Some threads do not exit and are still running when
trying to exit SITL running under Windows in Cygwin.
This problem was introduced with:
- posix shell #10173 because of strating a child process
for the startup script and mixing up the signal handling
(only Ctrl+C broken)
- lockstep #10648 because of simulator threads not
stopping gracefully anymore with timing race conditions
(no graceful exit possible anymore)
I leave the SIGINT handler on its default implementation for
Cygwin which kills the process and all its threads when pressing
Ctrl+C.
This hotfix at least allows the usage of Ctrl+C again instead
of forcing the user to use the task manager to shut down
px4.exe going crazy on CPU load instead of exiting
everytime.
instead of the derivative of the vertical position which is
not the same value when using the ekf2 estimation.
Using the position derivative resulted in delayed landing because
the value for some reason often bumped over the theshold before
slowly converging towards zero while the velocity was within expected
accuary in all these cases.
There were accidents because when missions lead through an obstacle
and it should be avoided but the setpoints from the external obstacle
avoidance module are suddenly missing the mission is continued into the
obstacle which results in a crash.
In 5bb9babc20 I made MAVLink send a
rounded up integer instead of rounded down. This makes sense in practice
because the low battery reactions happen exactly when the reported number
switches in the UI. But here we want to provoke an exact 50% in the UI so
we stop counting at 49.9%, it get's rounded up and we see the expected
result.
In 3e6e1f5c2b the simulated battery
percentage was reversed. I'm assuming because of the possibly
missleading variable name. Now I'm fixing it by switching the
maximum and minimum voltage such that the name is not misleading anymore
but it still works as expected.
Setting PWM_DISARMED to 0 results in no PWM output to the ESC for the pusher motor.
Most ESCs start beeping endless in short intervals if they don't get a signal.
I remove changing this parameter which is 900 by default to always command the motor to stand still.
Necessary because NuttX defines pollfd differend from
the POSIX standard. The difference of field order makes
use of C++ designated initializers to compile for both
platforms impossible.
If the landed state was set before the position controller starts it
would get updated with the first poll_subscriptions() anyways. But it's
not on all posix and nuttx platforms I tested and the deleted statement
just copies the zero message into the struct which is already zero.
Because it was always failing from the beginning on and
we want to make sure no other tests break in the meantime
by running the currently passing tests also on Windows CI.
to save the time running the installer for every build
which was necessary before because the folder exceedes
the maximum of 1GB cache of the appveyor free plan
to fix Cygwin upload. It failed silently but when catching it prints
"non-standard baudrates are not supported on this platform".
Discussion about platform independet FTDI detection is in issue #10429.
Latest QGC daily supports sending an orbit command. This commit applies
the correct radius, rotation direction, velocity and center position.
Still missing is a check to have global position available and applying
x,y coordinates and altitude independently.
For the geo.h C library to convert between global and local frame
we need an up to date global reference which is currently inside the
vehicle_local_position message.
Switch to positive tangential velocity being clockwise because with NED
frame z-axis points down and mathematically positive around the z-axis
results in clockwise yaw rotation when seen from above. Also MAVLink
messages and other components in PX4 are defined this way.
The handling for vehicle commands inside the FlightTasks library
was already there but commented out because the previously used
MAVLink messages were tied to the specific application we used and
there was no clean common message definition. Because there is now
a well defined message for the orbit task I'm uncommenting and using
the working message handling again here.
Nice smooth altitude control is already provided by the existing
flight task for altitude mode. We inherit from it to resuse the
functionality and just all its update first not altering the setpoints
for the z-axis it produced.
Maximum velocity achievable in 4 seconds,
maximum radius achievable in 8 seconds for futher testing.
These should be parameters as soon as it works fine.
It scales the yawspeed setpoint arbitrarily by default with 0.5.
This makes no sense because when you give a setpoint of 1rad/s then
you expect the setpoint to get executed. If you want manual yawspeed
response to be less agressive on the stick use the scaling parameter
for the stick MPC_MAN_Y_MAX.
More than a year ago I started the easy to use math::Functions to handle
the always used mathematical SISO functions to be tested and available.
I switched x, y and z stick input to the freesh programmed deadzone and
exponential functions from the library to unify and clarify their use.
I just realized yaw was left over because it lead to a drift problem in
certain new use cases.
Now I'm just adding the yaw stick to the already well working method.
to disable automatic evaluation of /dev/ttyS* expression in cygwin.
On linux verbatim seems to be default anyways. At least it doesn't get
evaluated on linux.
In the Cygwin environment the native Windows serial COM# ports get mapped
to /dev/ttyS# for POSIX compatibility. While # is one number lower inside
the environment than the COM port number because it's 0 indexed instead
of 1.
I added the necessary handling to all the dependent parts I found which
allows uploading to /dev/ttyS# when the cygwin platform is detected.
Now the usual "make px4fmu-v4 upload" and
"./Tools/upload.sh build/px4fmu-v4_default/px4fmu-v4_default.px4" work.
Thanks to @bkueang 's review comment I switched to an explicit check for
an empty sting instead of a condition that could be theoretically true
in other cases and is less readable.
Type "man test" on your terminal to read up what -n stands for.
It seems that on linux only inside a makefile the parameter after the echo
command gets printed if no single quoted sting comes afterwards so I had
to switch to single quotes such that I can use the parameter.
Force interpretation of backslash escapes with the parameter -e of echo.
Switch to a lighter blue because on certain terminals default blue is
hard to read on black background.
The shell script which checks the style relies on greping for the
keyword "Formatted" in the output of astyle. But the program has
localization support and will output in other languages e.g. german.
This leads to all style checks always succeeding. I only tested this
on Windows in Cygwin but I can imagine the problem also exists in
non-english Ubuntu installations.
Solution is the parameter --formatted of astyle which only produces
any output if there was something to fix. This allows for a display
language independent condition for an empty string inside the shell
script.
Matthias Grob