[kixelated]

Completes quest/m2/mux-sync.md and deletes it.

Problem

moq_mux::timeline, the native playout clock behind moq play, was audio-master. While audio played, only the speaker's position moved the anchor. A stalled speaker that came back on a fresh sink re-pinned the anchor and held video back with it. That contradicts the 2026-11-05 ruling on moq-dev/moq#7941: lip sync in the steady state, desync under congestion, and one slow track never stalls or moves another. The module also carried its own copy of Kind.

Approach

This ports js/watch/src/sync.ts (per-track timelines on one reference) to moq_mux::sync:

• One wall-clock reference. Each track keeps the earliest arrival on its own timeline (arrival time against timestamp), and the clock's reference is the earliest of those. An early frame pulls playback toward live. A late one only arrives late. Resetting or dropping a track gives up its part, so the reference falls back to the earliest track left. With no track left, it waits for the next frame to anchor. The last track to go leaves its schedule behind, so pictures already queued still present on time.
• Delay. Playback trails the reference by the largest target a live track holds (Track::set_target). While no track holds one, it uses the Producer::new delay. Audio holds its jitter target. Video has no estimator, so it holds none, which keeps today's latency (100 ms video-only, audio's estimate otherwise).
• Audio converges, it doesn't own. moq play folds each audio frame's arrival into the clock. Each write sets the sink's stretch target to due(end) - now, clamped to the 50 ms device floor and the age budget. In the steady state that is the jitter target. After a stall it is shallower than what the sink holds, so the callback speeds up, or past NetEq's fast threshold trims, until audio is back on the clock. Video never waits for it.
• Held replacements. A replacement segment can wait behind the old segment's tail at a declared end. Its arrivals are then shifted by that hold (recorded at plan, applied at apply), so media we held back doesn't read as early and drag the other tracks ahead.
• Transitions over every Kind. The declared-end and encoder-restart machinery now covers container::Kind (audio, video, data). A paired restart takes its common origin from the member plan with the earliest source rather than always audio's. The two agree in every existing case.
• The speaker's playhead is the speaker's own state, not the clock's, so it moves into moq play as Playhead and feeds --metrics. --metrics' delay is now the clock's resolved delay, matching the browser's sync.out.delay.

Tests use mocked time throughout. New ones:

• sync::stalled_audio_no_longer_holds_video: a late audio burst and a reset leave video's schedule alone.
• play::media::stalled_audio_no_longer_holds_video: end to end through play_audio and Video::run. Audio stalls, its backlog lands late and forces a fresh sink, and every picture still presents at start + delay + pts.
• sync::a_video_burst_moves_audio_onto_the_live_edge, a_reset_hands_the_schedule_to_the_tracks_left, the_delay_is_the_largest_target_held, media_held_behind_a_tail_leaves_the_schedule_alone, a_data_track_joins_a_paired_restart.

The audio-owns-the-anchor tests are gone. The video burst tests with a speaker now assert that the burst reaches live past a trailing speaker. The playhead tests move to play/timeline.rs.

Impact

• Breaking, moq-mux: moq_mux::timeline is now moq_mux::sync. Release had an unrelated moq_mux::timeline (the join-interval one), so the cut's MIGRATE entry should say both are gone. Migration:
  • timeline::Kind → moq_mux::container::Kind. Producer::track takes it, and Kind::Data is accepted.
  • Producer::video(ts, now) → track.received(ts, now) on the video Track. Producer::audio(Audio { .. }, now) → track.received(ts, arrival) on the audio Track, plus track.set_target(jitter_target). Steer the speaker toward producer.due(end) yourself.
  • Producer::video_restarted and Producer::audio_restarted → track.reset(). Producer::audio_stopped → drop or retire the track.
  • timeline::Audio is removed.
  • Consumer::target() → Consumer::delay(), which always returns the resolved delay. Consumer::playhead() is removed: the speaker tracks what it is sounding.
  • Producer::new(delay) now means the delay while no track holds a target.
• Downstream: moq-dev/studio's moq-scene playout uses this module through a path dependency. It adapts on its next pin with the migration above.
• moq play: no flag changes. A stalled audio track no longer holds video. --metrics writes delay on every sample. Audio-only steady-state depth after a write is one packet deeper than before, because a frame's end now sounds when the clock says it is due.
• Bench timeline is renamed to sync (sync_present, sync_received), in rs/justfile and the nightly list.
• Wire: none.

