timestamps, shifted to continue the timeline. continue_rtp_from carries the timeline over and adds the time since the stop.
take_pcm read pcm_written in the middle of a tick and cleared it at the end. A write that set the flag between the two was lost, and the next tick could flush the resampler while the stream continued. take_pcm now swaps the flag to false at the start of the tick.
The track has pace off, so no pacer starts. The test itself takes the frames.
dangusev
force-pushed
the
fix/flush-and-rtp-timeline-follow-ups
branch
from
October 13, 2026 15:02
72907b8 to
20a3935
Compare
aliev
approved these changes
Oct 13, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stack created with GitHub Stacks CLI • Give Feedback 💬