PX4 Maintainers Call: November 07, 2023

:calendar: November 07, 2023

The maintainers meeting is a meeting for the developer team to coordinate on pressing issues and to plan the development of the PX4 Autopilot project, the community is welcome to join and listen, but won’t be able to speak unless specific access is granted ahead of time.

:selfie: Meeting Link

:notebook_with_decorative_cover: Agenda

  • Finalize maintainer role description
  • Identify maintainer opportunities
  • Timeframe/approach for Gazebo on macOS - maybe VM?
  • Open mic for maintainers

Code style decision

@hamishwillee sorry I dropped this forever. Let’s just decide once and for all in this call what we’re doing here and lock it in. (maybe try to see when we can plan the linting update as well)

Rover architecture

Need to sync on short term vs long term output mapping strategy when we have vehicle types that don’t adhere to the force/torque inputs control allocation takes.

Bounty system for PX4

What do you all think of adopting a bounty system in PX4 Repository?

Essentially, we would have either Dronecode or private organizations (any companies using PX4) create a bounty to resolve a bug / create a new feature, and community members would try to claim it by submitting PRs, and once verified, would get the payout :money_with_wings:.

I think this could be cool way to incentivize active contributions & enable companies of PX4 Ecosystem to support the project organically :smiling_face_with_three_hearts:

My questions are:

  1. Can dronecode support this system? (Initial budget of maybe around $1000 to try it out?)
  2. Can we publicize this so that PX4 community companies can provide bounties to develop features?

Read more how it works here: Bounty workflow - Algora Docs

Check out an example fake bounty I created in one of our Issues to see how it would work: RC Switch and Button support and refactor · Issue #19341 · PX4/PX4-Autopilot · GitHub

Duplicated from the discord message: Discord

QGC V6X binary not updated (still v1.13.3)

There’s yet another numerous reports on the v6x target binary for stable release (v1.14) has not been updated properly: [Bug] QGC does not flash release v1.14.0 on Pixhawk 6X · Issue #22253 · PX4/PX4-Autopilot · GitHub

Other v1.14 issues

:memo: Meeting Notes

Excellent - good to have that code thing clear. @rroche I also added discussion of PX4 development on macOS toolchain.

  • IMO this is a precondition to dropping jMAVSim support as that is the only SIM that works everywhere
  • Gus suggests that we might use VM for this. That is toolchain he uses with parallels
  • FWIW I prefer native setups and parallels is not free. But if it works then we should document it as an option
  • The setup should be scripted and in the PX4 tree, ideally the same as for the other setups so that they can be run in CI, maintained, and edited as needed.

Notes on coding style:

Notes on Algora Bounty:

  • We should make sure that companies cant create bounties that is just for their use case & doesnt align with general roadmap of px4
  • We should make sure that bounties dont end up with quick and dirty patch PRs only
  • Giving organic way for companies using px4 to contribute back by creating bounties is a good thing that wasnt possible before
  • As long as bounty bot doesnt interfere with main development flow, its worth giving a try
  • 20% commision fee is very high. Perhaps we can negotiate to a lower fee for higher volume bounty operation
  • About flight testing dashboard idea: We would most likely have small specific flight testers contributing most flights, so maybe this will eventually not be needed, but could be worth it to have all required tests shown in single dashboard conveniently

Notes on terrain hold fix PR:

  • We need to have good tests to catch when these sorts of festures are broken
  • Perhaps something that doesnt depend on gazebo (as it comes with a lot of extra things), since it has dependency & takes a lot of computation. Can we have something quick & detrrministic, like SIH maybe?

Was there any discussion of the agenda item " * Timeframe/approach for Gazebo on macOS - maybe VM?" ?

Nope I don’t think we had that discussion

@rroche Could we have a discussion on initial funding / publicization of the bounty process to try it out?

Nope I don’t think we had that discussion

Why do we have an Agenda then?

At least from what I can recall we didnt discuss it. But now reading back on it I think since it was a single bullet item, it may have been looked over by mistake

Would be cool to have the bounty a pot as well (if possible at all) so the people could also create a worthwhile bounty instead of just companies - 100 people chipping in $5-$20+ each could make something worth a devs time

1 Like