- Mar 29, 2016
-
-
Lorenz Meier authored
-
Julian Oes authored
`sdlog2 start` lead to a segfault.
-
Julian Oes authored
Only free the logbuffer if you exit sdlog2. After disarming, just reset the pointers. This way, we don't take any risks doing malloc and free. The actual "double free" was caused by trying to free the perf counter inside logbuffer_free.
-
Julian Oes authored
This allows resetting of the logbuffer without having to free and malloc again.
-
Lorenz Meier authored
-
Julian Oes authored
-
Julian Oes authored
-
- Mar 28, 2016
-
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Nic authored
added descend into navigation using existing land helper, downstream logic handles ignoring position setpoints (despite being populated), tested with multirotor in gazebo
-
- Mar 27, 2016
-
-
James Goppert authored
-
Lorenz Meier authored
-
Julian Oes authored
-
Julian Oes authored
Timestamps didn't exist when commander was run on the Linux side. However, commander is now always run on the QURT side, so the hack is no longer needed.
-
Julian Oes authored
-
Julian Oes authored
PWM output is now stopped if the mavlink actuator_controls stop being sent. This means props will stop in case the Snapdragon crashes or the cable from Snapdragon to Pixhawk/Pixracer breaks.
-
Julian Oes authored
-
Julian Oes authored
-
Julian Oes authored
The mavlink instance USB can now be shutdown, so this hack is not needed anymore.
-
Julian Oes authored
-
Julian Oes authored
This includes taking care of the arming state and the limit ramp.
-
Julian Oes authored
This is by far the easiest solution for now.
-
Julian Oes authored
-