Plane: backport support for Python 3.12 to 4.5

[jonhoo]

Backport of #26889.

[jonhoo]

Apologies for the force pushes – my local Plane-4.5 was behind so wanted to make this apply cleanly before review.

[peterbarker]

This is a single commit - was this really a single commit into master? That would be contrary to our usual standards as it's got a mavlink commit in there.

We need to know what else went into that mavlink branch as part of that new submodule reference. We might need to reference a new commit in there if we are uncomfortable with other stuff implied by that mavlink submodule update

[jonhoo]

The original PR was indeed three commits, though the commits that updated the individual subprojects weren't any more informative than "python 3.12 support":

• 31e15c1
• c523dc0

[jonhoo]

The CI failure seems pretty weird (and maybe unrelated?):

ld: building fixups: pointer not aligned at __ZN12AP_FWVersion5fwverE+0x34 from /Users/runner/work/ardupilot/ardupilot/build/sitl/AntennaTracker/Tracker.cpp.47.o

[jonhoo]

Now, it is true that for the backport, this ends up picking up a few extra commits to mavlink, but a cursory look at them suggests they'd be okay to bring in here (i.e., not worth doing a cherry-pick):

ArduPilot/mavlink@1e04d8b...2825252d6d13eb347520f028b3c580ab29bd78e5^

[peterbarker]

@jonhoo this is unlikely to be merged as a single commit.

Thanks for checking out the mavlink commits.

The CI failure happens at random on MacOSX. Fixed in #26700 - I've marked that for backporting.

[jonhoo]

Okay, I have now cherry-picked the individual commits instead.

[jonhoo]

Is it worth also backporting this to 4.3 and 4.4?

[rmackay9]

Hi @jonhoo,

Thanks for this. I don't think it's worth creating a 4.3 or 4.4 backport because we don't plan to do any more releases for those version.

[rmackay9]

This is included in Copter-4.5.7-beta1