Dev Call Sept 9, 2019 2300 UTC

Issues & Pull Requests




1 Like

Attendee count (max): 19

UTC1103 - Peter Hall is now a member of the dev team

  • Welcome!

UTC1104 -

  • Let GPS log itself rather than replay logging
  • Logging things twice when cores are offset
  • Will also log all instances rather than just the primary core
  • Replay only kinda works at the moment
  • In the future, make Replay JUST log its inputs - not the batched sensor data
  • Merged

UTC1109 -

  • Rejig so it is all contained just within SITL
  • Single threading SITL discussion
    • We create an OSD thread and avoidance
    • But there’s only one thread generally
  • Can be merged once everything’s self-contained

UTC1113 -

  • Fix warnings
  • Merged

UTC1116 -

  • Merged

UTC1117 -

  • Sport PR
  • Needs flight testing a bit
  • Needs a longer review
  • Register_param_table method coming
  • Very light container format
  • Variable-size
  • Make it mavlink compatible?
  • Alex Peter and tridge will have to discuss the approach later in the week
  • Really looking forward to this going in

UTC1126 - AP_Vehicle

  • Appeal from scripting side
    • Common interface at the top allows scripting to be at top
    • Set_mode interface at the top
  • Need to make sure things like the boardconfig stuff still works
  • Need to make sure all the boardconfig stuff works
  • Relatively small changeset
  • Check flash size - fmuv2 before and afterwards
  • Lots of testing ideas added down the bottom
  • Not the start of a “single binary” project :slight_smile:
  • Maybe merge next week

UTC1134 -

    • People have wanted these forever
    • WS LEDs
  • Tom’s old support for them
  • 100 bytes per LED
    • 32x16 LEDs supported
    • But you don’t have any outputs left. Worth it?
  • Scripting support coming
  • Fancy patterns based on flight state
  • Some tweaks to API required
  • Kopis comes with chains of these LEDs
  • Chains on helicopter blades
  • Bunch of stuff on top of Tom’s patches
  • Blheli passthrough working?
    • Will test

UTC1146 -

  • Safety LEDs safety buzzer stuff
  • Adds a whole bunch of functionality to AP_Periph
  • Merge or review required
  • Includes the above neopixel stuff
  • Set buzzer volume
  • Getting reviews is a problem for this low-level stuff
    • Tridge is self-reviewing a bunch of stuff
      • Relying on his bad memory
  • Fix for dynamic node allocator merged
    • Vsnprintf didn’t do %llu

UTC1150 -

  • Needs more testing
  • Tridge wants to know that two connections both doing signing with different GCSs
  • Need to make sure replay attacks aren’t possible
  • Really need autotest stuff to happen

UTC1155 -

  • Randy has looked at it
  • Up to tridge to review again
    • He’ll need to do extensive testing himself given the first set of tests which went on
      • Testing was done prior to the review and was broken during the initial review
        • So tridge reviewed broken code
      • Bill’s flown it again
  • Some perf testing has been done
  • Some problems on Pixhawk v1
    • Bootloader issue?
    • Maybe just the tool from the PX4 repo bailing out?
  • Should be available on the f405 boards due to fewer supported sensors
    • Pixhawk1 1M board will stop working first

UTC0004 -

  • Abort VTOL takeoffs that take too long
  • Rebase and we were happy to merge it?
    • Can’t actually remember
  • merge after deleting the file and passing ci

UTC0010 - Plane update

  • release notes for 4.0.0 beta1 to be written
  • More changes in Copter than Plane
  • Tuning stuff with new loiter controller in quadplane though
  • New filtering code
    • Affecting people’s tuning

UTC0012 - Copter update

  • OA and stay-out zones are still in progress
  • Inclusion and exclusion zones etc
  • Better reporting to the user
    • Too many points
    • No path
    • Etc
  • Randy’s been pushing the limit
  • 33 stay-out circles and it worked
    • 45 did not
  • MAVProxy changes need to go in
  • QGC has a fantastic interface
    • Beautiful
    • But we need the MAVProxy changes in
      • And the Mission Planner stuff
  • ::frprintf debugging in threads is very, very nasty
    • hal.console->printf might have worked better
    • Should we redirect it?
  • Debugging symbols are allowed to change which optimisations are done
  • A minimal program to reproduce is required the FPE problem

UTC0030 -

  • Currently only works on first instance
  • Serial0 serial1 etc?
  • “Physical serial port 2”?
    • Telem2 on a cube?
    • Serial2 is just what’s on the case
    • One of several different names
  • Serial numbers are better than SR parameters
    • SR stuff should go away
  • Serial0 Serial1 Serial2 Serial3
    • 0-indexed vs 1-indexed
      • PWM channels are 1-based
      • Relevant for the LED stuff from earlier
      • User-friendly or programmer norms

UTC0042 - Close

1 Like