From f0e85662da65581d7eef772b819ceed4fbf81e42 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Tue, 27 Oct 2026 21:47:20 -0700 Subject: [PATCH 1/4] quest: plan follow-ups from the main-vs-release audit Add interop nightly and recurring quinn sync quests, and bound the export jitter buffer by bytes as well as time. Co-Authored-By: Claude Opus 5.5 --- quest/m0/README.md | 2 +- quest/m0/nightly/README.md | 5 ++- quest/m0/nightly/interop.md | 41 +++++++++++++++++++++ quest/m1/README.md | 1 + quest/m1/cut-audit/README.md | 2 +- quest/m1/cut-audit/jitter-admission.md | 9 +++++ quest/m1/quinn-sync.md | 51 ++++++++++++++++++++++++++ 7 files changed, 107 insertions(+), 4 deletions(-) create mode 100644 quest/m0/nightly/interop.md create mode 100644 quest/m1/quinn-sync.md diff --git a/quest/m0/README.md b/quest/m0/README.md index 8357acee985..6c51c188f83 100644 --- a/quest/m0/README.md +++ b/quest/m0/README.md @@ -20,7 +20,7 @@ edge, which embeds moq-relay; it moved there on 2026-09-28. ## Required -- [Nightly passes on main](/quest/m0/nightly/README.md) - the scheduled Nightly passes end to end, so red means a new break +- [Nightly passes on main](/quest/m0/nightly/README.md) - the scheduled Nightly and Interop runs pass end to end, so red means a new break - [Release shutdown backport](/quest/m0/ffi-shutdown-backport.md) - `release` publishes a non-breaking moq-ffi with an async `drain()` and qmux sessions on the QUIC stream limit - [Backport IETF close codes](/quest/m0/backport-ietf-close-codes.md) - `@moq/net` on `release` closes moq-transport sessions with the right error for three hostile cases - [Backport @moq/flate bounds](/quest/m0/backport-flate-budget.md) - `@moq/flate` on `release` stops inflating past its cap and stops pinning pako scratch diff --git a/quest/m0/nightly/README.md b/quest/m0/nightly/README.md index a153222320e..b75c3a8513c 100644 --- a/quest/m0/nightly/README.md +++ b/quest/m0/nightly/README.md @@ -2,8 +2,8 @@ ## Goal -The scheduled Nightly on `main` passes end to end, so a red run means a new -break instead of noise. Nightly is the backstop for what the PR gate skips on +The scheduled Nightly and Interop runs on `main` pass end to end, so a red +run means a new break instead of noise. Nightly is the backstop for what the PR gate skips on purpose (benches, examples, feature shapes; see [moq-dev/moq#5531](https://github.com/moq-dev/moq/pull/5531)), and it has failed every night since at least 2026-10-19, which let a bench and two @@ -34,3 +34,4 @@ Public API: none. Wire: none. - [moq-net compiles in Fuzz and send-depth](/quest/m0/nightly/send-overflow.md) - the ietf publisher future stops overflowing the Send check - [Probing beside Cubic](/quest/m0/nightly/competing-cubic.md) - the `competing_cubic` probe test stays within its deadline budget - [cargo deny passes](/quest/m0/nightly/cargo-deny.md) - Audit stops failing on the yanked crypto-bigint and moq-studio's egui licenses +- [Interop passes on main](/quest/m0/nightly/interop.md) - re-measure the stale wire-compat pin, fix the draft-14 harness test, and route each red cell to its owner diff --git a/quest/m0/nightly/interop.md b/quest/m0/nightly/interop.md new file mode 100644 index 00000000000..b56551707d7 --- /dev/null +++ b/quest/m0/nightly/interop.md @@ -0,0 +1,41 @@ +# [M] Interop passes on main + +## Goal + +The scheduled Interop workflow (`.github/workflows/interop.yml`) passes on +`main` again, so a red run means a new break. It has failed every night +since 2026-10-22; the last green run was 2026-10-22 11:47. + +## Plan + +Triage of run +[40517499786](https://github.com/moq-dev/moq/actions/runs/40517499786) +(2026-10-27) and the two before it. Four steps fail: + +- **Released wire compatibility**: `stale planned break: moq-lite-07-wip, + @moq/net changed from 0.4.5 to 0.4.6`. `@moq/net` 0.4.6 published on + 2026-10-23 and nobody re-measured. Run `just test wire-compat --remeasure`, + review the rewritten `test/interop/compat/planned-breaks.json`, and refresh + its reason. A patch release between cuts is not covered by + [the