test(capture): drive native capture through clock edge cases in CI

quest/m1/README.md

@@ -35,7 +35,6 @@ transport, benchmark tooling); worktrees isolate commits, not semantics.
 - [Broadcast close](/quest/m1/broadcast-close/README.md) - `close()` is the one way to end a broadcast in every language, a permanent retraction that leaves in-flight tracks alone
 - [Relay peer set](/quest/m1/relay-peer-set.md) - a wire consumer tells a client hop from a peer hop, and every mesh credential can mark a peer
 - [CLI import clock](/quest/m1/cli-import-clock.md) - fMP4, TS, and FLV imports publish on the shared broadcast clock across restarts
-- [Native clock fixtures](/quest/m1/native-clock-fixtures.md) - CI drives native capture through clock edge cases and asserts the published timestamps
 - [CLI inspection](/quest/m1/cli-inspect/README.md) - `moq ls` lists what is live and `moq fetch` reads a group over MoQ, and a guide shows how to inspect a relay
 - [JS caught up](/quest/m1/js-announce-caught-up.md) - @moq/net's announce consumer says when the initial set has landed, like Rust
 - [Bindings caught up](/quest/m1/announce-live-bindings.md) - moq-ffi, libmoq, and every wrapper yield the same flat announce event, `Live` included

quest/m1/cli-import-clock.md

@@ -30,5 +30,4 @@ archive playback. Update the import docs.
 
 ## Related
 
-- [Native clock fixtures](/quest/m1/native-clock-fixtures.md) - the same scenarios through native capture
 - [GStreamer clock](/quest/m1/3021-moq-gst-anchor-generated-media-timelines-to-wall-clock.md) - separate source adapter

quest/m2/README.md

@@ -82,5 +82,7 @@ upstream release waits in [m4](/quest/m4/README.md).
 - [Windows capture parity](/quest/m2/capture-windows.md) - system audio and screen cursor capture with a settled app-capture policy
 - [Linux capture parity](/quest/m2/capture-linux.md) - Wayland window/system-audio capture with explicit display-selection and app-capture limits
 - [Plan capture ergonomics](/quest/m2/capture-ergonomics.md) - scope independent crop and audio mixing quests
+- [Capture clock source](/quest/m2/capture-clock-source.md) - capture publishers stamp on the catalog's clock, with no separate clock to pass, on dev
+- [Audio capture time](/quest/m2/audio-capture-time.md) - native audio stamps a buffer's capture instant, not when the driver reads it
 - [X11 capture transport](/quest/m2/x11-capture-shm.md) - move X11 capture to shared memory and RandR events instead of a per-frame socket copy
 - [Capture frame buffers](/quest/m2/capture-frame-buffers.md) - stop rebuilding a full-frame buffer every tick in the X11 and Windows backends

quest/m2/audio-capture-time.md

@@ -0,0 +1,20 @@
+# [S] Native audio stamps acquisition, not arrival
+
+## Goal
+
+Native audio capture stamps a buffer at the instant its first sample was
+captured, like native video stamps a frame's acquisition. Today the driver
+stamps the broadcast clock when it reads the buffer, so audio lands at least
+one device buffer (plus callback and queue latency) later than video captured
+at the same instant.
+
+## Plan
+
+cpal reports `InputCallbackInfo::timestamp().capture` per callback; carry it
+through `capture::Samples` and map it onto the broadcast clock once per open,
+the way video maps its private capture timeline. Fall back to arrival where a
+host reports nothing usable, and keep the reset-on-gap behavior. Measure the
+skew before and after on at least one real device, and extend the moq-audio
+clock fixtures so a synthetic buffer's capture instant is what publishes.
+
+Public API: none. Wire: none.

quest/m2/capture-clock-source.md

@@ -0,0 +1,20 @@
+# [S] Capture publishers read the catalog's clock
+
+## Goal
+
+`moq_video::encode::publish_capture` and `moq_audio::encode::Publication` stamp
+on the clock their catalog advertises, with no separate clock to pass. Today
+each takes its own `moq_mux::Clock`, and `PublicationOptions::default()` builds
+a fresh one, so a caller relying on the default publishes audio against a
+mapping the catalog never advertised. `moq import capture` passes
+`catalog.clock()` to both, which is the only correct value.
+
+## Plan
+
+Drop the `clock` parameter from video `publish_capture` and the `clock` field
+from `PublicationOptions`, reading `catalog.clock()` instead. Update moq-cli and
+any binding that forwards a clock. The clock fixtures in both crates already
+pass the catalog's clock, so they keep grading the same path.
+
+Public API: breaking in published `moq-video` and `moq-audio`, so it targets
+`dev`. Wire: none.

quest/m2/js-discontinuity.md

@@ -21,4 +21,9 @@ and any other caller over, and keep data tracks skipping a sequence the way
 Rust's `discontinuity()` does if a JS data-track caller appears. Rust is
 untouched.
 
+Match Rust's break, too: without an explicit end, `discontinuity()` closes the
+group with no cadence-estimated duration marker. Whatever resumes can land
+sooner than one estimated frame later (a capture swap), and a marker past it
+reads as a rewind to every consumer.
+
 Public API: breaking in published `@moq/hang`, so it targets `dev`. Wire: none.

rs/justfile

@@ -495,9 +495,14 @@ check-changed $FILES $TEST="false":
 
     # moq-video's and moq-audio's device code is behind `capture`, which is off
     # by default, so the pass above never compiles it. Only nightly's
-    # `--all-features` did, and a break there lands on main.
+    # `--all-features` did, and a break there lands on main. With TEST=true the
+    # lint rides the test build, as in `check-test`.
     if [[ "$packages" == "ALL" ]] || just rs _wants-capture "$packages"; then
-    	just rs capture
+    	if [[ "$TEST" == true ]]; then
+    		CARGO_BUILD_WARNINGS=deny just rs capture-test
+    	else
+    		just rs capture
+    	fi
     fi
 
     # moq-relay's io_uring listener is behind `io-uring`, off by default, so the
@@ -628,6 +633,13 @@ macos *args:
 capture *args:
     cargo clippy --locked -p moq-video -p moq-audio --all-targets --features moq-video/capture,moq-audio/capture {{ args }} -- -D warnings
 
+# The capture publishers' tests (the clock fixtures among them) drive synthetic
+# devices, so they need no hardware and run with every PR that reaches either crate.
+
+# Test the capture feature, which the default test pass skips.
+capture-test *args:
+    just rs test -p moq-video -p moq-audio --features moq-video/capture,moq-audio/capture {{ args }}
+
 # Same idea as `windows`/`macos`, for Android: moq-video's MediaCodec encoder and
 # decoder, its `Surface::HardwareBuffer` variant, and the `frame::android` module
 # are all `#[cfg(target_os = "android")]`, so a host-target check compiles none of
@@ -886,6 +898,11 @@ test-changed $FILES:
     	*)     echo "rs: testing $(just rs _names "$packages")"; just rs test --no-tests=pass $packages ;;
     esac
 
+    # Mirrors `check-changed`: the capture tests are off the default feature set.
+    if [[ "$packages" == "ALL" ]] || just rs _wants-capture "$packages"; then
+    	just rs capture-test
+    fi
+
 # Compile and run the `/// ```` examples, which nextest skips.
 doctest *args:
     cargo test --locked --doc {{ args }}