this fixes an issue in selecting the best lane to change to when we
have 3 or more EKF cores. The bug is that if the current lane is
unhealthy it would always choose the last healthy lane instead of
choosing the lane with the lowest score
this fixes an issue in selecting the best lane to change to when we
have 3 or more EKF cores. The bug is that if the current lane is
unhealthy it would always choose the last healthy lane instead of
choosing the lane with the lowest score
this is the first step to using AP_RCProtocol for most (or all) RC
input on Linux boards. It also fixes a major bug on Disco where the
old SBUS decoder didn't use strict enough timing checks to filter out
bad frames when video capture is running. That led to erratic flight
mode changes and sometimes loss of control.
This allows to use any device to send RCInput to Disco via network.
Particularly this is useful when paired to additional changes on
SkyController 2 so it can be used to control Ardupilot.
This was tested with https://github.com/lucasdemarchi/dema-rc running
on SkyController 2.
g++-9 throws a lot of warnings for taking addresses of packed members.
We can have this warning for our ArduPilot code, but there's way too
many issues within MAVLink to put up with.
this is never expected to trigger unless we have a severe MCU error as
SPI transfers don't rely on a response from a device.
The only case that we could get a timeout is when a bug leads to use
doing transfers from memory that does not support the DMA
transaction (such as on H7). This change turns that from a immediately
fatal lockup into a bus error and failed sensor
CubeBlack-Solo hardware def to use FMUv3 as its parent rather than CubeBlack. Since this older cube doesn't have the same hardware as a new CubeBlack, it is causing conflicts. Nearly everything is the same.
Randy Mackay
Peter Barker
Peter Hall
Andrew Tridgell
Michael du Breuil
Tom Pittenger
Jaaaky
Lucas De Marchi
Pierre Kancir
Vinicius Knabben
Mark Whitehorn
Matt Lawrence
murata
Jeremy Feltracco
Guglielmo