From f26bfb3ea129f5afb29aedc16b11b5db19474c3a Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 24 Oct 2026 07:39:25 -0700 Subject: [PATCH 1/3] chore(quest): claim quest/m0/ffi-shutdown-backport.md Quest-Claim: 562ae8ea0e358352 From 4d4ce0d9dae1e780cd3dafe2c2683b92e87db96f Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 24 Oct 2026 08:36:03 -0700 Subject: [PATCH 2/3] quest: finish the release shutdown backport Co-Authored-By: Claude Opus 5.5 --- quest/m0/README.md | 1 - quest/m0/ffi-shutdown-backport.md | 51 ------------------------------ quest/m0/hls-admission-backport.md | 2 +- 3 files changed, 1 insertion(+), 53 deletions(-) delete mode 100644 quest/m0/ffi-shutdown-backport.md diff --git a/quest/m0/README.md b/quest/m0/README.md index f704c9a3310..c1d7fbd2222 100644 --- a/quest/m0/README.md +++ b/quest/m0/README.md @@ -43,7 +43,6 @@ Each lands its fixes in its own PR and plans only what needs a decision. - [Broadcast epochs](/quest/m0/broadcast-epoch/README.md) - every first-party publisher that can restart mints a fresh route epoch, the newest wins a path, and only routes with the same epoch resume a subscription - [64-bit Hop IDs](/quest/m0/hop-id-64.md) - a moq-transport peer with any non-zero 64-bit Hop ID keeps its session, and Hop IDs are `u64` in Rust and JS - [MALFORMED_AUTH_TOKEN session code](/quest/m0/malformed-auth-token-code.md) - a 0x16 close reports a malformed token instead of `unknown code=22`, and moq-net sends it for an unparseable token -- [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 - [H.265 aspect backport](/quest/m0/h265-aspect-backport.md) - `release` publishes the #5574 fix, so a square-pixel H.265 stream plays at its picture's aspect - [HLS admission backport](/quest/m0/hls-admission-backport.md) - `release` refuses the hostile HLS playlists `main` does, without breaking `import::Config::client` - [Adopt quest resolve](/quest/m0/quest-resolve.md) - local merges resolve concurrent quest link-list edits once kixelated/quest ships its merge driver diff --git a/quest/m0/ffi-shutdown-backport.md b/quest/m0/ffi-shutdown-backport.md deleted file mode 100644 index 1d486ed7461..00000000000 --- a/quest/m0/ffi-shutdown-backport.md +++ /dev/null @@ -1,51 +0,0 @@ -# [S] Backport draining shutdown and the qmux stream limit to release - -## Goal - -The `release` line publishes a non-breaking moq-ffi 0.4.x (wheel, Swift, and -Kotlin) and moq-tokio whose sessions can drain before closing and whose qmux -sessions accept the QUIC stream limit: - -- An async `MoqSession.drain()` drains finished tracks within one second - and returns any delivery error, as the earlier async shutdown implementation did - (7bbac3272, with 03fb0d1c1, f9de31167, and #5558's explicit Lite 07 test). - Explicit `moq-lite-07-wip` waits for the peer to receive the tail; older - Lite versions, including the default lite-06, and IETF wait only for - transport ACKs. The sync `shutdown()` stays `cancel(0)`. -- qmux sessions get the QUIC stream limit (#5609) instead of the 100-stream - default. - -Check: release-plz publishes the release commits, and the newest moq-ffi on -PyPI, the Swift package, and Maven Central contains both. Swift and Kotlin -publish from their own tag workflows, so check each. The newest hand-written -wrappers (`moq-rs`, the Swift and Kotlin libraries, and the other wrappers -that gain `drain()`) expose it too. - -## Plan - -Decided 2026-10-20 (maintainer): additive `drain()` on `release` rather than -backporting async `shutdown()` as-is. Why: making `shutdown()` async breaks -the 0.4.x Python, Swift, and Kotlin bindings. `main` requests the drain with synchronous `shutdown()` and reports its result -through `closed()`; it gains no separate `drain()`. Callers replace the release -line's awaited `drain()` with `shutdown()` followed by awaiting `closed()` at the -next breaking cut. - -Decided 2026-10-20 (maintainer): keep #5609 in the backport, on a qmux -0.5.3 published from web-transport's `qmux-v0.5.x`. Why: `release` locks -qmux 0.5.2, which lacks the builder session config #5609 uses, and -moq-dev/web-transport#427 already carries that config for 0.5. - -Follow the backport workflow. #5609 moves `release` to qmux 0.5.3 and takes -its builder session config instead of 0.6.3's. The drain backport -adapts 7bbac3272's session close to the `release` line's moq-net and keeps -its tests, renamed to `drain()`. Follow the Cross-Package Sync row for -`rs/moq-ffi`: every binding and its docs gain `drain()`, or the PR says why -one has none. Wrapper workflows skip versions already published, so bump -each changed wrapper's human-owned version (CONTRIBUTING's Versions). Leave -`go/wrapper/VERSION` alone: CI derives Go's patch, and `drain()` is -additive. Run each binding's tests and `just test interop --all`. The -qmux credit fix is out of scope: main has it from the in-tree qmux -([moq-dev/moq#6231](https://github.com/moq-dev/moq/pull/6231)). - -Public API: additive (`MoqSession.drain()` and its binding wrappers). Wire: -none. diff --git a/quest/m0/hls-admission-backport.md b/quest/m0/hls-admission-backport.md