ci: use dev-container for nuttx targets

[alexcekay]

Change

This PR changes all CI workflows that previously used a NuttX container to use the new dev container.
Build all targets was tested here: https://github.com/PX4/PX4-Autopilot/actions/runs/15064605660/job/42346543522?pr=24832
FLASH savings on v5x and v6x are ~44K

Open topics

Tagging strategy

As there is no latest or stable of the dev container yet, I used the most updated dev container. This would thus require updates when new dev containers are created in the future.

Is there a container tagging strategy yet?
Just adding latest automatically on each dev container build workflow run is not a complete solution as this would also trigger when creating a RC for an old PX4 release. I see the following options:

• Have some magic in the workflow that sets latest to the highest SemVersion
• Manually set a latest or stable tag for the dev container

@mrpollo: Can you manually create a new updated dev container that we can reference in this PR? Or do you guys prefer another strategy.

Checks not working

Fail 1:
https://github.com/PX4/PX4-Autopilot/actions/runs/14971242279/job/42052457639?pr=24832
This seems to be fixable by adding a safe directory. Does not depend on GCC as it also fails when using the old GCC, thus related to the newer git contained in the dev container.

Fix is contained in this PR.

Fail 2:
https://github.com/PX4/PX4-Autopilot/actions/runs/14971242273/job/42052457618?pr=24832
This seems harder to fix, had a look, but don't see the reason yet. Seems to be a problem with the new version of gcov, so no visible functional problem.

This needs to be fixed in a separate PR as it will require a deeper investigation. Thus I did not use the dev container in the checks.yml.

Flash Analysis not working

https://github.com/PX4/PX4-Autopilot/actions/runs/14971242276/job/42052457576?pr=24832
Known problem with some GCC versions, see here: google/bloaty#362

Easiest fix is not displaying compileunits but symbols instead.
The output will then look like this, which is as helpful as the previous one in my opinion:

    FILE SIZE        VM SIZE    
 --------------  -------------- 
  +0.0%    +296  +0.0%    +296    .text
    +0.8% +1.29Ki  +0.8% +1.29Ki    g_cromfs_image
    [NEW]    +152  [NEW]    +152    land_detector::RoverLandDetector::RoverLandDetector()
    [NEW]    +122  [NEW]    +122    land_detector::RoverLandDetector::_update_topics()
    [NEW]    +120  [NEW]    +120    commands.26190
    +8.2%     +60  +8.2%     +60    AirspeedValidator::check_first_principle()

Fix is contained in this PR.

[alexcekay]

FYI: @sfuhrer, @MaEtUgR

[mrpollo]

Thanks @alexcekay, this helps a lot. I intended to review all of the CI jobs and replace them in one single pass, but I was waiting for an updated container tag.

From now on, it will be best if we only push containers for tagged PX4 releases. However, we should have the flexibility to push a container tag on demand, which is why I made #24821. Hopefully we can get it into main soon, and push a new tag for CI

[dagar]

Looking good, let's just be sure to review a few flight test logs.

[dagar]

The flash savings is incredible.

[sfuhrer]

Looking good, let's just be sure to review a few flight test logs.

@dagar anything specific you look out for in flight testing?

[DronecodeBot]

This pull request has been mentioned on Discussion Forum for PX4, Pixhawk, QGroundControl, MAVSDK, MAVLink. There might be relevant details there:

https://discuss.px4.io/t/px4-dev-call-may-14-2025-team-sync-and-community-q-a/45516/3

[DronecodeBot]

This pull request has been mentioned on Discussion Forum for PX4, Pixhawk, QGroundControl, MAVSDK, MAVLink. There might be relevant details there:

https://discuss.px4.io/t/px4-dev-call-may-14-2025-team-sync-and-community-q-a/45516/1

[mrpollo]

Here's logs from flight testing done by @afwilkin (originally posted here)

Pixhawk 6C - PASS

• manual control: https://review.px4.io/plot_app?log=7484e0c9-14d9-4576-a43c-491a62576b0a
• altitude control: https://review.px4.io/plot_app?log=1a04ce17-8238-4680-9433-9bef010d316c
• position control: https://review.px4.io/plot_app?log=a70618d4-d2e1-4eee-b3a2-697b35c502a9

NXP - PASS

• arm and disarm: https://review.px4.io/plot_app?log=9f0d571b-fc9b-484f-a35b-ab26a6ffcdb4
• Manual Flight: https://review.px4.io/plot_app?log=65b5212a-a503-40fc-abb7-b1d8856e0e38
• Altitude Flight: https://review.px4.io/plot_app?log=05b671a6-ba82-42e0-8cd7-f6b60046910e
• Position Flight: https://review.px4.io/plot_app?log=5a29030c-acfe-4a92-8a91-27de5ba3194a

Pixhawk 4 - PASS

• arm/disarm: https://review.px4.io/plot_app?log=f2939a7e-3c7a-4c7a-93d2-451956efb0db
• manual: https://review.px4.io/plot_app?log=7a04e334-e68d-45f5-94ec-b094dc54c53f
• altitude: https://review.px4.io/plot_app?log=37b05b11-0d04-4277-81a8-d6664741b6df
• position: https://review.px4.io/plot_app?log=d334618f-b15d-428a-80ae-53d2e174395a

[alexcekay]

I manually built a new dev-container using the newly action of @mrpollo (thanks 👍). Then I updated my change to use that newly built container.

The build all targets run is here and looks good: https://github.com/PX4/PX4-Autopilot/actions/runs/15064605660/job/42346543522?pr=24832. Also downloaded a build artifact and flashed it onto my v5x, which also looks good (GCC13 + ITCM runs fine).

[sfuhrer]

I wouldn't know what's holding us back from getting this in, thanks a lot for pushing this forward @alexcekay !

[mrpollo]

I think I can fix the failed build pretty easily, give me a few to push a couple of commits