Alternatives

• Rename only, keeping the audio anchor. The maintainer rejected this on 2026-11-07.
• Unshifted arrivals for held replacements. Folding them as they arrived would skip every track toward live after each quick restart. That claws back the hold's latency, but it moves continuous video when only audio restarted, which breaks the decoupling. The shift preserves today's restart behavior.
• Video holding --delay's auto default (100 ms) as a target. This would raise A/V latency above audio's own estimate, so video holds no target.

Choices made without asking

• Kept the transition/mapping machinery and generalized it to three kinds, rather than narrowing sync to audio and video. The quest asks for container::Kind, which includes Data.
• moq play's two playback halves keep a private two-variant Kind, since the player has no data half.
• A constant offset between tracks (audio routed 200 ms slower than video, say) plays audio late by the part the delay doesn't cover, as in the browser. The catalog delay field is how a publisher states that offset (see doc/concept/audio-jitter.md).

Follow-ups

• Plan: audit moq play's use of the catalog rendition delay against doc/concept/audio-jitter.md's max(delay + target). The browser holds it, and this PR doesn't change the native side. Recommendation: plan it, since a slow video encoder otherwise plays behind audio natively.
• Drop: publisher clock drift slower than the local clock slowly erodes the delay, because the reference only moves earlier. The browser has the same property, and nothing has measured it mattering.

Models

• Writer: Claude Opus 5.5 (high)
• Reviewer: gpt-6.1-sol (xhigh)

Review

Rewritten by quest review after each round.

• Reviewer: Codex, gpt-6.1-sol, xhigh effort
• Session: 01a1ba42-5819-7ae2-8aae-77b707195bca, round 3
• Head: bb45de4
• Base: origin/main at 35b9909
• Score: 10/10, The previously reviewed implementation is unchanged, and static review of the merge resolution found no new issues; no builds or tests were run.
• Findings: none

🤖 Generated with Claude Code

(Written by Claude Opus 5.5 with high effort)

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 4 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used all 4 included reviews currently available.

Learn how review limits work.

Review configuration:

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 13153c3f-3f36-415d-84b9-e171e6aa2b70

Commits

Reviewing files that changed from the base of the PR and between 7898979 and bb45de4.

Files selected for processing: 20

• doc/bin/cli.md
• doc/lib/rs/moq-mux.md
• quest/m2/README.md
• quest/m2/api-shape/media.md
• quest/m2/mux-sync.md
• rs/justfile
• rs/moq-cli/src/play/args.rs
• rs/moq-cli/src/play/headless.rs
• rs/moq-cli/src/play/media.rs
• rs/moq-cli/src/play/metrics.rs
• rs/moq-cli/src/play/playback.rs
• rs/moq-cli/src/play/timeline.rs
• rs/moq-cli/src/play/video.rs
• rs/moq-cli/src/play/window.rs
• rs/moq-mux/Cargo.toml
• rs/moq-mux/benches/sync.rs
• rs/moq-mux/src/lib.rs
• rs/moq-mux/src/sync.rs
• rs/moq-mux/src/timeline.rs
• sh/gh/nightly.sh

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

• X
• Mastodon
• Reddit
• LinkedIn

[kixelated]

Automated review by review (OpenAI)

Reviewed a0c65b7. No actionable findings. Checked per-track earliest references, reset/release and target resolution, held-replacement offsets, three-kind transition mapping, and CLI audio/video integration. Audio now steers its buffer toward the shared due time without owning video's anchor; speaker playhead tracking remains separate for metrics. This matches the explicit direction of steady-state sync with independent progress during stalls.

Verification: GitHub-only static review against the previous timeline implementation and JS sync semantics, including regression tests and updated benchmark wiring. No builds, tests, audio-device playback, timing reproduction, or benchmarks run.

[pullfrog]

✅ No new issues found.

