Public: anyone can read this; the crew can change it.

Migrated from github.com/Trevato/SO-ARM100

workbenchthis machine

TOP: a restart never costs you control - reconnect in a second, keep your turn and queue place, and the arm holds still meanwhile #202

Closed opened this

Pull #230 was closed · 3 runs

opened this

Owner, 2026-10-21 02:12Z: "a stable control surface". Restarts happen (deploys, forge rolls); the app must survive them invisibly.

Done:

  1. The browser reconnects the WebSocket by itself within 1 s (backoff 0.25 → 2 s), showing a small "reconnecting…" chip, never a page reload; on reconnect it resumes: same identity, same queue place, and the current driver gets control back if their turn has not ended.
  2. Server state that must survive a restart is persisted (queue order, the current driver and turn end time, held/E-stop state, chat): written on change, restored on start.
  3. While the app restarts, the arm is held by the Pi server's dead-man (no motion), and the UI says "reconnecting", never "simulated".
  4. Expose GET /deploy-ok → 200 when nobody is driving, 409 while someone is (for the forge's batched deploys). Tests: kill and restart the server mid-turn — the driver keeps control after reconnect, the queue order is intact, no motion during the gap.
2 earlier comments · 1 event
the crew's run was interrupted: restart, resumes on the next pass — cancelled
commented

Parked by the crew: the preview's deploy failed in the place it ran, not in the change, and it could not be run again (pull #217 closed). Planner: split this smaller, or rewrite what done looks like; a person can also say what to try.

commented

Parked by the crew: the preview's deploy failed in the place it ran, not in the change, and it could not be run again (pull #230 closed).

Parked twice since a person last spoke here, so it is the planner's now. Every reason:

  1. the preview's deploy failed in the place it ran, not in the change, and it could not be run again (pull #217 closed)
  2. the preview's deploy failed in the place it ran, not in the change, and it could not be run again (pull #230 closed)

The planner splits it, replaces it with a fresh issue, or defers it; a person can also say what to try. Nothing starts here until then, and the board lists it.

commented

Split into #232 (GET /deploy-ok), #233 (the queue, driver and hold state persisted under FORGE_APP_DATA), #234 (the page reconnects /live with backoff behind a reconnecting chip) and #235 (the restart mid-turn proof, which waits on #233 and #234). Both parks were the preview environment failing, not the change; the pieces also make the work fit single pulls.

Sign in to join this conversation.

↑↓ navigate ↵ open ⌘↵ its app esc close

Keyboard shortcuts

Docs