Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions quest/m0/ci/land.md
Original file line number Diff line number Diff line change
Expand Up @@ -76,6 +76,12 @@ Decided 2026-10-26 (maintainer):
`release`), `dependabot.yml`, and `smoke-toolchain.yml` that use it
reference the Environment. Only the hosted land job reads it, never the job
that runs PR code.
- Open question, added 2026-10-29 (maintainer) from
[moq-dev/moq#7166](https://github.com/moq-dev/moq/pull/7166): the `publish`
Environment admits `main`, whose ruleset lets repository role 2 bypass
review, and moq-bot joins that bypass list here. Decide with the bypass list
whether `main` keeps deploying to `publish`, or publishing narrows to
`release` and release tags. Ask the maintainer; settings changes are theirs.

Work in this repository:

Expand Down
6 changes: 6 additions & 0 deletions quest/m1/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -136,6 +136,7 @@ removed).
- [moq-c at parity](/quest/m1/c/README.md) - the hand-written `moq-c` ships at the cut with every m1 binding feature moq-ffi gains
- [Lite-05/06 timestamps pass through](/quest/m1/lite-timestamps-through.md) - a timed track received over lite-05/06 stays timed and is forwarded with its timestamps over every version
- [Exact timestamps in every binding](/quest/m1/binding-timestamps.md) - bindings carry `{value, scale}` timestamps instead of lossy microseconds
- [A broadcast stamps everything from one clock](/quest/m1/broadcast-clock.md) - catalog, MSF, and HLS record groups are stamped from the broadcast's clock, never the ambient one
- [Announcement takeover](/quest/m1/announce-takeover.md) - a new independent route explicitly continues instances served through its predecessor, with shared Rust/JS semantics and lite-07 wire
- [Subscribe ranges](/quest/m1/subscribe-ranges/README.md) - a lite-07 SUBSCRIBE asks for past and live ranges with absolute or relative ends, in either order, and replaces FETCH; merged ranges never starve a subscriber; relays fill misses by range, including over moq-transport
- [Drop set](/quest/m1/drop-set.md) - each cursor judges missing groups against its own budget, so a narrow reader drops promptly and a widening re-offers groups already reported dropped on lite-07
Expand Down Expand Up @@ -172,6 +173,11 @@ removed).
- [TS fidelity and recovery](/quest/m1/ts-reliability/README.md) - both modes have explicit profiles, automated validation, loss tests, and a dedicated guide
- [WebGPU on Safari](/quest/m1/webgpu-safari.md) - the WebGPU renderer is verified on Safari 26 for macOS and iOS
- [CI housekeeping](/quest/m1/ci-housekeeping.md) - remove the actionlint queue ignore once a release after 1.7.12 reaches the dev shell
- [crates.io trusts the workflow](/quest/m1/crates-trusted-publisher.md) - maintainer action: every published crate gets a trusted publisher for `release-rs.yml`
- [Rust releases publish through OIDC](/quest/m1/release-rs-oidc.md) - release-plz publishes with trusted publishing and `CARGO_REGISTRY_TOKEN` is revoked
- [moq-clock's tests pass under cargo test](/quest/m1/clock-test-process.md) - `paused_runtimes_own_their_anchors` stops depending on being first in its process
- [kio's log tests pass in a workspace build](/quest/m1/kio-workspace-tests.md) - the eviction tests pass under the workspace's unified features, or eviction is fixed
- [The JS stalled-consumer eviction test ages its groups](/quest/m1/js-cache-age-test.md) - the test drives the clock the cache reads, so it can fail
- [Main is no slower than release](/quest/m1/release-perf/README.md) - every benchmark where main is slower than release is fixed or accepted with a reason, checked by a final paired run before the cut
- [Interop coverage](/quest/m1/interop-coverage/README.md) - the riskiest moq-transport features moq-net exposes get cross-implementation tests over drafts 18 and 22
- [Calibrate the paired nightly benchmarks](/quest/m1/bench-ci.md) - A/A and deliberately slowed Nightly runs set the paired comparison's threshold, so its regression alerts can be trusted
Expand Down
45 changes: 45 additions & 0 deletions quest/m1/broadcast-clock.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
# [S] A broadcast stamps everything from one clock

## Goal

Every group a moq-mux broadcast writes, its media and also its hang and MSF
catalog groups and HLS records, is stamped from the broadcast's own clock,
captured once when it is created. None reads the ambient `Timestamp::now()`,
so a call made from another runtime or clock context cannot move a track's
timestamps backwards.