index 8db37c735be..49f99a6014a 100644 --- a/quest/m0/hls-admission-backport.md +++ b/quest/m0/hls-admission-backport.md @@ -16,7 +16,7 @@ its own redirect handling; the checks apply to the client moq-hls builds. Decided 2026-10-23 (maintainer): backport once the checks have been validated on `main`, following the shape of -[the shutdown backport](/quest/m0/ffi-shutdown-backport.md): cut from +[the shutdown backport](https://github.com/moq-dev/moq/pull/6642): cut from `release`, no breaking change, a patch release. Public API: none on `release`. Wire: none. From 8b9d748f4093c064e2a08d7af68723b4026915fe Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Fri, 30 Oct 2026 21:22:49 -0700 Subject: [PATCH 3/3] quest: rewrite the shutdown backport to its remaining publishes The drain() backport and moq-ffi, Swift, and Kotlin have shipped. What remains is publishing Python moq-rs 0.5.2 and Dart moq 0.1.1 from release, which the maintainer runs in the combined patch cut. Co-Authored-By: Claude Opus 5.5 --- quest/m0/README.md | 1 + quest/m0/ffi-shutdown-backport.md | 54 +++++++++++++++++++++++++++++++ 2 files changed, 55 insertions(+) create mode 100644 quest/m0/ffi-shutdown-backport.md diff --git a/quest/m0/README.md b/quest/m0/README.md index f0b0fdc300e..201fb724c7a 100644 --- a/quest/m0/README.md +++ b/quest/m0/README.md @@ -23,6 +23,7 @@ edge, which embeds moq-relay; it moved there on 2026-09-28. - [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 +- [Publish the release line's drain() wrappers](/quest/m0/ffi-shutdown-backport.md) - Python `moq-rs` 0.5.2 and Dart `moq` 0.1.1 reach PyPI and pub.dev from `release` - [Backport audio groupDuration](/quest/m0/backport-audio-group-duration.md) - `@moq/publish` 0.5.4 on `release` can pack several Opus frames per group, making Cloudflare's relay drop audio less often - [Backport the SETUP-path and iroh-key fixes](/quest/m0/backport-security-fixes.md) - `release` refuses a conflicting SETUP Path and keeps persisted iroh keys private, in one patch release - [CI cost and latency](/quest/m0/ci/README.md) - cloud CI runs only in a batched merge gate as one `just ci` pass per batch, expensive checks run nightly and gate `release`, caches move to R2, and runners are chosen last diff --git a/quest/m0/ffi-shutdown-backport.md b/quest/m0/ffi-shutdown-backport.md new file mode 100644 index 00000000000..26047141e2a --- /dev/null +++ b/quest/m0/ffi-shutdown-backport.md @@ -0,0 +1,54 @@ +# [XS] Publish the release line's drain() wrappers + +## Goal + +The `release` line's hand-written wrappers that gained `drain()` in +[moq-dev/moq#6642](https://github.com/moq-dev/moq/pull/6642) are on their +registries: + +- Python [`moq-rs`](https://pypi.org/project/moq-rs/) 0.5.2 on PyPI. +- Dart [`moq`](https://pub.dev/packages/moq) 0.1.1 on pub.dev. + +Check: each registry lists the version. Both versions are already set on +`release` (`py/moq-rs/pyproject.toml`, `dart/moq/pubspec.yaml`), so this needs +a publish, not code. Each publish is outward-facing, so the maintainer runs it; +agents do not push tags, dispatch workflows, or bump versions. + +## Plan + +Already published (checked 2026-10-31): moq-ffi 0.4.19 carries `drain()` and +qmux 0.5.3 on PyPI, Maven Central, and pub.dev; moq-ffi 0.4.20 followed. +Swift `Moq` 0.5.2 is tagged, and Kotlin `dev.moq:moq` 0.5.3 is on Maven Central +from `gh workflow run release-kt-lib.yml --ref release` +([run](https://github.com/moq-dev/moq/actions/runs/40904381781), after +[moq-dev/moq#7026](https://github.com/moq-dev/moq/pull/7026)'s fix). + +Decided 2026-10-30 (maintainer): `moq-rs` 0.5.2 and Dart `moq` 0.1.1 ship in +the next combined patch cut on `release`, which the maintainer cuts. + +Python: `workflow_run` runs `main`'s `release-py.yml` against the `release` +checkout, which lacked `.github/actions/nix`. The maintainer picked the +backport of that action to `release`, which landed in +[moq-dev/moq#7421](https://github.com/moq-dev/moq/pull/7421). The +`moq-ffi-v0.4.20` tag's chained run +([failed](https://github.com/moq-dev/moq/actions/runs/41038297141)) came before +that merge, and a re-run checks out the same tag commit, so it fails again. +`release-py.yml` has no `workflow_dispatch`. What publishes 0.5.2 now: + +- The next `moq-ffi-v*` tag at a `release` commit with the nix action, which + chains Release Py FFI into Release Py. +- A `release` push touching `py/moq-rs/**`, `.github/workflows/release-py.yml`, + or `sh/gh/release.sh`: the push gate publishes, since `release`'s moq-ffi + 0.4.20 is tagged and on PyPI. + +Dart: pub.dev publishes `moq` only from a `moq-dart-v*` tag, and only +`moq-dart-v0.1.0` exists. Pushing `moq-dart-v0.1.1` at `release`'s head +publishes 0.1.1 and exercises the unproven path in +[A Dart tag publishes unattended](/quest/m2/release/dart-publish.md). Decided +2026-10-30 (maintainer): push it during the combined cut, not before. + +Public API: none. Wire: none. + +## Related + +- [A Dart tag publishes unattended](/quest/m2/release/dart-publish.md) - proves the `moq-dart-v*` tag path this publish uses