cut](/quest/m2/release/cut.md)'s re-measure step, so also find where + CONTRIBUTING.md's release procedure should say to re-measure after any + publish of a pinned package, and fix that inline. +- **Harness**: `test/interop/clients/js/publish-ui.browser.ts:390` still + sets `moq-transport-14` on a ``, which + [moq-transport 18 and 22 only](/quest/m1/ietf-drop/README.md) removed, so + `checkVersions` refuses it. Update the expected versions. +- **`interop --all`**: the failing cell changes each night (`cpp -> + js-native-node` tone "timestamp gap", `go -> c`, `rust -> js` past its + 30 s budget, `goaway rust <-> js-native-*`). Route each to its owner: + FFI-publisher tone gaps belong to + [FFI publisher audio gaps](/quest/m0/test-flakes-2/ffi-audio-gap.md). + Fix the rest here at the cause, or plan a quest per cause through + `/quest-plan` if one is large. +- **Tail**: `tail js-native-* -> rust` belongs to + [JavaScript FETCH](/quest/m1/js-fetch.md); confirm it still fails only + for that reason. + +Done when a scheduled Interop run passes, or every remaining red cell has +an owning quest in m0. + +Public API: none. Wire: none. diff --git a/quest/m1/README.md b/quest/m1/README.md index 9c356f35460..d9aa65bf2c6 100644 --- a/quest/m1/README.md +++ b/quest/m1/README.md @@ -116,6 +116,7 @@ surface and zero-copy texture import, land before the cut. - [Address groups by timestamp](/quest/m1/time-addressing/README.md) - group start timestamps replace sequence numbers in the moq-net API and lite-07 wire, while old subscribers, old publishers, and published JS clients keep working - [Release audit findings](/quest/m1/release-audit/README.md) - resolve newly identified resource, parser, playback, and interoperability risks before the cut - [Pre-cut audit findings](/quest/m1/cut-audit/README.md) - fix the crashes, hostile-input costs, and correctness bugs the 2026-10-24 audit of main found +- [moq-quic has every quinn fix it needs](/quest/m1/quinn-sync.md) - quinn's fixes since the fork point are ported with their tests, once per release cycle - [Cancelled moq-transport requests stop](/quest/m1/ietf-request-cancel.md) - a cancelled request releases its state even when the peer withholds credit - [Cluster routing](/quest/m1/cluster-routing/README.md) - prove shared routing contracts, implement per-origin routes, and connect outbound local clusters - [Routing guide](/quest/m1/routing-guide.md) - concepts and practical outbound federation setup diff --git a/quest/m1/cut-audit/README.md b/quest/m1/cut-audit/README.md index 6773b915fc7..64afdc87d4b 100644 --- a/quest/m1/cut-audit/README.md +++ b/quest/m1/cut-audit/README.md @@ -60,7 +60,7 @@ leaving JS-only redesigns of `js/net` to [rs2ts](/quest/m2/rs2ts/README.md). - [Withdrawing a moq-transport FETCH or TRACK_STATUS ends it](/quest/m1/cut-audit/ietf-withdraw-resolve.md) - withdrawal during resolution releases the request - [A description-less AAC rendition means one thing](/quest/m1/cut-audit/aac-description-less.md) - every player and exporter agrees on the frame format - [RTMP sessions follow one strict path](/quest/m1/cut-audit/rtmp-session-strict.md) - idle sockets, repeated requests, rewinds, and stalled servers are handled -- [The export jitter buffer admits only frames due within its window](/quest/m1/cut-audit/jitter-admission.md) - a fast track is backpressured, not buffered without bound +- [The export jitter buffer admits only frames due within its window](/quest/m1/cut-audit/jitter-admission.md) - a fast or dense track is backpressured by time and bytes, not buffered without bound - [Peer-driven receive loops yield on a per-message budget](/quest/m1/cut-audit/peer-loop-budget.md) - small-message floods cannot monopolize a session driver - [Removing a stale front is O(1)](/quest/m1/cut-audit/weak-cache-remove.md) - a restart wave is linear under the origin lock - [Rust and JS accept the same catalogs](/quest/m1/cut-audit/catalog-parity.md) - one strictness, per rendition, with a shared fixture corpus diff --git a/quest/m1/cut-audit/jitter-admission.md b/quest/m1/cut-audit/jitter-admission.md