Plane Guided Alt Command: Use frame instead of bools for setting alt frame

[Ryanf55]

Purpose

When a plane receives an altitude command from mavlink SET_POSITION_TARGET_GLOBAL_INT, we were using the booleans in the Location field to set the frame. This switches to using the altitude setter with the frame which is more explicit.

We now convert the frames with mavlink_coordinate_frame_to_location_alt_frame.

Conveniently, deleting code saves 40b flash.

Tests performed

./Tools/autotest/autotest.py test.Plane.

It seems like this might be covered by SetpointGlobalPos, but that test is not run by Plane.

I also used the changealt and changealt_abs command in GUIDED mode while flying in SITL to verify we can change altitude. Those commands do not use this interface.

I do not have a great way to test this as is. Suggestions welcome!

Size Compare Durandal

Binary Name      Text [B]        Data [B]     BSS (B)      Total Flash Change [B] (%)      Flash Free After PR (B)
---------------  --------------  -----------  -----------  ----------------------------  -------------------------
arducopter-heli  0 (0.0000%)     0 (0.0000%)  0 (0.0000%)  0 (0.0000%)                                      14[28](https://github.com/ArduPilot/ardupilot/actions/runs/11563094404/job/32185663719?pr=28479#step:10:29)72
antennatracker   0 (0.0000%)     0 (0.0000%)  0 (0.0000%)  0 (0.0000%)                                      615404
arducopter       0 (0.0000%)     0 (0.0000%)  0 (0.0000%)  0 (0.0000%)                                      142344
ardurover        0 (0.0000%)     0 (0.0000%)  0 (0.0000%)  0 (0.0000%)                                      [29](https://github.com/ArduPilot/ardupilot/actions/runs/11563094404/job/32185663719?pr=28479#step:10:30)6528
arduplane        -40 (-0.0022%)  0 (0.0000%)  0 (0.0000%)  -40 (-0.0022%)                                   145824
blimp            0 (0.0000%)     0 (0.0000%)  0 (0.0000%)  0 (0.0000%)                                      593368
ardusub          0 (0.0000%)     0 (0.0000%)  0 (0.0000%)  0 (0.0000%)                                      352236

[Ryanf55]

@IamPete1 Mind taking a look? This is starting to move towards treating altitude frame booleans as private.

It will make this PR smaller: #26328 (comment)

[Ryanf55]

Switched to mavlink_coordinate_frame_to_location_alt_frame , now there's much less code. changealt still works.

[peterbarker]

Please outline what testing you have done on this PR.

[Ryanf55]

Please outline what testing you have done on this PR.

I did in the description. Do you want me to upload logs, copy the exact terminal output, a video, or something more specific?

[Ryanf55]

I'm working on adding an autotest in master to preserve existing behavior. Once that's in, we can revisit this PR.

The test methodology I used for this was not hitting the API I thought it was and I want this checked in CI before touching it.

See #28487

[peterbarker]

So PH and I weren't going to scope creep you on that bit :-)

OTOH, now that you've started - why not early-return on the alt-mask check too? It's going to be a nice function after that.

[Ryanf55]

So PH and I weren't going to scope creep you on that bit :-)

OTOH, now that you've started - why not early-return on the alt-mask check too? It's going to be a nice function after that.

If you have a way to easily test it and are comfortable, maybe we skip CI 🤣

The reason I don't early return on the alt mask is there will be other masks to test against coming up. The alt can be controlled in the upcoming path guidance mode at the same time as other fields in this function.

[IamPete1]

This looks correct to me.

fixed