fix(desktop): preserve recovered media when predecessors become obsolete - #75
Merged
Merged
Conversation
rldyourmnd
marked this pull request as ready for review
October 1, 2026 23:44
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
A predecessor arriving after a newer recovery picture was treated as a rejected reader, requesting another key. Its stream stop could also invalidate sender recovery; a stop during payload writing ended the video writer.
Receiver and sender now distinguish a valid obsolete predecessor with one exact application STOP_SENDING code. Disposal releases admission without another repair, writer termination or a fresh delivery ACK for bitrate growth. Unknown stops, malformed current headers and actual failures keep their existing behavior. No serialized greeting/control/frame layout or authority changes; older senders retain their generic-stop fallback.
Validation: the delayed-header real-stream regression failed on old behavior by requesting another key; a separate unfinished-payload stop regression failed by ending the writer. Both passed afterward on Iroh and Noq, retaining malformed-frame repair, unknown-stop failure, admission release, no false bitrate growth and connection usability. All 74 macOS desktop units, 253 focused integration tests and strict full-workspace/all-target/all-feature lint passed. Both optimized builds were installed with executable/identity/IPC/service receipts. Linux focused qualification also completed; exact counts are retained privately. A sequence-correlated receipt regression also rejects a late failure invalidating an independently recovered chain and a second repair request.
The installed predecessor still showed multi-second media pauses, and controlled native timing was confounded by occlusion/targeting. This PR does not claim to fix every freeze or close native latency, quality and sustained stability gates. Scoped evidence: docs/reports/rds-obsolete-frame-20261002.md.
Ubuntu full-workspace CI exposed an unrelated fixture port reservation race before candidate identity assertions; its exact failure is retained and a repair is in progress. That job is not reported as passed.