Plane: update alt error for every target change

[robertlong13]

This is related to #23458. This PR should make it much harder to let a bug like that one slip through.

Any of the functions that touch target_altitude should update altitude_error_cm before returning.

[IamPete1]

Looks like a sensible change. However all the altitude stuff is extremely touchy in plane, we would need to go through and test each case.

[magicrub]

@robertlong13 I'm with @IamPete1 . Looks like a sensible change but Altitude is some serious spaghetti code in Plane. Since @tridge has already responded on the linked issue I'll let him handle this.

[robertlong13]

@tridge, I've gone ahead and rebased this in case you're still interested in it. If we think it's too risky though, I'm also happy to close it. Other than autotest and flying around a bit in SITL, I'm not sure the best way to test.

[IamPete1]

I had another look at this. As far as I can tell we don't actually use altitude_error_cm for anything important. By my count its logging, two MAVLink messages and a check in RTL and VTOL assist.

I think it would be a clearer change to remove the variable completely. Then there is nothing to go out of date. The few places that need it can call calc_altitude_error_cm directly.

The key thing is that the diff will show all the users of the value which is the thing which actually matters.

[robertlong13]

I had another look at this. As far as I can tell we don't actually use altitude_error_cm for anything important. By my count its logging, two MAVLink messages and a check in RTL and VTOL assist.

I think it would be a clearer change to remove the variable completely. Then there is nothing to go out of date. The few places that need it can call calc_altitude_error_cm directly.

The key thing is that the diff will show all the users of the value which is the thing which actually matters.

This sounds like a much better idea to me, and way easier to verify correctness.