Reviewed changes

• moq_mux::sync replaces moq_mux::timeline — per-track arrival anchors with the earliest as the reference, left retained for already-queued media, reset/drop releasing a track's part, set_target driving the resolved delay, and container::Kind (Audio/Video/Data) replacing the module's own Kind.
• Generalized transition machinery — declared ends and encoder restarts now span three roles; paired restarts take their origin from the earliest source plan, and Segment.held shifts arrivals from media that waited behind an old tail.
• moq play audio no longer owns the clock — each write folds the frame's arrival, sets the track's jitter target, and steers the sink toward due(end) - now (clamped to the device floor and age budget); the Playhead moved into play/timeline.rs and is reset per broadcast, with --metrics reporting the resolved clock delay.
• Wake handling and tests — audio task completion now wakes video when the released target lowers the delay; new mocked-time tests cover stalled audio, reset handoff, target selection, held replacements, and a data-track paired restart.
• Renames and docs — bench timeline → sync across Cargo.toml, rs/justfile, and nightly; doc/bin/cli.md, doc/lib/rs/moq-mux.md, crate docs, and quest references updated; quest/m2/mux-sync.md deleted. MIGRATE.md is correctly left for the release cut.

I traced the anchor/left fallback, reset/release paths, transition pairing and pruning, the held shift, Playhead locking and per-generation isolation, and the audio stretch-target math; no stale moq_mux::timeline callers remain in the tree. Local cargo test --release -p moq-mux --lib sync:: was blocked by a held target-directory lock in this sandbox, so CI is the build/test authority for this head.

  ^{｜ View workflow run ｜ Using deepseek-v4.1-flash-1017 (free via Pullfrog for OSS) ｜ 𝕏}

[moq-bot]

quest land watch: the merge queue removed this pull request at a0c65b7 because its merge_group build failed (failed_checks).

The land fixer works on it in the background, and the watcher re-enqueues it at the back of the queue once it is fixed or judged flaky, or hands it back here with what it needs.

Queue / Queue (job) failure at step Check:

…ent_consumes_only_the_request
        PASS [   0.012s] ( 2156/13788) moq-ffi server::transport_tests::converts_supported_transports
        PASS [   0.092s] ( 2157/13788) moq-ffi session::tests::maps_native_auth_connect_errors
        PASS [   0.092s] ( 2158/13788) moq-ffi session::tests::maps_closed_errors_without_flattening_the_reason
        PASS [   0.363s] ( 2159/13788) moq-ffi producer::metadata_tests::dropping_a_group_binding_ends_its_unfinished_group
        PASS [   0.689s] ( 2160/13788) moq-ffi media::timebase_tests::an_unused_timebase_does_not_withhold_the_catalog
        PASS [   0.262s] ( 2161/13788) moq-ffi socket::tests::protocols_are_independent_and_endpoint_close_wakes_both_queues
        PASS [   0.561s] ( 2162/13788) moq-ffi origin::supervision_tests::readers_derived_from_a_routed_broadcast_report_the_origin_panic
        PASS [   0.895s] ( 2163/13788) moq-ffi media::read_tests::encoded_media_preserves_breaks
        PASS [   0.275s] ( 2164/13788) moq-ffi socket::tests::cancelled_accept_wait_leaves_the_queue_and_rejection_is_one_shot
        PASS [   0.274s] ( 2165/13788) moq-ffi socket::tests::cancelling_a_dial_resets_its_incoming_but_the_endpoint_can_dial_again
        PASS [   0.197s] ( 2166/13788) moq-ffi test::a_closed_route_leaves_its_replacement_alone
        PASS [   0.285s] ( 2167/13788) moq-ffi socket::tests::tcp_halves_outlive_the_stream_and_preserve_fin
        PASS [   1.002s] ( 2168/13788) moq-ffi media::catalog_tests::catalog_handle_updates_sections
        PASS [   0.199s] ( 2169/13788) moq-ffi socket::tests::track_subscription_survives_stream_release_and_has_no_default_age_limit
        PASS [   0.376s] ( 2170/13788) moq-ffi socket::tests::dropping_an_unfinished_tcp_writer_resets_with_a_deliberate_cause
        PASS [   0.286s] ( 2171/13788) moq-ffi socket::tests::tcp_reset_is_not_fin_and_does_not_close_another_session
        PASS [   0.104s] ( 2172/13788) moq-ffi test::a_hidden_route_is_invisible_until_announced
        PASS [   0.189s] ( 2173/13788) moq-ffi test::an_instance_pins_request_and_follow
        PASS [   0.292s] ( 2174/13788) moq-ffi test::a_new_route_at_a_path_restarts_its_followers
        PASS [   0.190s] ( 2175/13788) moq-ffi test::an_exact_route_yields_no_requests_and_a_prefix_route_no_broadcast
        PASS [   0.191s] ( 2176/13788) moq-ffi test::abort_after_finish_reaches_group_consumer
        PASS [   0.191s] ( 2177/13788) moq-ffi test::accepting_a_request_after_close_is_refused
        PASS [   0.192s] ( 2178/13788) moq-ffi test::abort_after_finish_keeps_the_track_handle
        PASS [   0.190s] ( 2179/13788) moq-ffi test::announce_and_withdraw_toggle_discovery
        PASS [   0.293s] ( 2180/13788) moq-ffi test::a_waiting_ready_ends_on_abort
        FAIL [   1.193s] ( 2181/13788) moq-ffi media::read_tests::raw_media_catalog_warmup_reaches_the_subscription
  stdout ───

