Plane: disallow mavlink disarm while flying

[tridge]

this relies on is_flying(), and we will need to watch for reports of
the heuristics failing
in simple SITL tests the is_flying() checks seem to be good for VTOL modes

[peterbarker]

Looks right

[IamPete1]

LGTM. If users really want to disarm from mavlink they can still force.

[IamPete1]

CI failures look genuine. Can no longer disarm in airmode since that forces flying. Just a matter of updating the tests to reflect the new behavior.

[tridge]

CI failures look genuine. Can no longer disarm in airmode since that forces flying. Just a matter of updating the tests to reflect the new behavior.

fixed

[davidbuzz]

seems quite reasonable.

[WickedShell]

I really think this needs an opt in (or at least have an opt out) check. I'd prefer not to have this. I've had to disarm aircraft in the past from the GCS and the flying estimate has been quite wrong before. This definitely doesn't look like a check I'd like to have on, as disarming is used to safe the aircraft in the event of a crash or anomaly.

Random log I have here shows that after crashing is flying held true for 5 seconds before falling down to be not true, and that only happened because it was disarmed.

[tridge]

@WickedShell as discussed, I'd be happy to add a FLIGHT_OPTIONS bit to disable the new behaviour, but we do need to change the default because of the way the mission planner arm/disarm button is implemented

[timtuxworth]

What about RC disarm? I accidentally disarmed my plane today about 200m up in the air. The log is quite spectacular and crazy that the plane is still in one piece. Will this PR prevent this?
This is the log: https://www.dropbox.com/s/78uk9sdjyswfw2e/log_81_2022-3-10-15-17-00.bin?dl=0