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

Migrated from github.com/Trevato/SO-ARM100

workbenchthis machine

A restart costs a blink, not a session: turn, queue and chat survive it and the page reconnects in under a second #440

shipped merged opened this has it

Landed in pull #453 · run

opened this

Owner: "is there a bug with new versions that break my connection? I really want a stable control surface". Today the app restarted with every forge release (forge #4496), about one an hour, on top of its own deploys.

The arm app should come through a restart without anyone noticing. Until the forge stops restarting it, a restart costs a fraction of a second, not a session.

Done when:

  1. The queue survives a restart. The turn holder, the queue order, each person's name, chat history and the moderation state are written to the app's data directory and restored on start. The person whose turn it was keeps it, with the remaining time intact.
  2. The page reconnects at once. On a socket close it reconnects with backoff starting at 100 ms (not seconds), shows a slim "reconnecting…" that never covers the camera, and resumes the same seat (browser id) and the same turn.
  3. The camera and the 3D model reload without a page reload.
  4. The arm is safe throughout. A lost relay link still parks the arm (that's correct), and after reconnecting the driver's next movement works with no extra click.
  5. Test: restart the app server in a test while a fake driver holds a turn; within 1 s the same driver has the turn back and a command reaches the fake Pi.
Sign in to join this conversation.

↑↓ navigate ↵ open ⌘↵ its app esc close

Keyboard shortcuts

Docs