index f9590ab10de..caebd2f784d 100644 --- a/quest/m1/cut-audit/jitter-admission.md +++ b/quest/m1/cut-audit/jitter-admission.md @@ -30,6 +30,15 @@ frames pile up behind it, and `Stream::fill` keeps reading while `jitter.full(name)` is false. No frame-count cap. Alternatively charge a track whose lead keeps growing as a source fault. +Bound bytes too (decided 2026-10-27, maintainer, from a second audit of +main against release). Timestamp distance does not bound memory: frames +that share one timestamp all fall inside the window, and +[container backpressure](/quest/m1/container-backpressure.md)'s budget stops +at frames the jitter buffer has not pulled yet. So a track is also full while +its queued bytes exceed a per-track byte budget. Like the time bound, it +holds the track back in its source as backpressure, never refuses a frame. +Test a track that packs many frames into one timestamp beside an honest one. + Then evaluate the larger simplification: about a dozen fixes to the export clock (f85547ffd0, 5d5974f083, 3fdf46160b, e84ffb3de3, a1465d0d00, 11bfa30461, 3cf978cdbf, 493d8eee99, 7a3890cd30, 3b77d23015, 7c4c48d08b, diff --git a/quest/m1/quinn-sync.md b/quest/m1/quinn-sync.md new file mode 100644 index 00000000000..bf1ae5b5b4d --- /dev/null +++ b/quest/m1/quinn-sync.md @@ -0,0 +1,51 @@ +# [S] moq-quic has every quinn fix it needs + +## Goal + +moq-quic, moq-sock's `udp`, and moq-tokio's `quic::driver` carry every +security and correctness fix quinn has landed since the last sync, each +ported with its regression test or recorded as not applying. The fork +ledger in `rs/moq-quic/README.md` says how far upstream was checked. + +## Plan + +Decided 2026-10-27 (maintainer), from a second audit of main against +release: the hard fork never merges upstream and cargo-deny cannot see a +quinn advisory in in-tree code, so checking upstream is recurring work. + +Each run: + +- Compare quinn-rs/quinn `main` against the last checked commit (the fork + point [`7616e6b2`](https://github.com/quinn-rs/quinn/commit/7616e6b2782722f3ce4a1b181ef08817d7f545e4) + for the first run), plus quinn's GitHub security advisories. +- Port each fix that applies to code the fork still carries, with its + regression test, and list it under the README's ported commits. Record + one that does not apply with its reason, as the README already does. + Skip dependency bumps. +- Record the newest upstream commit checked in the README, so the next run + starts there. + +First run: three fixes on quinn `main` are not in the ledger yet: +[`f233510c`](https://github.com/quinn-rs/quinn/commit/f233510c) (honor a +disabled idle timeout when accepting), +[`6c69f9f4`](https://github.com/quinn-rs/quinn/commit/6c69f9f4) (don't +credit a reset stream's unread data twice on stop: flow-control accounting a +peer drives), and +[`623a0369`](https://github.com/quinn-rs/quinn/commit/623a0369) (don't make +specific IPv6 endpoints dual-stack). + +Recurrence: the PR that finishes a run deletes this file and creates the +next run at `/quest/m2/quinn-sync.md`, linked from `/quest/m2/README.md`'s +`Required`. It keeps this Goal and recurrence paragraph, carries forward +what this run learned, and requires [the release cut](/quest/m2/release/cut.md). +The cut removes itself from every `Required` and shifts m2 into m1, so the +next run becomes ready in the new m1, once per release cycle. Only +`/quest-delete` ends the recurrence. + +Public API: none expected; a ported fix that changes one says so. Wire: +QUIC behavior follows upstream's fixes. + +## Related + +- [moq-quic's API improves on quinn's](/quest/m1/quic-api/README.md) - renames that make each port a hand translation +- [Dependency audit](/quest/m2/audit-deps.md) - covers crates cargo-deny can see From 76499520d8b003be4958a7f01459521c26e50847 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Tue, 27 Oct 2026 21:53:58 -0700 Subject: [PATCH 2/4] quest: check each quinn candidate against the fork before porting Co-Authored-By: Claude Opus 5.5 --- quest/m1/quinn-sync.md | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/quest/m1/quinn-sync.md