    running 1 test
    test media::read_tests::raw_media_catalog_warmup_reaches_the_subscription ... FAILED

    failures:

    failures:
        media::read_tests::raw_media_catalog_warmup_reaches_the_subscription

    test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 321 filtered out; finished in 1.18s

  stderr ───

    thread 'media::read_tests::raw_media_catalog_warmup_reaches_the_subscription' (1988497) panicked at rs/moq-ffi/src/media.rs:1635:13:
    assertion `left == right` failed: video=false
      left: [4400000, 5000000]
     right: [4400000, 4700000, 5000000]
    note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

  Cancelling due to test failure: 23 tests still running
        PASS [   0.302s] ( 2182/13788) moq-ffi test::a_track_is_adopted_by_one_writer
        PASS [   0.306s] ( 2183/13788) moq-ffi test::a_failed_finish_ends_the_track_with_its_cause
        PASS [   0.017s] ( 2184/13788) moq-ffi test::audio_rejects_bad_init_bytes
        PASS [   0.019s] ( 2185/13788) moq-ffi test::announced_hides_dot_paths_unless_asked
        PASS [  12.659s] ( 2186/13788) moq-cli::integration report::a_late_sink_reads_the_final_report
        PASS [   3.069s] ( 2187/13788) moq-ffi audio::tests::retired_audio_producers_keep_clean_tails_for_late_readers
        PASS [   0.109s] ( 2188/13788) moq-ffi test::announced_filters_patterns_and_reports_captures
        PASS [   0.108s] ( 2189/13788) moq-ffi test::audio_raw_publish_rejects_bad_frames
        PASS [   0.111s] ( 2190/13788) moq-ffi test::ascending_subscription_reads_a_cached_backlog_oldest_first
        PASS [   0.110s] ( 2191/13788) moq-ffi test::audio_consumer_follows_a_sibling_broadcast_reference
        PASS [   0.110s] ( 2192/13788) moq-ffi test::audio_subscribe_configuration_boundary
        PASS [   0.301s] ( 2193/13788) moq-ffi test::an_unpaced_writer_evicts_a_slow_reader
        PASS [   0.120s] ( 2194/13788) moq-ffi test::audio_raw_publish_from_many_threads
        PASS [   0.777s] ( 2195/13788) moq-ffi session::tests::new_rejects_invalid_config
        PASS [   1.291s] ( 2196/13788) moq-ffi media::timebase_tests::containers_of_one_source_keep_their_relative_media_time
        PASS [   1.185s] ( 2197/13788) moq-ffi session::tests::setting_tokens_refuses_an_empty_token
        PASS [   1.191s] ( 2198/13788) moq-ffi session::tests::new_accepts_the_defaults_and_known_versions
        PASS [  13.727s] ( 2199/13788) moq-cli::integration rtsp::every_rtsp_session_is_a_fresh_broadcast
        PASS [  14.636s] ( 2200/13788) moq-cli::integration listen::an_export_stage_ending_during_drain_waits_for_every_session
        PASS [   2.879s] ( 2201/13788) moq-ffi server::lifecycle_tests::accept_on_a_foreign_thread
        PASS [  17.332s] ( 2202/13788) moq-cli::integration import::a_short_import_drains_its_media_to_a_waiting_subscriber
        PASS [  17.703s] ( 2203/13788) moq-cli::integration import::import_delivers_the_catalog_finish_at_eof
        PASS [  17.701s] ( 2204/13788) moq-cli::integration report::import_reports_count_what_each_rendition_wrote
────────────
     Summary [  28.268s] 2204/13788 tests run: 2203 passed, 1 failed, 151 skipped
        FAIL [   1.193s] ( 2181/13788) moq-ffi media::read_tests::raw_media_catalog_warmup_reaches_the_subscription
  stdout ───

