20190918 Dev Meeting
Hamish Willee edited this page Sep 19, 2019
·
2 revisions
General:
- Update from relevant previous calls
- Tagged: issues, PR.
-
Common Flight Modes Microservices
- Lorenz
- Discussion of broad goals. Note that this is scratchpad - in no way defines mechanism, just first broad suggestion of approach that might work.
- Aim of first discussion mostly to get agreement that this is something that we want to be able to do
- Should be able to add a new mode to flight stack without having to recompile the groundstation to support it.
- Hamish: Also nice to have easy mechanisms to set/get modes without having to do logs of setup - e.g. takeoff should not require user to think about arming. Lorenz agreed but is orthogonal problem.
- JamesP - we should start with small set of non controversial modes.
- JamesP - modes have same name on different flight stacks and vehicles. Do we go something truly common or do we have a full set and then choose.
- JamesP - Camera API seen as complicated due to XML compared to the old messages
- Lorenz noted that the complicated bits are optional
- you have a basic API for simple camera API that is similar to old camera messages.
- XML file is optional, but allows complicated scenario to be supplied by the vehicle.
- We want something that captures this same idea for flight modes - simple methods for what is common, with optional flexible methods for extensions.
- JamesP - Perhaps also flight-stack-definable modes that can be part of the basic API too.
- Lorenz noted that the complicated bits are optional
- General agreement that any XML format would be optional part of the spec.
- Action on Lorenz to do first evolution of spec/requirements.
- Lorenz
-
#207 Document Tunnel protocol
- Lorenz interested in providing a generic end to end tunnel that implementers can use without having to think about the encoding or anything else.
- No particular objection to way ids for tunnel encoding formats are being reserved. Just waiting on Julian to complete discussion.
- Some objection to enabling support for documenting opaque tunnel encoding formats; we don't like tunnels, so we think they should be rare and publishing encoding format not needed.
- Action on HamishW to document this for:
- #1233 Register 10 tunnel payload codes for STorM32
-
#1232 Add component ID for a TUNNEL node
- Lorenz OK with this if it is effectively required to provide a heartbeat for a router.
-
DONE #1231 common.xml: NMEA 0183 passthrough message
- Strong opposition - either messages are useful in MAVLink context or they are not. So lets see whether the set does as needed. Action on Hamish to share that info.
- In particular, NMEA is a pretty common standard that might be useful to capture in MAVLink.
-
#1228 common.xml: add AIS message
- Action: JamesP to do due diligence and confirm these are sane.
- James managed to get the two fuzz testing PRs in, but still working on the multiple-level includ PRe. Not likely to be in immediate future.
Did not discuss:
- JulianOes Absent did not discuss "MAVLink connection semantics
-
#929 No way to transmit airspeed reference
- JamesP: ArduPilot does not use
NAV_CONTROLLER_OUTPUT
, so will need to investigate how it sends this information in order to surface possible options.
- JamesP: ArduPilot does not use
- #1234 Add messages for power regulation and temperature status reporting
- #1216 SET_POSITION_HOME: Add frame for local coordinates
- JulianO to discuss microservices proposal.
See above!
- HamishW, JamesP, DennisM (stifael),
- JulianO unable to attend