indent is not available on our documentation server:
[build_parameters.py]
Traceback (most recent call last):
File "./param_parse.py", line 422, in <module>
do_emit(XmlEmit())
File "./param_parse.py", line 415, in do_emit
emit.close()
File "/home/wiki/build_wiki/ardupilot/Tools/autotest/param_metadata/xmlemit.py
", line 24, in close
etree.indent(self.paramfile)
AttributeError: module 'lxml.etree' has no attribute 'indent'
If the RTC time we reset parameters (STAT_RESET) is zero then we reset all parameters.
In SITL we only set that time if we see a GPS due to the AP_RTC RTC_TYPES parameter default bitmask value of 1
So if you reboot SITL fast enough while STAT_RESET is zero you can see that STAT_BOOTCNT does not increase. Easily reproducible in SITL; start with -w and type reboot in several times before SITL sees the GPS come up and you can see STAT_BOOTCNT doesn't increase.
Fix is to ensure STAT_RESET is non-zero before rebooting.
This option is currently ignored, but if given allows the script to
continue normally.
This allows us to change the autotest server to pass the option in - and
then we can change autotest.py so that unless the option is given we
spit out a help message rather than completely stuff up the user's
development environment
We end up where we started, so when we start to play the mission back we
might immediately be at the first waypoint. That's a problem as we may
never see the NAV_CONTROLLER_OUTPUT mention waypoint 1 and thus we
fail the test
We're now waiting for the vehicle simulation to provide us a heartbeat
for a non-generic frame before considering it good to fly.
Unfortunately, Copter relies on the parameter file to tell it which
frame to use - and we don't apply parameters from parameter files until
after we've checked the heartbeat.
Passing the file into ArduPilot on the commandline means we don't have
this problem.
Until ArduCopter allocates its motors backend we emit a generic type,
for which there is no mode map in pymavlink.
So don't consider the reboot complete until we see a valid heartbeat.
This was still relying on heartbeats coming from MAVProxy. As speedup
increased those heartbeats may not come fast enough - and they really
should be coming from autotest as that's who's doing the commanding.
autotest: set SYSID_MYGCS in AFS test
autotest: set SYSID_MYGCS before setGCSfailsafe
Peter Barker
Randy Mackay
Michelle Rossouw
yaapu
Stavros Korokithakis
Samuel Tabor
Pierre Kancir