    running 1 test
    test media::read_tests::raw_media_catalog_warmup_reaches_the_subscription ... FAILED

    failures:

    failures:
        media::read_tests::raw_media_catalog_warmup_reaches_the_subscription

    test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 321 filtered out; finished in 1.18s

  stderr ───

    thread 'media::read_tests::raw_media_catalog_warmup_reaches_the_subscription' (1988497) panicked at rs/moq-ffi/src/media.rs:1635:13:
    assertion `left == right` failed: video=false
      left: [4400000, 5000000]
     right: [4400000, 4700000, 5000000]
    note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

warning: 11584/13788 tests were not run due to test failure (run with --no-fail-fast to run all tests, or run with --max-fail)
error: test run failed
mbx[cache]: object cache: 44 hits, 179 misses, 90 bypassed; 0 B downloaded, 0 B uploaded, 6.2 GiB stored locally
mbx[cache]: 6m 12s session; 1m 57s estimated compiler time avoided (summed across compilations)
mbx[cache]: bypass reasons: 89 unsupported-crate-type, 1 unportable-native-link
mbx[cache]: Cargo artifact reuse and CI cache archive transfers are not included above
mbx[savings]: has quietly disposed of 13.7 TiB logical of build leftovers. nobody saw anything.
error: recipe `test` failed on line 367 with exit code 100
error: recipe `test-changed` failed on line 86 with exit code 100
##[endgroup]
##[error]ci: test rs failed
ci: seconds per module
   192  check rs
   372  test rs (failed)
     5  check biome
   191  check js
    17  test js
    10  check markdown
    13  check gh
   323  test gh
    46  check flake
     1  check drill
     5  check quest
     1  check ts_paths
    27  check shell
     0  check large_files
     0  check text
     1  check toml
     0  check nix
     0  check justfile
   564  wall time
error: recipe `ci` failed on line 119 with exit code 100
##[error]Process completed with exit code 100.
##[group]Run actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9
with:
  name: cargo-timings
  path: target/cargo-timings/cargo-timing-*.html
  if-no-files-found: ignore
  retention-days: 14
  compression-level: 6
  overwrite: false
  include-hidden-files: false
  archive: true
env:
  BASE: 79960a9dd64e48a2fe51a06fc1a3e9b4692dfb22
  CHECK_FAIL_FAST: 1
  NEXTEST_TEST_THREADS: 24
  MBX_CACHE_LINKS: 1
  MBX_REMOTE_URL: 
##[endgroup]
With the provided path, there will be 4 files uploaded
Artifact name is valid!
Root directory input is valid!
Uploading artifact: cargo-timings.zip
Beginning upload of artifact content to blob storage
Uploaded bytes 219686
Finished uploading artifact content to blob storage!
SHA256 digest of uploaded artifact is b7ac63dd3f4323c840a5e56bc91dc5d4d620ffc380c7b45cec9393539d9fd904
Finalizing artifact upload
Artifact cargo-timings successfully finalized. Artifact ID 13626696125
Artifact cargo-timings has been successfully uploaded! Final size is 219686 bytes. Artifact ID is 13626696125
Artifact download URL: https://github.com/moq-dev/moq/actions/runs/42502335110/artifacts/13626696125
Post job cleanup.
##[start-action display=Run jdx/mr-boxington-action@88acf2bb50552c7c508cbac13e5f8c576b95b1de;id=__02ee65a1-2453-4945-9f1d-3fda0167bbcd.mbx]
##[end-action id=__02ee65a1-2453-4945-9f1d-3fda0167bbcd.mbx;outcome=skipped;conclusion=skipped;duration_ms=0]
Post job cleanup.
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git version
git version 2.55.0
Temporarily overriding HOME='/var/cache/github-runner-work/_temp/13e73b94-5cad-4daf-b809-abc6517733ab' before making global git config changes
Adding repository directory to the temporary git global config as a safe directory
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git config --global --add safe.directory /var/cache/github-runner-work/moq/moq
Removing SSH command configuration
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git config --local --name-only --get-regexp core\.sshCommand
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git submodule foreach --recursive sh -c "git config --local --name-only --get-regexp 'core\.sshCommand' && git config --local --unset-all 'core.sshCommand' || :"
Removing HTTP extra header
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git config --local --name-only --get-regexp http\.https\:\/\/github\.com\/\.extraheader
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git submodule foreach --recursive sh -c "git config --local --name-only --get-regexp 'http\.https\:\/\/github\.com\/\.extraheader' && git config --local --unset-all 'http.https://github.com/.extraheader' || :"
Removing includeIf entries pointing to credentials config files
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git config --local --name-only --get-regexp ^includeIf\.gitdir:
[command]/nix/store/lg46w7hrjx7kylsh6645c9l975i8h43b-git-2.55.0/bin/git submodule foreach --recursive git config --local --show-origin --name-only --get-regexp remote.origin.url
A job completed hook has been configured by the self-hosted runner administrator
##[group]Run '/nix/store/gx2v6q84b0w43bafm0whkdx93qyxbpps-ci-cache-gc.sh/bin/ci-cache-gc.sh'
shell: /nix/store/1mv3qz005gkbalxfghk4y3s8ayv4d2dl-bash-interactive-5.3p15/bin/bash --noprofile --norc -e -o pipefail {0}
##[endgroup]
evicted 0 objects and 0 action results (0 B logical); 50.0 GiB logical remain
removed 0 learned incremental units and 2 checkout directories (3.7 KiB logical); 1 units (220.9 MiB logical) remain
evicted 0 objects and 0 action results (0 B logical); 50.0 GiB logical remain
removed 3 target directories (34.4 GiB logical, 2 abandoned and 1 live); 0 B logical remain
Cleaning up orphan processes

