mikkeli
6a7de03773
obd-poller asks 0x7DF for the nine obd2 signals, only while the bus is alive, and publishes envelopes stamped with the reply's kernel time. The recorder subscribes to yukart/vehicle/* and writes vehicle.ndjson beside the frames. Tested against a fake ECU on vcan0. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DDwfd6Ki8UfG4hK6W5YpJz
1.6 KiB
1.6 KiB
type
| type |
|---|
| guidelines |
Core services guidelines
1. One CAN writer: obd-poller
obd-polleris the only thing that transmits on the car's bus (project-yukart decision 3): OBD-II requests, one at a time, only while the bus is visibly alive. Anything else that transmits is a bug; the recorder only reads.can_pcanis no longer listen-only (2026-10-12), so the adapter acknowledges frames like any scan tool. Going back is one line in orin-nano-env-setup's/etc/default/can-can_pcan.
2. Capture never waits
- A frame missed is gone. Producers hand frames to the writer without blocking; what doesn't fit is counted, recorded as an event, and never silently lost.
- Capture reads SocketCAN and gpsd directly. It does not depend on the broker, a display, or any other service being up.
3. Every failure leaves complete files
- Append, flush and
fdatasyncabout once a second. Power loss costs at most the last second. - Replace files by write-temp, fsync, rename. Never rewrite in place.
- Anything found half-done at startup is repaired or marked, not deleted.
4. Raw stays raw
- Recordings are
candump -Ltext plus the gpsd JSON as received, socan-utils,canplayerand anything already written against the first captures keep working. - Decoding happens downstream of the recording, never instead of it.
5. Time is evidence, not truth
- The car unit's clock can boot stale and be stepped mid-drive. Record what is needed to correct it later (sync state, monotonic time, observed steps, GNSS time beside local receive time) rather than trusting it.