Pin retina-simulation to the merged live-ADS-B feed sanity (c62f6e4)

[jehanazad]

Pin bump for offworldlabs/retina-simulation#15 (ground rows retire aircraft, altitude jump guard with self-consistent adoption, fix-age expiry, stale-row skip) — the fleet-side half of the fix for solve mnece44394c3 / N6389R, where adsb.retina.fm served another aircraft's cruise altitude in a low GA row and a numeric 0 for on-ground. The server-side half (known-lane publish rms gate) is #360, already on main.

docs/simulation.md reworded: "dropped 60 s after the feed last reported it" → after its last fix, plus the ground/jump/stale rules.

Verification: pin is the merge commit of #15 (git ls-tree HEAD libs/retina-simulation = c62f6e4); deployed to test from the droplet for a live-feed check.

🤖 Generated with Claude Code

[claude]

Claude finished @jehanazad's task in 1m 22s —— View job

---

Review: Pin retina-simulation to c62f6e4 (live-ADS-B feed sanity)

Tasks

• Gather context (PR body, changed files, CLAUDE.md)
• Review docs/simulation.md wording change
• Review libs/retina-simulation submodule pin bump
• Check repo-specific conventions (submodule policy, contract gen, etc.)
• Post final review

Summary

Small, self-contained change: a submodule pointer bump (libs/retina-simulation 6ffb678 → c62f6e4) plus a doc reword in docs/simulation.md:145-155 to match the new behavior. No backend/frontend code touched.

Findings

No blocking issues.

• docs/simulation.md changes read correctly and are internally consistent with the surrounding section (ground retirement, stale-row skip, altitude-jump guard with self-consistent adoption). Terminology ("coasted", "fix") matches usage elsewhere in the file.
• The submodule diff is a clean single-line pointer bump (Subproject commit 6ffb678... → c62f6e4...), which is the expected shape for this kind of change — nothing else in the repo needed to move since /v1/nodes, node_config.py, and the OpenAPI contract are untouched, so the "regenerate the contract" gate in CLAUDE.md doesn't apply here.
• No addresses/credentials introduced; consistent with the "public repo" rule.

Minor, non-blocking nit: "held off unless it persists, self-consistently, for a minute" (docs/simulation.md:152) reads a little awkwardly with the comma placement around "self-consistently" — consider "held off unless it persists self-consistently for a minute" (no commas). Purely stylistic.

Verification limitation

I was not able to independently confirm that c62f6e4 is in fact the merge commit of offworldlabs/retina-simulation#15 and reachable from that repo's main: this sandbox has no network access (submodule init / git ls-remote / gh api calls against offworldlabs/retina-simulation all require approval that isn't available in this headless run), and the submodule directory isn't checked out locally. The PR description states the pin was verified via git ls-tree HEAD libs/retina-simulation and a live deploy check to test — that's the right verification method, I just couldn't re-run it myself here. Worth a human double-check that #15 is merged (not just pushed to a branch) before merging this.