## Plan

Decided 2026-10-29 (maintainer), a follow-up of
[moq-dev/moq#7161](https://github.com/moq-dev/moq/pull/7161): one clock for
the whole broadcast is the fix, not stamping with `max(now, last)`, which
would hide the second clock.

Found 2026-10-29: media is stamped on the broadcast timeline
(`moq_mux::Clock`, held by the catalog producer), while the catalog's hang and
MSF groups (`rs/moq-mux/src/catalog/producer.rs`) and the HLS record track
(`rs/moq-mux/src/hls.rs`) stamp with ambient `Timestamp::now()`, which reads
whichever runtime or `Manual` clock is current and anchors on its origin. The
`Clock` itself also reads its instant from the ambient clock. In moq-ffi's
`containers_of_one_source_keep_their_relative_media_time`, the broadcast is
built under a paused runtime while some calls enter the FFI worker's
runtime, so one catalog track got groups from two anchors and was refused
with `TimestampMismatch` (moq-net checks group order per track).

- Find every stamp a broadcast writes that reads the ambient clock, and route
it through the broadcast's clock.
- The catalog's timestamps change meaning if they move onto the media
timeline. Check that against
[Address groups by timestamp](/quest/m1/time-addressing/README.md) and
record the choice.
- Fix the moq-ffi test's mixed clock contexts, or make it prove the new
guarantee.

Tests: a broadcast written from two clock contexts keeps every track's
timestamps non-decreasing.

Public API: none expected; record any change. Wire: none.

## Related

- [Address groups by timestamp](/quest/m1/time-addressing/README.md) - defines what a group's timestamp means
26 changes: 26 additions & 0 deletions quest/m1/clock-test-process.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,26 @@
# [XS] moq-clock's tests pass under cargo test

## Goal

`cargo test -p moq-clock` passes, not only under nextest.
`paused_runtimes_own_their_anchors` checks the same behavior without
depending on being the first test in its process to read the clock.

## Plan

Decided 2026-10-29 (maintainer), a follow-up of
[moq-dev/moq#7161](https://github.com/moq-dev/moq/pull/7161), whose agent hit
it. The test (`rs/moq-clock/src/runtime.rs`, from
[moq-dev/moq#5681](https://github.com/moq-dev/moq/pull/5681)) builds its two
paused runtimes before anything reads the clock, "so in a fresh process each
counts from its own origin". `cargo test` runs a crate's tests in one
process, so another test usually takes the origin first and the assertion fails,
every run. CI runs nextest, one process per test, so it
passes there and the failure only bites local `cargo test` runs.

Make the test independent of process state, for example by giving it a
fresh origin through the module's own test hooks or by running its body in a
child process, rather than marking it nextest-only. Keep what it proves: a
paused runtime counting from its own origin anchors on its own clock.

Public API: none. Wire: none.
Original file line number Diff line number Diff line change
Expand Up @@ -6,13 +6,18 @@ Condition: every crate that `release-rs.yml` publishes has a crates.io trusted
publisher for `moq-dev/moq`, workflow `release-rs.yml` (and its environment,
if the job uses one). This is a maintainer action in each crate's crates.io
settings and blocks
[Rust releases publish through OIDC](/quest/m2/release-rs-oidc.md).
[Rust releases publish through OIDC](/quest/m1/release-rs-oidc.md).

Check: for each crate that release-plz publishes, its crates.io settings page
lists that configuration. Delete this quest once every crate has one.

## Plan

Moved from m2 to m1 on 2026-10-29 (maintainer): after
[moq-dev/moq#7166](https://github.com/moq-dev/moq/pull/7166) moved publish
secrets into the `publish` Environment, `CARGO_REGISTRY_TOKEN` is the
remaining long-lived registry token where npm and PyPI already use OIDC.

Decided 2026-10-22: the human step is split from the workflow change so it
stays visible as ready work. A crate that has never been published still needs
its first publish done by hand before it can have a trusted publisher; list any
Expand Down
18 changes: 18 additions & 0 deletions quest/m1/js-cache-age-test.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
# [XS] The JS stalled-consumer eviction test ages its groups

## Goal

`js/net/src/broadcast.test.ts` "a stalled consumer does not pin evicted
groups" ages its groups past `maxAge` on the clock the cache actually reads,
so it fails if a stalled consumer does pin evicted groups.

