Repository navigation
quest: audit the tree against main (2026-10-22) - #6273
Conversation
Correct facts that main has moved past: landed PRs, closed PRs, moved files, and finished quests. Delete m1/linger-ffi (landed in #6157) and its references, drop Required entries that have landed, and remove customer wording from lazy-discovery. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Identity proof extends the merged AUTH stream instead of Requiring token in band; three quests leave the C line's Required rule; lite-07 finalize waits on timestamp provenance and lite-06 max sequence, and lite-07 ranges carry a relative start's resolution. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Fold the Opus near-zero budgets into capped budgets, delete the window allocation and faster-runner quests, narrow the R2 and Nix cache checks to what still uses them after the queue moved in house, and record the remaining maintainer decisions and stale facts in their quests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Correct m2 facts that main or upstream has moved past, and record the accepted decisions: TS hitless waits on stream routes and arrival stamps, the io_uring and QUIC sim quests are rescoped, qmux reports acks as unsupported, and the UniFFI namespace issue is noted as unfixed upstream. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Warning Review limit reachedYou'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 57 seconds. View limit details
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. Comment |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c87627fa92
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
…pires Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Take main's deletions of finished quests and #6268's wording where both sides edited, drop links to quests main finished, and keep the window-alloc removal. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Merged main at c1e2be6: took main's deletions of finished quests and #6268's wording where both sides edited, dropped links to finished quests, and kept the window-alloc removal (details under "Merged with main" in the body). Codex reviewed c1e2be6 with no findings; the Codex thread on nix-cache-verify was fixed in 93d8a31 and is resolved. Auto-merge is enabled on this head. (Written by Claude Opus 5.5) |
Take main's deletion of the finished lazy-discovery epic (#6165). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 90fcd56780
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Audits the quest tree against
mainon 2026-10-22: facts that main has moved past are corrected, finished quests are deleted, and landed blockers leaveRequired. Quests with in-flight PRs are left alone.Public API: none. Wire: none. Only
quest/and one spike report (test/wasm/generator.md) change.Changes
m1/refresh: drop "flate modes" from the non-goals (compress removes moq-flate); the passthrough Related line names the NEW_GROUP_REQUEST relay (#5708) that #5930 promoted.m1/README.md: refresh's ranking note no longer cites the passthrough path; theRelatedsection linking merged #5289 is gone; lite-time-provenance moves above moq-transport passthrough, since serve-timed requires it.m1/quic/receive-memory: #5878 is ready for review, onquest/m1/quic/receive-memory-lease.m1/test-flakes-2: the 2026-10-20 batch landed (#6127, #6125, #6056); the 2026-10-21 child holds two flakes, since RTT and interop landed in #6213.m1/quic: native probing is m1 (probe-native.md); only padding and the browser rung disable stay in the m2 probe epic. The realtime and L4S pointers are kept.m1/lazy-discovery: customer wording removed (public repo rule).m1/ts-reliability: raw export and its packet-level proof landed in #5856.m0/README.md: #4741 is on main; the Goal drops "two independent tracks" and names the release-line, stall, CI, and audit work it lists.m0/broadcast-epoch:Relatedremoved (only merged #5254).m0/ci/nix-cache-verify: Cachix retirement includessmoke.yml's substituter and push step (.github/workflows/smoke.yml:195-225).m2/release/{cpp-release,dart-publish}: the 2026-10-10 wait on Codecs and the epoch rename is gone (landed #5694, #5839, #5146); OBS parity landed in #5726 and #6131. Dart keeps itsmoq_ffifloor bullet.m2/intra-refresh/consumer-warmup: the rung fetches N-K through N as one range subscription; Related links fetch-range.m2/net-wrap/traits: the track dynamic wrapsrequested_fill/track::Fill, perm1/subscribe-ranges/dynamic.md.m2/qos: prefix tracks landed in #5811; only the lag histogram is deferred.m2/README.md: the QoS line names the egresslaghistogram anddroppedcounters.m2/route-health: since #6183, resume recovers