Dev Call Dec 17 2018 2300 UTC

Issues & Pull Requests



Common

  • AP_Periph work
  • NuttX removal

Plane

  • Onion Update

Copter

  • Release update

Rover

  • Release Update

Attendee count (max): 21

UTC2302 - https://github.com/ArduPilot/ardupilot/pull/10023

  • Merged!

  • Bisection for flight characteristic regressions coming

UTC2306 - slew rate limitting

UTC2308 - https://github.com/ArduPilot/ardupilot/pull/9844

  • Separate fixed wing landing radius etc

  • No MdB

  • MdB appears to have addressed tridge’s initial comments

    • Particularly for non-achieved stuff

      • Extra check added
  • merged!

UTC2309 - AP_Periph

  • PhillipK, Jani and tridge have been working on CANBus periperhals

  • Build target for AP which builds a firmware which isn’t ardupilot but is a peripheral firmware

  • Provides a CAN node

  • Initial target is for the little generic CAN nodes that Jani is selling

  • Hwdef to define pinout

    • Easily retargetable
  • Passthrough of i2c, spi, uarts

    • Appear in HAL as being local i2c etc buses

      • So existing drivers should just work

      • I2c remotely over CANBus…

        • That’s the plan, anyway

        • I.e. doesn’t use the remote-magnetic-field CAN messages

      • No new sensor drivers needed?

      • Won’t be able to share the device across multiple autopilots when remoting the bus like this

  • IO expanders, essentially

    • More functionality than existing CAN units as AP drivers are more full-featured

    • Pavel’s not a great fan of this way of doing things…

  • Currently just comes up and blinks LEDs

  • Allocation of node/bus ids?

    • Canbus parameters set via mavlink
  • Will probably need to delay bootup for compasses and the like so the remote devices are ready

  • Using Libcanard

    • Very easy to work with and nice
  • Still uses CAN messages, so interoperable with other CAN things on the same bus at the same time

  • Softserial

    • Coming at some stage….

    • F7 and F3 both support pin inversion

    • Softserial means no external inverter needed on F4

    • We already have some softserial decoders

      • The RC stuff, for example

      • Blheli passthrough too

      • Abstraction needed!

  • We have option bits on serials now

    • Swap RX and TX in software on F7….

      • Should be able to do ows easily

        • F4 should also be able to do this half-duplex stuff
  • PixRacer has internal inverter

  • Full support only on F7s for these serial options

UTC2328 - NuttX removal

  • Just ONE conscientious objector at the moment

  • Px4v2 no longer builds on master

  • ChibiOS builds fmuv2 no worries

  • CI failing for every PR

  • Could put effort into making v2 smaller again

    • But really just want to remove it
  • PR wasn’t rebased

    • So it didn’t fail CI
  • New landing stuff for Plane was the objection?

  • 8k a few months back when we disabled the bootloader on px4v2 and 1

  • Chewed through that…

  • Can we just pull the pin now?

  • A branch is probably going to have to happen until the conscientious objector can use ChibiOS

  • RM: a fair bit of time has gone by since we discussed this

    • Warnings went out a long time ago

    • Leonard’s moved over

    • MissionPlanner build has been solved

    • Randy has no concerns any more

      • How much of tridge’s time do we want to waste?
  • This is master, our playground for the next release

    • Tridge doesn’t want to do another release of Plane with NuttX supported

      • Confusing for users
  • No MdB on the call

UTC2335 - Plane onioning

  • Rebasing is hell for Tom ATM

    • Tridge could help

    • Might just restart instead

  • TP is off after a week

    • Hopefully get it done before then

UTC2336 - Randy and Copter

UTC2339 - Randy and Rover

  • Peter is looking at more message splitting

  • See Jaime’s video of Aion Robotics rover driving around using Cartographer

    • It’s great

    • And there’s documentation floating around for it on the Wiki!

UTC2344 - circling back around to NuttX

  • Need a decision!

  • Discussion / browbeating to happen

  • Will go in in next few days if no response can be had

UTC2346 - https://github.com/ArduPilot/ardupilot/issues/9372

UTC2354 - next meeting?

  • Probably not xmas morning….

  • 1st will be a “casual” meeting

    • No agenda as such

    • Probably lightly attended

  • Next official Jan7/8

  • Thanks everyone for all your work this year!

    • Vast amounts of stuff done!
  • Congrats to tridge on his new degree

    • Credit for the AP