[jrgutier]

Summary

charging.session.trip_target is decoded with the wrong field layout. The schema has one
field, soc_limit = 2. The app's message for this topic (d5l, dispatched from
c97.java:581 in app 3.16.0) has three:

#	app field	what it is
1	soc	target state of charge, percent
2	time_estimate	minutes until that SoC is reached
3	status	enum, values not recoverable from the app

The app stores #1 and #2 unchanged as chargingTripTargetSoc and
chargingTripTargetMinsRemaining (qp2.java:130-131), the same names as the two gateway
fields in schemas/gateway.graphql. So today the decoder reports the minutes as the SoC
and never reads the SoC at all.

The same bytes, an 80% target with 45 minutes to go (08 50 10 2d 18 01):

	result
before	{"chargingTripTargetSoc": 45}
after	{"chargingTripTargetSoc": 80, "chargingTripTargetMinsRemaining": 45, "_chargingTripTargetStatus": 1}

The one frame we have captured from a vehicle is 10 ff ff 03: #2 = 0xFFFF and no #1. The
current comment, "0xFFFF when not set", comes from frames like this one, and the
decoder gives None for it before and after this change. The 0xFFFF belongs to the minutes,
though, not to the SoC.

Changes

• charging.proto: TripTarget gets soc = 1, time_estimate = 2 and status = 3
  (int32, enum type unknown). Regenerated charging_pb2.py/.pyi. No other module changed.
• decode_trip_target:
  • chargingTripTargetSoc keeps its key. It is set only when GraphQL Endpoints #1 is 1-100, the same check
    the app makes before it shows a trip target on any screen (np4.java:81,
    y13.java:397, j23.java:1146). Otherwise it is None, as before.
  • New chargingTripTargetMinsRemaining (minutes), with the gateway's name. It is set only
    when there is a target and the value is 1 to 0xFFFE. Otherwise it is None: Add CI to run tests #2 absent
    or 0 (the app hides the minutes when they are <= 0), or 0xFFFF.
  • _chargingTripTargetStatus is the raw value, following _unexpectedStopReason.
  • Fields section of the docstring updated.
• tests/parallax/test_charging.py: the 80%/45 min case, with a check that it
  serializes to 0850102d1801, so the field numbers are pinned; the 0xFFFF minutes case; an
  80% target with Add CI to run tests #2 absent (08 50), which gives None minutes rather than 0; no target at
  SoC 0 and 101; and the captured 10ffff03 frame.

Testing

PYTHONPATH=src pytest -q -p no:cacheprovider && ruff check . && ruff format --check . && mypy src/rivian

236 passed (232 before, plus 4 new; test_trip_target was rewritten). Ruff and mypy are
clean.

Notes

• The field numbers come from the app's schema. No captured frame has a trip target
  set yet. The only capture is the "no target" frame above. We will check one live frame with a
  target set before marking this ready.
• 0xFFFF minutes becomes None because the captured frame sends it. The app itself never
  tests for 0xFFFF. It only hides the indicator when the SoC is outside 1-100, and the
  minutes when they are <= 0, which the decoder follows by reporting None there too, so
  a consumer shows unknown instead of "0 min".
• status (Fixes error in publish #3) has no accessor in any of the 54 app versions we checked, so it stays raw.