diff --git a/quest/m1/README.md b/quest/m1/README.md index fab044d765c..66ee3556b89 100644 --- a/quest/m1/README.md +++ b/quest/m1/README.md @@ -62,6 +62,7 @@ were done or not worth their cost. - [Dogfood hosted worklets](/quest/m1/dogfood-assets.md) - moq.dev and the moq.pro dashboard host the worklets and call `assets()` after the release - [Media audio-tone check](/quest/m1/media-audio-tone.md) - the media lane's audio-tone check passes under load, fixed at its cause - [More tests under load](/quest/m1/test-flakes-2/README.md) - the second round of load-only failures, one quest per flake, fixed at the cause +- [Test TypeScript check](/quest/m1/test-ts-check.md) - `just check` type-checks the TypeScript harnesses under test/ - [IETF request headers](/quest/m1/ietf-dispatch-headers.md) - each bidi request reads its header in its own task, so a slow one never blocks the next - [END_OF_TRACK placement](/quest/m1/ietf-end-of-track-placement.md) - END_OF_TRACK rides the upstream's Location, and Rust's header stops claiming END_OF_GROUP - [moq-uring tests under load](/quest/m1/uring-tests-under-load.md) - uring tests pass while parallel checks share locked memory @@ -176,3 +177,4 @@ were done or not worth their cost. - [Datagram replay bound](/quest/m1/datagram-replay-bound.md) - a new Rust datagram subscriber starts within its max delay of the newest datagram, not at a minutes-old buffer - [Catalog colour](/quest/m1/color-catalog.md) - the catalog describes a rendition's colour and HDR properties, which the WebGPU HDR renderer reads - [WebGPU HDR](/quest/m1/webgpu-hdr.md) - HDR renditions play as HDR where the browser and display can show it, and tone-map to SDR elsewhere +- [Request ID order](/quest/m1/request-id-order.md) - drafts 14 to 16 refuse a reused or lower Request ID in both languages diff --git a/quest/m1/ietf-legacy-updates.md b/quest/m1/ietf-legacy-updates.md index 2529e0a5857..32f505589a8 100644 --- a/quest/m1/ietf-legacy-updates.md +++ b/quest/m1/ietf-legacy-updates.md @@ -14,10 +14,9 @@ it with REQUEST_OK or REQUEST_ERROR, and a refused d14 update ends the subscription with INTERNAL_ERROR. Remaining (maintainer, 2026-10-07, from #4961 and #5011): -- JS: #5011 routes updates to their target, but the d14-16 publisher never - reads the subscribe stream again, so an unread update keeps the stream from - reporting closed and a later UNSUBSCRIBE is lost. Read framed messages and - apply updates as Rust now does. Rebase #5011 onto that or fold it in. +- JS: #5011 routes updates to their target, and the d14-16 publisher reads + and discards them so a later UNSUBSCRIBE still closes the stream. Apply them + as Rust now does. - Drafts 15 and 16: an update aimed at a namespace or fetch request is skipped with no REQUEST_OK or REQUEST_ERROR; answer it. - Draft 14: every field is mandatory, so a narrowed start or end group looks diff --git a/quest/m1/request-id-order.md b/quest/m1/request-id-order.md new file mode 100644 index 00000000000..086701934dd --- /dev/null +++ b/quest/m1/request-id-order.md @@ -0,0 +1,25 @@ +# [S] A reused moq-transport Request ID is refused + +## Goal + +On drafts 14 to 16, `@moq/net` and `moq-net` refuse an incoming request whose +Request ID is not above every ID the peer already used, instead of accepting +any ID below the window's bound. A peer can no longer reuse an ID that named +an earlier request. + +## Plan + +Found while landing [#5011](https://github.com/moq-dev/moq/pull/5011); it +predates it and covers every request type. `js/net/src/ietf/adapter.ts` +deliberately tolerates reuse today ("A peer reusing IDs below the bound still +cannot hold more than the window open"), bounding only how many are open. + +- First confirm what drafts 14 to 16 require of an out-of-order or repeated + Request ID and which session error they name; follow that. +- Keep a high-water mark of the IDs the peer allocates and refuse anything at + or below it, in both languages, beside the existing window checks. +- Tests: a repeated ID and a lower ID are each refused with the drafts' error; + in-order IDs and the MAX_REQUEST_ID grant are unchanged. + +Public API: none. Wire: no format change, but a stricter receiver closes +sessions with peers that reuse IDs today; report it as an interop change. diff --git a/quest/m1/test-ts-check.md b/quest/m1/test-ts-check.md new file mode 100644 index 00000000000..6d60d5178a5 --- /dev/null +++ b/quest/m1/test-ts-check.md @@ -0,0 +1,24 @@ +# [S] Checks type-check the TypeScript under test/ + +## Goal + +Every TypeScript harness under `test/` (such as `test/drain/drain.ts`) is +type-checked by `just check`, so a removed or renamed `@moq/*` API breaks the +PR that removes it instead of a nightly run. + +## Plan + +Found by [#5070](https://github.com/moq-dev/moq/pull/5070): `test/drain/drain.ts` +still called the deleted `writeString`, and only its review caught it. Today +`js/justfile`'s `check` runs each root workspace's `check` script. +`test/drain` is a root workspace but has no `check` script or tsconfig, so it +is skipped; `test/wasm` and `test/interop/clients/js-native` also lack a +`check` script. Standalone harnesses such as `test/max-age/client.ts` and +`test/interop/*.ts` sit outside any workspace tsconfig, and `sh/dispatch.sh`'s +`js` scope skips a change that touches only `test/drain/`. + +- Type-check every `test/` TypeScript file, workspace or standalone, with + `tsc --noEmit`. +- The scoped `just check` runs it when a harness or a `js/` package it imports + changes. +- Prove coverage once with a deliberate type error in each kind of harness.