diff --git a/quest/m0/test-flakes-2/README.md b/quest/m0/test-flakes-2/README.md index b9f9cb1d409..3c5a5b829dd 100644 --- a/quest/m0/test-flakes-2/README.md +++ b/quest/m0/test-flakes-2/README.md @@ -125,7 +125,6 @@ Public API: none. Wire: none. - [dispatch-query's terminate test survives load](/quest/m0/test-flakes-2/dispatch-query-terminate.md) - `test_terminating_the_run_stops_the_running_phase` passes reliably under host load, re-enabled once fixed - [Test PulseAudio servers stay off the desktop's D-Bus](/quest/m0/test-flakes-2/pulse-dbus.md) - concurrent media runs on one desktop never collide on `org.PulseAudio1` - [The Python audio playout-delay test reads its frames](/quest/m0/test-flakes-2/audio-playout-delay.md) - `test_audio_consumer_reports_its_playout_delay` stops timing out on its first read in the merge queue -- [The dev host's locked-memory limit fits parallel sessions](/quest/m0/test-flakes-2/uring-memlock.md) - the maintainer's host raises the user's `RLIMIT_MEMLOCK`, which Node.js rings from parallel agent sessions fill before io_uring tests start - [The FFI concurrent-ready test wakes both calls under load](/quest/m0/test-flakes-2/ffi-concurrent-ready.md) - `concurrent_ready_calls_all_wake` stops stalling its reader drain under load and is re-enabled - [The datagram-end close test survives load](/quest/m0/test-flakes-2/datagram-end-close.md) - `a session close right after a datagram track's end keeps the end` stops aborting the track under load and is re-enabled - [lite-05 copy fill checks idleness without wall time](/quest/m0/test-flakes-2/copy-fill-idle.md) - the copy-fill test proves no FETCH went out on the `Manual` clock, not a 200 ms real-time wait diff --git a/quest/m0/test-flakes-2/uring-memlock.md b/quest/m0/test-flakes-2/uring-memlock.md deleted file mode 100644 index 419676d4d2c..00000000000 --- a/quest/m0/test-flakes-2/uring-memlock.md +++ /dev/null @@ -1,54 +0,0 @@ -# [XS] The dev host's locked-memory limit fits parallel sessions - -## Goal - -moq-relay's and moq-uring's io_uring tests stop failing ring setup with -`ENOMEM` on the user's shared `RLIMIT_MEMLOCK` when many agent sessions -share the maintainer's host. Done when `ulimit -Hl` in a new agent session -on that host prints at least 65536, the 64 MiB hosted CI already uses. - -## Plan - -Decided 2026-11-06 (maintainer, /quest-complete): ✅ raise the host limit and -keep this quest as a host action. Not chosen: cap the `io-uring` test group at -one thread (about +45 s per workspace check), or first make the 10 s endpoint -tests fast; add a public ring-geometry knob so tests use small rings. - -Measured 2026-11-06 in [moq-dev/moq#8051](https://github.com/moq-dev/moq/pull/8051) -on that host (kernel 7.0, 8 MiB limit). The tests are not what fills the -budget: - -- Processes other than tests held 4.1 to 4.9 MiB. Most of it was Node.js: - libuv gives every event loop a 256-entry epoll-batching ring (28 KiB), and - `UV_USE_IO_URING` does not turn that ring off. Each agent session's - chrome-devtools-mcp alone holds nine of them (about 250 KiB), so the - baseline grows with the number of sessions. -- One `just rs test -p moq-uring -p moq-relay --features moq-relay/io-uring` - run needs 0.5 to 1 MiB of headroom. Twelve such runs started at once all - passed at that baseline. -- The failures reported in [moq-dev/moq#7313](https://github.com/moq-dev/moq/pull/7313) - (eleven moq-uring tests in one workspace run) leave the baseline above - 7 MiB at the time. No test-side cap brings that back under 8 MiB. - -Measured and set aside: - -- Capping the `io-uring` nextest group at one thread halves the headroom a - run needs, but slows `just rs test --workspace --features moq-relay/io-uring` - from 101 to 108 s to 149 to 153 s. Three endpoint tests in - `rs/moq-uring/tests/integration/endpoint.rs` each sit out moq-quic's 10 s - handshake timeout while holding a ring, which makes the serial group the - tail. -- A lock shared by every invocation counts its wait against nextest's - hang timeout. It also needs nextest's experimental wrapper scripts or a - hook in every ring test. -- Giving integration tests the smaller unit-test rings needs a public - geometry knob and stops testing production geometry. - -## Waiting - -- Maintainer: raise the host user's hard `RLIMIT_MEMLOCK` to 64 MiB, for - example `LimitMEMLOCK=64M` in a `/etc/systemd/system/user@.service.d/` - drop-in plus ` hard memlock 65536` in `/etc/security/limits.d/` for - login shells. An agent cannot do this because it needs root; - `sh/rs/nextest.sh` raises the limit with `sudo prlimit` only on hosted CI. - Check with `ulimit -Hl` in a fresh session. diff --git a/quest/m1/README.md b/quest/m1/README.md index c772a56bdbb..39d2b73e86d 100644 --- a/quest/m1/README.md +++ b/quest/m1/README.md @@ -57,6 +57,7 @@ moves what remains of them once their claims merge. - [The over-limit PUBLISH refusal invites a retry](/quest/m1/retry-interval-overload.md) - `EXCESSIVE_LOAD` sends the may-pass retry interval, not 0 - [A relay forwards an upstream refusal's retry interval](/quest/m1/retry-interval-forward.md) - a forwarded SUBSCRIBE refusal keeps the upstream's retry interval - [Tests run on a mocked clock](/quest/m1/mock-clock-tests/README.md) - about 300 tests stop proving nothing happened by waiting, sleeping to settle, or asserting on elapsed time +- [moq-mux reads H.265 SPS without scuffle](/quest/m1/h265-sps-reader.md) - moq-mux's own SPS reader replaces scuffle-h265 and scuffle-expgolomb, and the fuzz regressions still pass ## Related diff --git a/quest/m1/h265-sps-reader.md b/quest/m1/h265-sps-reader.md new file mode 100644 index 00000000000..83113300d2f --- /dev/null +++ b/quest/m1/h265-sps-reader.md @@ -0,0 +1,55 @@ +# [S] moq-mux reads H.265 SPS without scuffle + +## Goal + +moq-mux's `codec::h265` module parses the whole SPS with its own bounded +reader and returns the cropped size, VUI colour signal, and aspect ratio that +moq-mux and moq-video read today. scuffle-h265 and scuffle-expgolomb leave +the dependency tree. The committed `fmp4`, `annexb`, and `ts` fuzz +regressions and moq-video's SPS regression test still pass. + +## Plan + +Decided 2026-11-07 (maintainer, /quest-unblock): ✅ drop scuffle-h265 and +make the parser a module in moq-mux. ScuffleCloud/scuffle looks unmaintained +(last commit to main 2025-10-31; scuffle-h265 0.2.2 and scuffle-expgolomb +0.1.5 are from mid-2025; outside PRs such as +[ScuffleCloud/scuffle#652](https://github.com/ScuffleCloud/scuffle/pull/652) +sit unreviewed), and +[ScuffleCloud/scuffle#653](https://github.com/ScuffleCloud/scuffle/issues/653) +and [#654](https://github.com/ScuffleCloud/scuffle/issues/654) have no reply. +Not chosen: fork scuffle under moq-dev and publish patched crates; keep +waiting on upstream releases. Milestone: ✅ m1 (not chosen: m2, m4 as before). + +moq-mux already walks every SPS field itself before handing it to scuffle +(`sps::check` and `walk` in `rs/moq-mux/src/codec/h265/sps.rs`, added by +[moq-dev/moq#7685](https://github.com/moq-dev/moq/pull/7685)), so turn that +walk into the parser: + +- Read the fields the callers use instead of only bounding them, in the + spec's layout (ITU-T H.265 7.3.2.2.1). Drop the places where the walk copies + scuffle 0.2.2's non-spec layout, such as `sub_layer_level_idc` read twice + and `iv_di_mc_enabled_flag` read once. +- Keep the head fallback for SPSs whose later fields break a semantic + constraint real encoders violate (`matrix_coeffs` 0 on 4:2:0), and keep + refusing truncated or out-of-range values. +- Replace the scuffle `NALUnitType`, `Profile`, `AspectRatioInfo`, and + `VideoSignalType` uses in moq-mux (`h265/`, `container/ts/video_timing.rs`, + `container/mkv/hvcc.rs`) and moq-video (`decode/vui.rs`) with moq-mux's own + types, crate-private where moq-video doesn't need them. +- Keep every regression input, including the `scuffle-h265-*` and + `scuffle-expgolomb` ones moq-video's tests include. + +scuffle-av1 stays, though it is the same upstream: its one known bug, a +UVLC leading-zero count that overflows a `u8`, is guarded at moq-mux's own +boundary (`validate_timing` in `rs/moq-mux/src/codec/av1/import.rs`, from +[moq-dev/moq#8042](https://github.com/moq-dev/moq/pull/8042)). Keep that +guard while scuffle-av1 is a dependency. + +Public API: none; moq-mux exports no scuffle type. +Wire: none. + +## Related + +- [Media fuzz findings](/quest/m1/fuzz-findings/README.md) - found the scuffle overflows +- [Re-run the SPS fuzzer on h264-reader](/quest/m4/sps-fuzz-h264-reader.md) - the H.264 half diff --git a/quest/m3/video-quality/corpus-bucket.md b/quest/m3/video-quality/corpus-bucket.md index 5c283e6e2de..b3376af06ef 100644 --- a/quest/m3/video-quality/corpus-bucket.md +++ b/quest/m3/video-quality/corpus-bucket.md @@ -9,16 +9,31 @@ public access exposes a whole bucket and a prefix cannot be private behind it. A read-only token for the private bucket is a secret the nightly video-quality jobs can read. -This waits on a Cloudflare account action. Check it by fetching one public -object anonymously, one private object with the secret from a nightly -`workflow_dispatch` run, and the same private object anonymously, which must -fail. Delete this quest once all three hold. - ## Plan +Decided 2026-11-07 (maintainer, /quest-unblock): ✅ a checked-in wrangler +setup creates the buckets. Not chosen: running the wrangler commands once by +hand with nothing checked in; leaving it all to the maintainer. + +- Add a `just create` recipe under `infra/`, the way `infra/nix/justfile` + creates the nix.moq.dev buckets: both buckets, with public access and a + moq.dev custom domain (such as `corpus.moq.dev`) on the open one only. + Run it with the desktop's existing `wrangler login`; wrangler there reaches + the Cloudflare account. - Follow the existing R2 secret naming (`MBX_R2_*`, `R2_*`), with a separate token that is read-only and scoped to the private bucket. The corpus must never share write credentials with the caches. - Upload is a maintainer action. The manifest the harness owns records each object's sha256, so a replaced object fails the run instead of silently changing the baseline. + +## Waiting + +- Maintainer: once the setup has created the private bucket, mint its + read-only token in the dashboard (R2, Manage API tokens, Object Read only, + scoped to that bucket) and store it with `gh secret set` under the names + the setup documents. Wrangler cannot create R2 API tokens, and an agent + never handles the secret. Check by fetching one public object anonymously, + one private object with the secret from a nightly `workflow_dispatch` run, + and the same private object anonymously, which must fail. Delete this quest + once all three hold. diff --git a/quest/m4/README.md b/quest/m4/README.md index 70c412b141b..217bbd487ae 100644 --- a/quest/m4/README.md +++ b/quest/m4/README.md @@ -71,7 +71,6 @@ deleted. - [moq-bench reports sub-millisecond latency](/quest/m4/bench-latency-resolution.md) - steady-window