## Plan

Decided 2026-10-29 (maintainer), a follow-up of
[moq-dev/moq#7076](https://github.com/moq-dev/moq/pull/7076). The test calls
`setSystemTime`, but the cache ages groups on the monotonic clock, so that
step does nothing. Drive a `Manual` clock past `maxAge` instead, as the
other converted tests do, and restore the ambient clock when it finishes.
Prove the test fails when the consumer is made to pin evicted groups.

Public API: none. Wire: none.
28 changes: 28 additions & 0 deletions quest/m1/kio-workspace-tests.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,28 @@
# [S] kio's log tests pass in a workspace build

## Goal

kio's log eviction tests, including `eviction_frees_whatever_readers_do`,
pass in `cargo test --workspace` as they do in `cargo test -p kio`. If the
workspace build exposes a real eviction bug, the code is fixed, not the
test.

## Plan

Decided 2026-10-29 (maintainer), a follow-up of
[moq-dev/moq#7161](https://github.com/moq-dev/moq/pull/7161), whose agent saw
two kio log tests fail only in a `--workspace` run.

Checked 2026-10-29: `cargo test -p kio --lib` passes with and without the
`_sim` feature, so the trigger is another dependency feature the workspace
unifies, such as one of `crossbeam-epoch`'s (kio frees evicted segments
through epoch reclamation). The cargo-hakari workspace-hack in
[moq-dev/moq#6985](https://github.com/moq-dev/moq/pull/6985) excludes kio, so
it will not fix this.

Find the feature that flips the result, then decide whether the test assumed
something the code doesn't promise (such as when a deferred free runs) or
whether eviction is wrong under that feature. Prove the fix by running the
tests under the unified feature set.

Public API: none. Wire: none.
7 changes: 6 additions & 1 deletion quest/m2/release-rs-oidc.md → quest/m1/release-rs-oidc.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,11 @@ secret is removed and revoked.

## Plan

Moved from m2 to m1 on 2026-10-29 (maintainer): after
[moq-dev/moq#7166](https://github.com/moq-dev/moq/pull/7166) moved publish
secrets into the `publish` Environment, `CARGO_REGISTRY_TOKEN` is the
remaining long-lived registry token where npm and PyPI already use OIDC.

Decided 2026-10-22 from the repository improvement audit. release-plz (0.3.172
in the dev shell) exchanges the GitHub OIDC token itself, so no separate
authentication Action is needed.
Expand All @@ -24,4 +29,4 @@ authentication Action is needed.

## Required

- [crates.io trusts the workflow](/quest/m2/crates-trusted-publisher.md) - every published crate has the trusted publisher configured
- [crates.io trusts the workflow](/quest/m1/crates-trusted-publisher.md) - every published crate has the trusted publisher configured
2 changes: 0 additions & 2 deletions quest/m2/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -105,8 +105,6 @@ lite-07; bitrate enforcement is not part of this release.
- [PulseAudio capture periods](/quest/m2/pulseaudio-capture-period.md) - capture on cpal's PulseAudio host arrives in short periods, not the server's 2 s default fragment, proven on the null sink's monitor
- [Bindings size profile](/quest/m2/ffi-size-profile.md) - a benchmark decides whether the moq-ffi builds ship at opt-level "s", which halves the dylib
- [The catalog publish path compiles once](/quest/m2/catalog-publish-once.md) - the catalog producer and its JSON diff codegen in `moq-mux` instead of once per importer crate, about 1 MiB of the CLI
- [crates.io trusts the workflow](/quest/m2/crates-trusted-publisher.md) - maintainer action: every published crate gets a trusted publisher for `release-rs.yml`
- [Rust releases publish through OIDC](/quest/m2/release-rs-oidc.md) - release-plz publishes with trusted publishing and `CARGO_REGISTRY_TOKEN` is revoked
- [Go module proxy](/quest/m2/go-module-proxy.md) - `moq.dev/moq-ffi` ships from a static GOPROXY we host, so its staticlibs stop growing git history by ~61 MiB per release
- [Go per-platform lib modules](/quest/m2/go-ffi-platform-modules.md) - a Go build downloads only the host's staticlib, and `go mod vendor` works
- [Go cancelled results](/quest/m2/go-cancelled-result.md) - a Go call cancelled as it completes destroys the handle it produced instead of leaving it to the finalizer
Expand Down
Loading