(written by quest land watch)

[moq-bot]

codex, gpt-6.1-sol, medium effort land fixer note:

The PR merges cleanly with base on a scratch branch. Its diff does not change the failing FFI test or its container read path; the log does not establish that this PR caused the failure. No blame assigned and no code changed.

Failure log lines 4629–4648 report:

FAIL moq-ffi media::read_tests::raw_media_catalog_warmup_reaches_the_subscription
assertion `left == right` failed: video=false
  left: [4400000, 5000000]
 right: [4400000, 4700000, 5000000]

I ran just ci base on the scratch merge, but it reported: ci: flake's NixOS VM tests need KVM, but /dev/kvm is missing; run them on a KVM host, as CI does. I stopped the remaining checks after that validation blocker. A targeted just rs test -p moq-ffi -E 'test(=media::read_tests::raw_media_catalog_warmup_reaches_the_subscription)' was also stopped during compilation, so there is no local reproduction or passing result. Complete validation and determine the original failure's cause on a KVM runner.

(written by codex, gpt-6.1-sol, medium effort)

[moq-bot]

codex, gpt-6.1-sol, medium effort land fixer: Queue warmup test lost the 4700000 frame; validation cannot complete: just ci base reports "ci: flake's NixOS VM tests need KVM, but /dev/kvm is missing; run them on a KVM host, as CI does".. Recommendation: Run the warmup test on base alone and repeat the full gate on a KVM runner to establish base versus intermittent failure before requeueing.

        FAIL [   1.193s] ( 2181/13788) moq-ffi media::read_tests::raw_media_catalog_warmup_reaches_the_subscription
  stdout ───

    running 1 test
    test media::read_tests::raw_media_catalog_warmup_reaches_the_subscription ... FAILED

    failures:

    failures:
        media::read_tests::raw_media_catalog_warmup_reaches_the_subscription

    test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 321 filtered out; finished in 1.18s

  stderr ───

    thread 'media::read_tests::raw_media_catalog_warmup_reaches_the_subscription' (1988497) panicked at rs/moq-ffi/src/media.rs:1635:13:
    assertion `left == right` failed: video=false
      left: [4400000, 5000000]
     right: [4400000, 4700000, 5000000]
    note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

(written by codex, gpt-6.1-sol, medium effort)