what max age still holds, so a max age below the idle timeout still loses media;design.mdstarts from fix(net): a resume through a silent outage keeps the backlog its max age holds #6183's mock-time diamond (route_change.rs).m2/rs2ts/wasm-bindingsandtest/wasm/generator.md: moq-ffi reports the negotiated version (MoqConnectionStats.version, #5723).m2/stats/rust:rtspis among the refused gateway imports (stats_refusalinrs/moq-cli/src/args.rs).m3/teleop-robot: the group caps hold until unbounded groups lands; then a lagging reader refills or seesEvicted. Related links that epic.m1/smoke-moxygen: reported upstream as facebookexperimental/moxygen#225, acknowledged 2026-10-16.m1/resolved-demand: #5931 closed unmerged; its listed-track workaround never reached main.m1/playwright-nix: npm pins 1.64.1, matching the current nixpkgs lock (the 1.63 downgrade is stale).m1/linger-ffi.mddeleted (landed in #6157), with its references inm1/README.md,m1/interop-finish.md,m1/c/README.md, andm2/go-first-dial-failure.md. This overlaps open #6229.m1/interop-finish:Requireddropped; Rust hasbroadcast::Producer::finish, JS landed in #6151, bindings in feat(ffi): finish a broadcast cleanly in every binding #6157.m1/close-tail-drain: the quic-qmuxRequiredis gone; the maintainer took the upstream fix (moq-dev/web-transport#437, qmux 0.6.5), per #5762.m1/audit-ci: Features clippy was fixed by #6117 (the 10-21 failure is tests);gh testfailed again 10-21, then #6154 removed the assertion;rs audio-devicesis owned byaudio-devices-arm.md; C++/Gradle cost bybinding-check-cost.md(#6247). Kept minimal since draft #6266 rewrites this file.m1/audio-devices-arm: Nightly 39600361153 (10-21, before #6175 merged) already passed on ARM after the 10-20 failure, so one green nightly proves nothing.m1/audit-size: #6090 restored the Linux release links; the firstsizesrun is the next Nightly.m1/listen-admission-test: the test lives atrs/moq-cli/tests/integration/listen.rs.m2/audit-docs: 31 crates set#![warn(missing_docs)], including moq-mux, moq-ffi, and moq-c since #6088 and #6091.m2/drop-dead-items: theassert_next_*helpers stay;model/origin.rscalls them since #5678.m2/bench-sources(now[S]): the binding, Rust bench target, and JS sources already feed the table (#5759); therust-testjob and optional series remain.m2/audio-decode-audiotoolbox: the docs bullet names thedecoderow indoc/lib/rs/moq-audio.md.m2/go-first-dial-failure: the wait on the linger and auth binding PRs is gone (#6157, #6137).m2/publish-announce-never:Closesdropped (#5703 closed as not planned; the Plan still cites it).m2/mux-vr-fields: Add spherical and stereo video metadata (sv3d, st3d, vexu, hfov) kixelated/mp4-atom#244 addedst3d/sv3d; take 0.17.0 (chore: release v0.17.0 kixelated/mp4-atom#247). Note: 0.17.0 is not on crates.io yet (latest 0.16.2, chore: release #247 open), so the quest says "not yet released".m2/starvation-frames: the group sampler is the lag histogram, whose code is onwip/4133-lag-histogram.m2/rs2ts-wasm2js: Charon #1513 and #1514 merged upstream; pin a revision carrying them.m2/rs2ts-js: the Bind the browser through moq-ffi/UniFFI instead of a second hand-written wasm API #2907 Related entry and the old-WASM-issues sentence are gone.m2/render-d3d11: the renderer is the first consumer ofd3d11::Texture'spub(crate)contents.m2/webgpu-hdr: the catalog colour fields landed in #5846.Left alone:
m1/quic-qmux.md(open #6231 deletes it),m0/redundant-failover-drift.md(open #6113 deletes it), and the quests with in-flight agents.Decisions
Round 1, asked explicitly:
m1/cluster-routing/identity.md)m1/c/README.md)m1/archive/subscribe-tracks-outbound, in flight, not edited here; forwarded to its agent)m1/lite07-finalize.md)Accepted as recommended:
m1/audio-jitter-target/capped-budgets.md: one pass over seven post-cap nightlies, which are also post-flush (#6164 merged 2026-10-22T00:10Z; none yet).m1/compress/migrate.mdRequires refresh JSON; both rewritesnapshot::Configand moq-stats'TrackPair(#6054).m2/intra-refresh: plan-boundary amends the hang draft so a parallel max-delay decoder MAY present before group start pluswarmup(keeps merged #5842); h264-import roundswarmupup per the draft (#5212).m2/net-wrap/traits.mdRequires the breaking release cut;e2ee.mdlinks compressed tracks.m1/simulcast-rung-disable.mdandm2/probe/rung-disable-js.md.m2/perf/relay-allocs.mdRequires dropping web-transport-trait, then re-profiles.m0/ci/mbx-r2-verify.mdnarrowed to hosted jobs on R2 (Nightly, release PRs, macOS/Windows); the queue runs on self-hostedmoq-queuewith local mbx since #6040.m0/ci/README.mdupdated.m0/ci/nix-cache-verify.md: ci: add reusable runner VMs with persistent