Merge main into dev#1848
Merged
Merged
Conversation
Co-authored-by: moq-bot[bot] <186640430+moq-bot[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Luke Curley <kixelated@gmail.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…7.3 (#1790) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…1792) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: moq-bot[bot] <186640430+moq-bot[bot]@users.noreply.github.com>
… aarch64-linux (#1793) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
#1806) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…1809) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…st the last seen (#1812) Co-authored-by: Luke Curley <luke.curley@discordapp.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Add `closed` to the pull_request types so a merge/close lands in the same concurrency group and cancels the superseded Check build, freeing the self-hosted runner. The job is skipped on close (it exists only to trigger the cancellation). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…ix://) (#1810) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…e cap (#1816) Co-authored-by: Luke Curley <luke.curley@discordapp.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…ion) (#1815) Co-authored-by: Luke Curley <luke.curley@discordapp.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Luke Curley <kixelated@gmail.com>
… metadata track (#1822) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
#1820) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Luke Curley <kixelated@gmail.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
…be (#1842) Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Brings 31 commits from main into dev, resolving 40 conflicted files plus a few silent both-sides-added duplicates. Where main added a feature on top of code dev had since refactored, the feature was hand-ported onto dev's architecture rather than dropped or taken wholesale. Notable resolutions: - CI (check.yml, flake.nix, nix/overlay.nix, rs/justfile): took main's cargo+sccache direction (it supersedes the crane checks dev only carried via an earlier merge), while preserving dev's wasm/vaapi devShell deps. - Cargo.toml/Cargo.lock: qmux 0.2 (verified 0.2 keeps dev's high-level ws::Client/Server API), kept dev's moq-hls; also pulled boytacean 0.12.1. - moq-native: kept dev's high-level qmux integration and Option-aware with_publisher/with_subscriber API; deduped a both-sides-added PeerIdentity in tls.rs. - moq-relay: kept dev's API, added main's unauthenticated internal listener (#1810, tcp:// + unix://) and ported its smoke tests and internal.rs onto dev's Option-returning publisher/subscriber and announced()/track() API; dropped main's cluster.root tests (dev removed that field). - moq-mux: hand-ported main's multi-SPS/PPS/VPS carriage (#1812) and generic verbatim MPEG-TS / mpegts catalog section (#1815) onto dev's split-based importer architecture, keeping dev's thiserror error types. moq-cli and moq-gst consumers updated to match (moq-gst also keeps main's broadcast-aligned timestamp fix). - js/watch: dev had already ported main's buffered-playback latency feature (#1620) onto its API, so dev's expression was kept. - demo/web: hand-ported main's #1822 multi-broadcast inspector / stats dashboard onto dev's @moq/hang and @moq/watch APIs. The metadata feature was re-expressed as a catalog "meta" section (dev's publish element cannot serve a custom track), a behavioral change from #1822's separate-track design. Verified: cargo check --all-targets (minus moq-gst, which needs gstreamer system libs unavailable here), clippy -D warnings, cargo fmt, biome, and the js type-checks/tests all pass. moq-mux (279), moq-net, moq-token, moq-json, hang, moq-cli, and moq-relay lib (124) tests pass. The only failing Rust tests are environmental (IPv6 [::] binds and the relay smoke integration tests, which fail identically on unmodified dev in this sandbox). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BQ9o9paZnRLpYRgbyjUwFV
The publish element's `Broadcast` created its `Moq.Broadcast` producer internally and never exposed it, so an application couldn't serve its own tracks alongside the built-in catalog/audio/video. Expose it as a `net` signal, (re)created on each (re)connection. A statically inserted track (`net.createTrack`/`insertTrack`) is served via the producer's fast path, so it bypasses the element's `requested()` loop instead of being rejected as an unknown track. Use it in the watch demo: the per-stream metadata now rides on a separate `meta.json` track (advertised in the catalog's `metadata` list) served via `net` and a `@moq/json` Producer, instead of the catalog `meta`-section workaround. The watch inspector subscribes to it off `broadcast.output.active`. This restores the separate-track design from the demo without the backwards-compat `publishTrack` shim. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BQ9o9paZnRLpYRgbyjUwFV
…r catalog frame The watch inspector read the advertised metadata track name straight off the catalog signal, so the subscription effect tore down and recreated the `meta.json` subscription on every catalog update (e.g. a live encoder-setting tweak), briefly flickering the metadata panel. Derive the track name through a memoized computed so the subscription only re-runs when the name or the active broadcast actually changes. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BQ9o9paZnRLpYRgbyjUwFV
