Dev Call Apr 14, 2025

Development Team
1

Flight Code Release Update

Issues & Pull Requests

2

Attendee count (max): 12

UTC1100 - vehicle updates

  • Another altitude bug found in Plane
    • Old bug, if you are terrain-following in eg. loiter and switch to circle it will climb to current height plus home height
    • Circle is relatively common for short failsafe
    • If someone has a long short-failsafe they might climb a lot
    • Tridge has removed all cases in Plane where we set a .alt field in a location directly
    • 4.6 has a lot of altitude fixes!
  • Plane Pitch level at takeoff alt has an issue if during the takeoff your groundspeed drops below your cruise speed
  • No ap-Message for mavlink id
  • Fov_status
    • Redundant message
    • Peter will make PRs to remove the “no message for id” statustext
  • Log_buf_size on a speedybee causes enomem
    • It probably really did just run out of memory
    • On a directory creation failure?!
      • Perhaps we’re not resetting errno?
    • We have a Wiki page which adds all possible arming errors
      • Adds some advice
        • AI food
  • Remove microbrain rangefinder?
    • We don’t support this?
  • Lua on windows sitl
    • “NEW_NOTHROW” probably requires DELETE_NOTHROW”
  • Velocity divergence before arming?
    • Waiting for Paul, this isn’t a regression
  • IOMCU servos not moving?
    • Possible cause is register corruption
    • We had one case in years where servos didn’t move
      • Some servos not moving
      • Eg. only one aileron movs
      • Reboot and both move
    • We might do a check-registers
      • Tridge will work up a PR
    • We’d assumed a hardware fault and rebooted
    • Noisy serial port between IOMCU and MCU?
    • Only an 8-bit checksum
    • Perhaps poll the registers and make sure they are as we expect them
    • Checking registers doesn’t really help for things like ICE which might stop on a bad value transfer

UTC1132 - GCS_MAVLink: Support requesting MESSAGE_INTERVAL via REQUEST_MESSAGE cmd by nexton-winjeel · Pull Request #28353 · ArduPilot/ardupilot · GitHub

  • Support requesting message interval via request message command
  • Can be merged on CI pass
  • We usually index from 0 in mavlink, but in the future we’ll index from 1 for things like battery, or just break a whole bunch of things by only sending first instance

UTC1143 - AP_NavEKF3: revert to primary compass on ground by rishabsingh3003 · Pull Request #29300 · ArduPilot/ardupilot · GitHub

  • Revert to primary compass in EKF on the ground
  • Merged!

UTC1143 - SITL: Add gausian noise and rate limit to VICON sim by rishabsingh3003 · Pull Request #29524 · ArduPilot/ardupilot · GitHub

  • Add gaussian noise to vicon sim
  • Now have parameter set broken out
  • Merged!

UTC1145 - AP_NavEKF3: Allow disabling lane switching by rishabsingh3003 · Pull Request #29623 · ArduPilot/ardupilot · GitHub

  • Allow disabling of lane switching
  • Allows LUA script to change the lane and not have it magically switch back
  • Question as to whether if the filter is in a very bad state then whether we should shift anyway
    • Eg. NaN quaternion
  • Our lane switching isn’t great
    • Anti-oscillation code isn’t great when we don’t check the switched-to lane hard enough
  • Not switching, which is what this PR does means we still log the lanes
    • Setting EK3_IMU_MASK means we don’t log those other lanes
  • Forcing lane switches with a three-position switch is possible after this PR
  • “Manual lane switching”

UTC1148 - AP_GPS: correct _port array-overread in EAHRS case by peterbarker · Pull Request #29647 · ArduPilot/ardupilot · GitHub

  • Fix _port array overread
  • Merged!

UTC0001 - AP_HAL_Chibios:add hwdef files for PixSurveyA2-IND by mikefenghao · Pull Request #29662 · ArduPilot/ardupilot · GitHub

  • Support for Pixsurvey AP
  • Merged

UTC0003 - Shadow walker8642 dmf4 by Hwurzburg · Pull Request #29727 · ArduPilot/ardupilot · GitHub

  • Dmf4 autopilot support
  • Merged!

UTC0003 - Add rc 2 sub by Hwurzburg · Pull Request #29742 · ArduPilot/ardupilot · GitHub

  • Add rc to sub
  • Preliminary reviews requested

UTC0009 - AP_NavEKF: Allow EK3 cores to run different sources by rishabsingh3003 · Pull Request #29750 · ArduPilot/ardupilot · GitHub

  • Allow ekf3 to run different sources
  • So EK3_SRCn* sources become per-core sources
    • This is the way a lot of users thought things worked anyway
  • Need Randy to review

UTC0015 - Correct filehandle handling in error paths by peterbarker · Pull Request #29755 · ArduPilot/ardupilot · GitHub

  • Correct filehandling in error paths
  • Resource leakage
  • Merged!

UTC0017 - IOMCU - surfaces not moving · Issue #29759 · ArduPilot/ardupilot · GitHub

  • Discussed as part of vehicle update

UTC0018 - AP_Mission: reset loop counter when loop completed by tridge · Pull Request #29763 · ArduPilot/ardupilot · GitHub

  • Reset loop counter if you complete all of the loops
  • This PR changes the behaviour
    • Do we need an option to have the old behaviour?
    • DO_LAND_START fly a circuit and fly a landing
    • If you go-around do you want it to come straight in?
  • We will add an option, with the default being the new behaviour

UTC0024 - add and use a Location::copy_alt_from by peterbarker · Pull Request #29764 · ArduPilot/ardupilot · GitHub

  • Is it valid to copy terrain altitudes?
  • Maybe it isn’t an internal error to take a terrain from another location?
  • next_WP_loc can be terrain which you can’t resolve right now
    • New behaviour is next_WP_loc is terrain you asked for
      • Terrain db can then continue to try to resolve it
      • After a few loops it can do the terrain following people
    • 4.5 could fail to terrain follow as we resolved once
  • MergeOnCIPass

UTC0045 - AP_InertialSensor: disable loop check on sitl by andyp1per · Pull Request #29777 · ArduPilot/ardupilot · GitHub

  • Disable loop check in SITL
  • Can SIM_LOOP_RATE be used instead?
  • … we should actually have an autotest which checks this pre-arm

UTC0051 - close