local caches #6040 deletedcache.yml, so first measure a hosted Nightly's dev-shell entry; restore a dev-shell push only if it rebuilds, else narrow to release tags and Cachix retirement.m0/ci/runners.mddeleted: the persistentnix/runnerVM is the queue runner (median 5.5 min per entry against 21 hosted), recorded inm0/ci/README.md; Blacksmith dropped. Nothing remained.m1/ts-reliability/arrival-stamps.md: after fix(mux): redundant TS failover survives a killed importer behind bufferbloat #6113, the container consumer's hole budget anchors on stamped arrivals.moq import --echo-pathinm2/stats/rust.md, anecho-pathattribute inm2/stats/js.md;m3/stats-encoder-feedback.mdkeeps only the adaptation loop.m2/site/move.md: moq.dev's biome-local and lean-agents-md are moot; astro-typecheck ports into thesiteCI module.m1/window-alloc.mddeleted (#6241 measured a join regression);m1/audit-perf.mdrecords why.m1/resume-rust-end.md: Rust refuses a different end with JS'sIncompatibleclassification, settled in #6224's review; merge order fix(net): end a resumed track the way a direct one ends #6136, then fix(net): deduplicate resumed datagrams by identity #6209 and fix(net): refetch a resumed group only when its route failed delivering it #6251, then fix(net): retain resume bounds across route copies #6224.m1/audio-devices-arm.mdis deleted only after three consecutive green Nightlies.m1/catalog-identity-rs.mdnotes its branch checkpoint predates the 10-15 rendition-name decision.m1/binding-timestamps.md: the value bound followsTimestamp::new(2^61 - 1 once #6168 lands).m2/audit-docs.md: the Go failure accessor is planned in go-first-dial-failure.m2/error-cleanups.md: keep the removal;model/track.rs'scloseconstructsStreamError::Session(SessionError::Cancel)directly.m1/track-priority-scope.md: no edit; open #6142 replaces it.m2/ts-hitless.md: keeps Closesexport ts --sync: a group that began behind the one before it lays apart on a leg that lost that earlier group #5864; Requires stream routes (#6252) and arrival stamps (#6249) before rebasing #5949;--synclanes cover or refusepesPacketandverbatim; release follows arrival stamps' lateness rule.m2/uring-transport-machines.mdrescoped to measuringLocal's thread check after #6176;Boxablestays the single bound.m2/quic-ack-hook.md: qmux'spoll_ackedreports unsupported; over qmux the histogram stays at group granularity.m2/quic-sim.mdrescoped to MoQ sessions and the relay overmoq-quic'sPair/Linksim (#5780, #5989).m2/uniffi-{namespaces,fork}.md: Library mode: a type cannot reference an external type with the same name mozilla/uniffi-rs#3026 was auto-closed by our #5498 and remains unfixed upstream.Housekeeping done outside the tree
Pending maintainer action
Blocked by permissions:
quest/m2/e2ee/rust-publishtoref/e2ee-mux-bridge. The e2ee README already links #5822 by PR, so it needs no edit.export ts --sync: a group that began behind the one before it lays apart on a leg that lost that earlier group #5864" to #5949.Follow-ups
test/wasm/uniffi/cases.tsandtest/wasm/baseline/consumer.tsstill comment that moq-ffi exposes no negotiated version; the cases could readMoqConnectionStats.versioninstead of offering one version.Merged with main (c1e2be6)
After #6268 and the night's merges, main's deletions and wording win where both sides edited:
m1/close-tail-drain(25),m1/listen-admission-test(29),m1/lite-time-provenance(2, D),m1/playwright-nix(22),m1/reference-epoch-bindings(B),m1/ts-reliability/arrival-stamps(N),m2/ts-hitless(Y).m1/linger-ffi(23) went in #6229.m1/refreshpassthrough line (1),m1/audit-cibinding cost (26),m1/audit-size(28),m1/interop-finish(24),m2/audit-docsGo wording (30's crate count stays),m2/bench-sources(32 dropped, stays[M]),m3/teleop-robotcaps (19's unbounded-groups sentence stays).m1/c/README: B's paragraph dropped; reference epochs (#6244) and linger (#6157) have merged, and main already records the dynamic-ranges exception.m1/README,m1/lite07-finalize(now waits only on lite-06 max sequence; provenance landed in #6150),m1/lite06-max-sequence, andm1/release-perf/recheck(window-alloc: #6241 regressed join time).m0/READMEGoal drops the publisher stall estimate (#6218);m1/resume-rust-endmerge order is fix(net): end a resumed track the way a direct one ends #6136 then fix(net): retain resume bounds across route copies #6224 (fix(net): deduplicate resumed datagrams by identity #6209 and fix(net): refetch a resumed group only when its route failed delivering it #6251 landed);m1/test-flakes-2notes only the OBS stop drain remains.quest checkandquest check --base origin/main --branch quest/audit-2026-10-22pass.(Written by Claude Opus 5.5)
🤖 Generated with Claude Code