b/quest/m1/quinn-sync.md index bf1ae5b5b4d..fb9213730ad 100644 --- a/quest/m1/quinn-sync.md +++ b/quest/m1/quinn-sync.md @@ -25,12 +25,14 @@ Each run: - Record the newest upstream commit checked in the README, so the next run starts there. -First run: three fixes on quinn `main` are not in the ledger yet: +First run: three fixes on quinn `main` are not in the ledger yet. Check +each against the fork's code before porting, since a missing ledger entry +does not prove the bug remains: [`f233510c`](https://github.com/quinn-rs/quinn/commit/f233510c) (honor a disabled idle timeout when accepting), [`6c69f9f4`](https://github.com/quinn-rs/quinn/commit/6c69f9f4) (don't -credit a reset stream's unread data twice on stop: flow-control accounting a -peer drives), and +credit a reset stream's unread data twice on stop; the fork rewrote `stop` +under the README's "Stream credit on stop", so it may already hold), and [`623a0369`](https://github.com/quinn-rs/quinn/commit/623a0369) (don't make specific IPv6 endpoints dual-stack). From 1cee2d10b59b796e5bbc42dae19f382fd69491f5 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Thu, 29 Oct 2026 08:23:50 -0700 Subject: [PATCH 3/4] Merge main; drop interop.md in favor of interop-green main's #7003 planned the same Interop failures as quest/m0/interop-green, so this branch's nightly/interop.md duplicated it. CONTRIBUTING already requires re-measuring wire-compat after any publish. Co-Authored-By: Claude Opus 5.5 --- quest/m0/README.md | 2 +- quest/m0/nightly/README.md | 5 ++--- quest/m0/nightly/interop.md | 41 ------------------------------------- 3 files changed, 3 insertions(+), 45 deletions(-) delete mode 100644 quest/m0/nightly/interop.md diff --git a/quest/m0/README.md b/quest/m0/README.md index 45a4fc92daf..c349fd57fa1 100644 --- a/quest/m0/README.md +++ b/quest/m0/README.md @@ -21,7 +21,7 @@ edge, which embeds moq-relay; it moved there on 2026-09-28. ## Required - [Adopt local reviews](/quest/m0/local-review.md) - PRs are reviewed locally by the other vendor's model before they are pushed, and merging never waits on an online reviewer -- [Nightly passes on main](/quest/m0/nightly/README.md) - the scheduled Nightly and Interop runs pass end to end, so red means a new break +- [Nightly passes on main](/quest/m0/nightly/README.md) - the scheduled Nightly passes end to end, so red means a new break - [Release credential leaks](/quest/m0/security-audit/README.md) - tokens stay out of logs and outbound HTTP clients verify their peer, on `main` and `release` - [Interop and Video pass on main](/quest/m0/interop-green/README.md) - the scheduled Interop and Video workflows pass, so red means a new break - [Release shutdown backport](/quest/m0/ffi-shutdown-backport.md) - `release` publishes a non-breaking moq-ffi with an async `drain()` and qmux sessions on the QUIC stream limit diff --git a/quest/m0/nightly/README.md b/quest/m0/nightly/README.md index 7e5e22ccc5c..fac16fe2dc5 100644 --- a/quest/m0/nightly/README.md +++ b/quest/m0/nightly/README.md @@ -2,8 +2,8 @@ ## Goal -The scheduled Nightly and Interop runs on `main` pass end to end, so a red -run means a new break instead of noise. Nightly is the backstop for what the PR gate skips on +The scheduled Nightly on `main` passes end to end, so a red run means a new +break instead of noise. Nightly is the backstop for what the PR gate skips on purpose (benches, examples, feature shapes; see [moq-dev/moq#5531](https://github.com/moq-dev/moq/pull/5531)), and it has failed every night since at least 2026-10-19, which let a bench and two @@ -35,6 +35,5 @@ Public API: none. Wire: none. - [The browser backend baseline completes on IETF22](/quest/m0/nightly/browser-baseline.md) - the WASM and UniFFI backend baseline completes every supported transport case without an internal error - [Resolve the paired catalog fanout regressions](/quest/m0/nightly/paired-shared.md) - the reported catalog fanout regressions are attributed and genuine extra costs are fixed with paired measurements - [Probing beside