[pullfrog]

Problem

moq-dev/moq#7691 made every JSON write take a broadcast-clock timestamp and refuses an untimed JSON track. It missed the in-tree CARLA status publisher: demo/carla/sim/moq_carla_sim/publish.py still created the status track with moq.TrackInfo(compress=True) and called SnapshotProducer.update(payload) without a timestamp. So on main, just check fails at check carla (pyright reports the missing argument), and the worker, which installs the bindings from this checkout, would fail with TimestampMismatch at startup.

Approach

• Give the status track timescale=1_000.
• Stamp every StatusPublisher.publish with moq.media.CatalogProducer(broadcast).now(). That is the broadcast clock the bridge's video tracks are shifted onto, so status and video share one timeline.
• Update the test stub's SnapshotProducer.update to take (value, timestamp) and add a CatalogProducer stub. A call that drops the timestamp now fails the tests instead of being hidden by the stub.

I searched the rest of the tree (Python, Go, Kotlin, Swift, Dart, C/C++, the interop clients, demos, and doc samples) for other JSON writers that #7691 left untimed. There are none: CARLA's status track was the only one. The JS @moq/json package has its own implementation, which #7691 did not change.

Impact

• Public API: none. This changes only the demo.
• Wire: the CARLA status track now carries a timescale and frame timestamps, as feat!: JSON writes take a broadcast-clock timestamp in every binding #7691 requires for every JSON track.

Checks

• just carla check: pyright found 0 errors, pytest passed 169 tests, and the launcher, server, systemd, and bridge GPU script tests all passed.
• ruff check and ruff format --check in demo/carla/sim: clean.

Review

• Reviewer: Codex (quest review), gpt-6.1-sol, medium effort.
• Head: aeac7c8, against origin/main at 7e4ed93.
• Round 1 reported no findings.

(Written by Claude Opus 5.5)

[kixelated]

Automated review by review (OpenAI)

Reviewed head aeac7c8.

No actionable findings. The timed status track and broadcast-clock timestamp address the constructor and update mismatch introduced by #7691, and the test stub now requires the timestamp. Using the shared broadcast clock is the right direction for alignment with media.

Verification: GitHub-only static review of the CARLA change, binding contract, and tests. I did not run the checks or exercise CARLA.