20190529 Dev Meeting

Agenda

General:

• Update from relevant previous calls
• Tagged: issues, PR.

Notes

Call 1

Discussed the following issues

• Clarify RFC process vs PRs - PRs for single message changes. AI HamishW to update the docs to clarify.
• Done #926 Add new PWM output message - Looks fine - just push in. Done by Hamish in #1148 Add Actuator output message
• Done #1142 OBSTACLE_DISTANCE: add higher angle resolution and angle offset - @HamishW to merge - looks sensible.
• Done 1101 MAVLink Link status - @Lorenz to review and merge
• #714 Add C++11 headers to c_library_v2 - probably better to create a CPP library. AI @Lorenz to create this.
• #1138 common: add ACCELERATION message - @Lorenz propose LOCAL_POSITION_NED_COV as alternative.
• #1147 Manual Control add 4 more joystick axes (channels 5-8)
  • Concern that this turns the MANUAL_CONTROL message which has unambiguous purpose independent of flight stack into one that requires separate parameter support/interpretation for flight stacks - ie it is more like RC_OVERRIDE
  • Also there are other messages that do the things this is intended for, so we can use those, and if needed fix missing cases.
  • We would consider an extension where the purpose of each channel is defined in an enum, but leaning towards extra messages to reduce the cost for cases where this is not needed.
  • We don't want to block Mike. He to go back and think on this some more.
• #1141 The latest released version is now six years old Dev Call
  • The "release story" for MAVLink is confusing for users because there are multiple things that can change
    • source code for generators (in mavlink/mavlink and ArduPilot/pymavlink repos), and these have different release cycle and numbering from each other and from messgaes
    • built libraries - e.g. for C and Pymavlink
    • MAVLink version - 1 and 2
    • MAVLink subversion - from messages.
  • The source for MAVLink/mavlink is on 1.0.11. This matches nothing, so there is no harm on making release (given above confusion). It is just a blip on the release tab.
  • Done AI Hamish to approve a release - see https://github.com/mavlink/mavlink/releases/tag/1.0.12
  • AI Lorenz to clarify release process (or delegate) - issue tracked as #1149
• #1146 Add a constant to MAV_AUTOPILOT for a COEX Pelican charging station - Needs more detailed discussion in dev call. Invited them to attend

Discussed Common Flight Modes proposal

• Proposal is pre-alpha/early thinking
• Idea is that:
  • there are a few common flight modes across every stack - mission, position hold, loiter. We would like to make these part of the spec so that a GCS could implement them without having to know anything about the flight stack.
  • Custom modes could be defined in an definition file (like camera API) so that a GCS could provide a UI without having to understand the semantics.
  • Hope to further refine the idea in the next call.

Brief discussion of "MAVLink 3"

• Hope to put together a working group in July
• MAVLink 3 is a misnomer - there are a bunch of things that we'd like to do (e.g. in security/encryption) but it may be possible to do them within the MAVLink 2 architecture. TBD

AI:

See above!

Attendees

HamishW, JamesP, JonasV, JulianO, Lorenz, Stone, Mike Lyons, MatejF