Cubic](/quest/m0/nightly/competing-cubic.md) - the `competing_cubic` probe test stays within its deadline budget -- [Interop passes on main](/quest/m0/nightly/interop.md) - re-measure the stale wire-compat pin, fix the draft-14 harness test, and route each red cell to its owner - [Gate stops hanging](/quest/m0/nightly/gate-hang.md) - `just gate` finishes instead of hanging to the timeout, and a hang names its module - [moq-quic rustdoc](/quest/m0/nightly/quic-rustdoc.md) - `--all-features` rustdoc for moq-quic builds and a lane checks it diff --git a/quest/m0/nightly/interop.md b/quest/m0/nightly/interop.md deleted file mode 100644 index b56551707d7..00000000000 --- a/quest/m0/nightly/interop.md +++ /dev/null @@ -1,41 +0,0 @@ -# [M] Interop passes on main - -## Goal - -The scheduled Interop workflow (`.github/workflows/interop.yml`) passes on -`main` again, so a red run means a new break. It has failed every night -since 2026-10-22; the last green run was 2026-10-22 11:47. - -## Plan - -Triage of run -[40517499786](https://github.com/moq-dev/moq/actions/runs/40517499786) -(2026-10-27) and the two before it. Four steps fail: - -- **Released wire compatibility**: `stale planned break: moq-lite-07-wip, - @moq/net changed from 0.4.5 to 0.4.6`. `@moq/net` 0.4.6 published on - 2026-10-23 and nobody re-measured. Run `just test wire-compat --remeasure`, - review the rewritten `test/interop/compat/planned-breaks.json`, and refresh - its reason. A patch release between cuts is not covered by - [the cut](/quest/m2/release/cut.md)'s re-measure step, so also find where - CONTRIBUTING.md's release procedure should say to re-measure after any - publish of a pinned package, and fix that inline. -- **Harness**: `test/interop/clients/js/publish-ui.browser.ts:390` still - sets `moq-transport-14` on a ``, which - [moq-transport 18 and 22 only](/quest/m1/ietf-drop/README.md) removed, so - `checkVersions` refuses it. Update the expected versions. -- **`interop --all`**: the failing cell changes each night (`cpp -> - js-native-node` tone "timestamp gap", `go -> c`, `rust -> js` past its - 30 s budget, `goaway rust <-> js-native-*`). Route each to its owner: - FFI-publisher tone gaps belong to - [FFI publisher audio gaps](/quest/m0/test-flakes-2/ffi-audio-gap.md). - Fix the rest here at the cause, or plan a quest per cause through - `/quest-plan` if one is large. -- **Tail**: `tail js-native-* -> rust` belongs to - [JavaScript FETCH](/quest/m1/js-fetch.md); confirm it still fails only - for that reason. - -Done when a scheduled Interop run passes, or every remaining red cell has -an owning quest in m0. - -Public API: none. Wire: none. From bc57263932d554abc4c291c7512f35f260061e0a Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Thu, 29 Oct 2026 08:24:09 -0700 Subject: [PATCH 4/4] quest: point jitter-admission at the landed container max buffer Co-Authored-By: Claude Opus 5.5 --- quest/m1/cut-audit/jitter-admission.md | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/quest/m1/cut-audit/jitter-admission.md b/quest/m1/cut-audit/jitter-admission.md index caebd2f784d..fcf76a0ed97 100644 --- a/quest/m1/cut-audit/jitter-admission.md +++ b/quest/m1/cut-audit/jitter-admission.md @@ -32,9 +32,10 @@ a track whose lead keeps growing as a source fault. Bound bytes too (decided 2026-10-27, maintainer, from a second audit of main against release). Timestamp distance does not bound memory: frames -that share one timestamp all fall inside the window, and -[container backpressure](/quest/m1/container-backpressure.md)'s budget stops -at frames the jitter buffer has not pulled yet. So a track is also full while +that share one timestamp all fall inside the window, and the container +consumer's max buffer +([moq-dev/moq#7056](https://github.com/moq-dev/moq/pull/7056)) only charges +frames the jitter buffer has not pulled yet. So a track is also full while its queued bytes exceed a per-track byte budget. Like the time bound, it holds the track back in its source as backpressure, never refuses a frame. Test a track that packs many frames into one timestamp beside an honest one.