dev advanced 3 commits while this PR was open (#1845 PTS-exposing TS Export + PCR-paced SRT egress, #1847 moq-lite-05 wire sync, #1819 always-on hardware encoders). Re-merged dev and reconciled two conflicts: - nix/overlay.nix: kept this PR's cargo+sccache CI direction (dev's tip still carries the crane `moqChecks`; the cargo switch lives on main, #1821). - rs/moq-mux/src/container/ts/export.rs: combined dev's #1845 Frame-returning `Export::next` (PTS/keyframe-stamped) with this PR's generic `catalog::Catalog` trait (main's #1815 mpegts rename). Updated the moq-cli drain helper to read `frame.payload` and refreshed two stale `scte35::Ext` comments. Verified: moq-mux (279), moq-cli, moq-srt tests pass; workspace clippy -D warnings and fmt clean (excluding the crates that need libva/gstreamer system libs unavailable in this sandbox: moq-video/libmoq/moq-boy/moq-gst); JS type-checks and biome clean. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BQ9o9paZnRLpYRgbyjUwFV
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merges the 31 commits on
mainthat weren't yet indev, resolving 40 conflicted files plus a couple of silent both-sides-added duplicates that git didn't flag. The guiding principle throughout: wheremainadded a feature on top of codedevhas since refactored, the feature was hand-ported onto dev's architecture rather than dropped or taken wholesale.Conflict resolutions
CI (
check.yml,flake.nix,nix/overlay.nix,rs/justfile) — took main's cargo+sccache direction. dev only carried the crane checks via an earlier main merge (#1801); main has since superseded it with the cargo/sccache approach (#1821), so main's is the newer decision. Preserved dev's wasm32 target +wasm-bindgen-cli+libvadevShell deps.Cargo —
qmux0.2 (confirmed 0.2 still exposes dev's high-levelws::Client/ws::Serverbuilder API, so dev's cleaner integration compiles against it), kept dev'smoq-hls, and the lockfile picked upboytacean0.12.1.moq-native — kept dev's high-level qmux integration and its Option-aware
with_publisher/with_subscriberAPI. Deduped a both-sides-addedPeerIdentitystruct/impl intls.rs(a silent auto-merge artifact).moq-relay — kept dev's API + brought in main's unauthenticated internal listener (#1810,
tcp://+unix://). Portedinternal.rsand the internal-listener smoke tests onto dev'sOption-returningpublisher()/subscriber()andannounced()/track()API. Dropped main'scluster.rootmigration tests (dev removed that field entirely).moq-mux (the big one) — hand-ported main's two features onto dev's
split.rs-based importer architecture (#6419a702), keeping dev'sthiserrorerror types instead of reverting toanyhow:Vecwithpush_distinct/reconcile_keyframe_params;build_avcc/build_hvcctake lists.mpegtscatalog section (feat(moq-mux): generic verbatim MPEG-TS carriage (mpegts catalog section) #1815): reconciled main's newts/catalog.rs+ts/adts.rswith dev's TS layout.moq-cliandmoq-gstconsumers updated to dev'simport::Track/decodeAPI;moq-gstalso keeps main's broadcast-aligned timestamp fix (#678a250c).js/watch — dev had already incorporated main's "latency range with buffered playback" feature (#1620) and refactored it onto dev's API, so dev's expression was kept across all 8 files.
demo/web — hand-ported main's #1822 multi-broadcast inspector / stats dashboard onto dev's
@moq/hangand@moq/watchAPIs (catalog/backend output reshaping). Accepted main's deletion of the now-orphaneddiscover.ts.New public API:
@moq/publishBroadcast.netThe publish element's
Broadcastcreated itsMoq.Broadcastproducer internally and never exposed it, so an application couldn't serve its own tracks. This adds anetsignal exposing that producer (re-created per connection). A statically inserted track (net.createTrack/insertTrack) is served via the producer's fast path, bypassing the element'srequested()loop.With this, the demo's per-stream metadata now rides on a real separate
meta.jsontrack (served vianet+ a@moq/jsonProducer, advertised in the catalog'smetadatalist, read on the watch side offbroadcast.output.active) — restoring #1822's separate-track design rather than the catalog-section workaround.Re-merge of dev (catch-up)
devadvanced 3 commits while this PR was open, re-merged and reconciled:Export::nextreturningFrame+ PCR-paced SRT egress): combined dev'sFrame-returning API with this PR's genericcatalog::Catalogtrait (thempegtsrename from feat(moq-mux): generic verbatim MPEG-TS carriage (mpegts catalog section) #1815); updated the moq-cli drain helper to readframe.payload.nix/overlay.nixover dev's still-present cranemoqChecks.Notes for review
with_mode/decode_frame/Mode(removed in dev's refactor); the asserted behavior is now covered at thesplit.rslevel for both H.264 and H.265.Audio.Statshas nosampleCount).Verification
cargo clippy --workspace --all-targets -D warningsandcargo fmt --checkclean; the JS per-package type-checks andbiome checkpass.moq-gst/moq-video/libmoq/moq-boyneed gstreamer/libva system libs that aren't installed here (post-feat(moq-video)!: make hardware encoders always-on (openh264 stays the software fallback) #1819 vaapi is always-on); themoq-relaysmoke integration tests and themoq-nativeIPv6 dual-stack tests fail on[::]binds / QUIC client sockets, and fail identically on unmodifieddevin this container.(Written by Claude)