latency resolves finer than 1 ms - [Benchmark series stay meaningful](/quest/m4/bench-hygiene.md) - a warm `burst` variant, stale series pruned, and slow-group `bytes` kept out of series names - [Re-run the SPS fuzzer on h264-reader](/quest/m4/sps-fuzz-h264-reader.md) - the H.264 SPS overflow is rechecked on the new parser -- [Drop the H.265 SPS pre-walk once scuffle is fixed](/quest/m4/scuffle-sps-bump.md) - waiting on ScuffleCloud/scuffle#653 and #654 releases, then bump and drop the walk - [PCM encodes in one pass](/quest/m4/pcm-bulk-encode.md) - the PCM encoder converts a frame in bulk instead of one sample at a time - [moq-quic's internal dev-dependencies are path-only](/quest/m4/quic-dev-dep-path.md) - moq-congestion is a path dev-dependency - [moq-mux's AAC config reports SBR](/quest/m4/aac-sbr-config.md) - moq-audio reads backward-compatible SBR from the config diff --git a/quest/m4/scuffle-sps-bump.md b/quest/m4/scuffle-sps-bump.md deleted file mode 100644 index 25f36b133bf..00000000000 --- a/quest/m4/scuffle-sps-bump.md +++ /dev/null @@ -1,39 +0,0 @@ -# [XS] Drop the H.265 SPS pre-walk once scuffle is fixed - -## Goal - -moq-mux (and moq-video, once it shares the walk) parse H.265 SPS with -fixed scuffle-h265 and scuffle-expgolomb releases and no longer pre-walk the -SPS to refuse inputs that overflowed the old ones. The committed regression -inputs still pass. - -## Waiting - -- ScuffleCloud/scuffle: fixed releases for - [ScuffleCloud/scuffle#653](https://github.com/ScuffleCloud/scuffle/issues/653) - (scuffle-h265) and - [ScuffleCloud/scuffle#654](https://github.com/ScuffleCloud/scuffle/issues/654) - (scuffle-expgolomb) on crates.io. Check the issues - and `curl -s https://crates.io/api/v1/crates/scuffle-h265` (and - `scuffle-expgolomb`) for a version newer than 0.2.2 and 0.1.5. Posting - on those repos still needs the maintainer's approval. - -## Plan - -Decided 2026-11-03 (maintainer, follow-up planning after the 2026-11-03 /quest-afk and /quest-complete runs), milestone: - -- m0 -- m1 -- ✅ m2 - -[moq-dev/moq#7685](https://github.com/moq-dev/moq/pull/7685) added the walk -(`sps::parse_full`) because scuffle-h265 0.2.2 and scuffle-expgolomb 0.1.5 -overflow on hostile SPS fields. - -- Bump both crates, run the `fmp4`, `annexb`, and `ts` fuzz regressions and - moq-video's SPS regression test against the bare parser, and delete the - walk only if they pass without it. -- Keep every regression input. - -Public API: removes the shared walk if moq-video no longer needs it. -Wire: none. diff --git a/quest/m5/rind-instant-acme.md b/quest/m5/rind-instant-acme.md index 1fe3102b1d3..1b149c43559 100644 --- a/quest/m5/rind-instant-acme.md +++ b/quest/m5/rind-instant-acme.md @@ -3,13 +3,18 @@ ## Goal `rind-instant-acme` is on crates.io, built on a published instant-acme 0.9. -It holds once moq-dev/rind's -[Publish rind-instant-acme](https://github.com/moq-dev/rind/blob/main/quest/m3/instant-acme-publish.md) -quest is deleted and `cargo search rind-instant-acme` shows it. This quest -only waits; delete it once that holds. +This quest only waits; delete it once that holds. ## Plan rind's adapter waits on instant-acme's 0.9 release, which carries the pluggable `CryptoProvider`. Only [relay ACME](/quest/m5/relay-acme.md) needs it (maintainer, 2026-10-31). + +## Waiting + +- moq-dev/rind: its + [Publish rind-instant-acme](https://github.com/moq-dev/rind/blob/main/quest/m3/instant-acme-publish.md) + quest is deleted and `cargo search rind-instant-acme` shows the crate. + Checked 2026-11-07: the rind quest is still open, crates.io has no + `rind-instant-acme`, and instant-acme's latest is 0.8.5.