Dev Call Oct 26, 2020

Issues & Pull Requests




1 Like

Attendee count (max): 17

UTC1100 - Copter update

  • twitching pitch works on 4.0.4
    • So not a blocker for 4.0.5 any more!
    • 4.0.5 will be out today or tomorrow

UTC1102 - Plane update

  • Tridge looking at Nate’s log

UTC1103 -

  • Bi-directional telemetry logs
  • This has been a very long time coming
    • Should we hold this up?
  • Wire Protocol isn’t as good as it could be
  • Use mavftp?
    • No memory on the transmitter side!
  • Community probably won’t mind too much if we change the protocol later
  • Flash cost?
    • Why do we have so much flash free at the moment?

UTC1118 -

  • Show frame type as text via mavlink
  • Have a const-char-* on the motors backend
  • Randy will push this one along

UTC1123 -

  • Adds missing betaflight frame reversed
  • merged!

UTC1123 -

  • Enables proper OOP type stuff
  • Uses a lot less memory than e.g. tridge’s PID example
  • MdB will try to get an example together for using this

UTC1125 -

  • Fport2 support
  • And cleanups so the flash cost is minimal
  • Completely different protocol to fport1, near-enough
  • Test-suite for fport and fport2 added
  • Should run the rc protocol tester as part of CI
  • Fport and fport2 are inverted 115200
  • Peter will merge after the call after looking at it

UTC1127 -

  • Fix css on
  • What testing has been done?
  • Needs to be run on the autotest server via ssh

UTC1135 - looking at flash

  • MO looked at DB and found a big gap
  • Need to look at why the big drop…

UTC1137 -

  • Race-condition causing watchdogs
    • Probably only on ground
  • Stop two threads doing filesystem operations at the same time
  • All filesystems operation must go via AP_FileSystem and this enforces that

UTC1145 - fix build on 9.3

  • Merged!

UTC1146 -

  • Needs review comments to be looked at

UTC1146 -

  • Add midposition option
  • Merged!

UTC1148 -

  • Merged!

UTC1148 -

  • Log banner
  • How does this actually help?
  • Does the obfuscation actually help?
  • Randy doesn’t think it adds value or will dissuade people
  • Inflicting this on all users
  • Only to dataflash logs…
  • No appetite for this
  • We do NOT want people running AP on manned stuff, but this isn’t really going to help
  • Craig reached out to the EAA article author to point out our policy and they might print a retraction of some description

UTC1151 -

  • AP_Scripting add parameterto disallow scripts from places
    • E.g. don’t load from ROMFS
    • Defaulting to not running scripts on SITL on Linux
    • Tridge likes scripts on ROMFS on Linux
      • So change default to retain current behaviour
  • Tridge had a patch to have scripts on SD card override scripts in ROMFS
    • ROMFS forces the script to run

UTC0003 -

  • Fix issue previously added with tailsitters
  • Merged

UTC0005 -

  • First conservative GPS_AUTO_SWITCH
  • Behaviour change…
  • Primary == 5 will force to zero
  • Should be able to fly with only a second GPS
  • Blended as primary?
  • Naming issue?
  • Choosing what to switch to can be problematic
    • E.g. F9 not using SBAS only etting lock-type-3 and M8 giving type 4
    • E.g. RTK?
    • More work will be done on this

UTC0017 -

  • AP_Airspeed: add dedicated wind max param
  • Better logic around GPS going away added since last time
    • you will get your airspeed sensor back
  • Parameters are global

UTC0021 -

  • Add support for NMEA water sample
  • Adds matlab nmea water speed sensor
  • AP_Airspeed/ models
  • How does this work on negative waterspeeds?
  • Doesn’t fuse this into the EKF at this point
    • Logged only

UTC0024 -

  • Merged!

UTC0028 - closed

1 Like