Christoph Fröhlich

We know of the failing CI of the binary builds, and the coverage diff of 0.05% is some "noise" you don't have to worry about.

> Hi @bmagyar ! After fixing the format issues, I synced my fork with the master since its been a while. I now have a build error in rolling binary...

Hi. Thanks a lot for your code example. I haven't used diagnostics messages in ROS 2, and it doesn't seem to be very common having in mind that there isn't...

Maybe it's worth it to implement a `RealtimeTFBroadcaster` based on the [realtime_publisher](https://github.com/ros-controls/realtime_tools/blob/master/include/realtime_tools/realtime_publisher.h)?

Thanks to the detailed and verbose description. I totally agree to `state_msg` and `state_current` issues. > The first update of the controller after it was activated does not generate any...

> * only one test (state_consistency test) really tests the `state_msg` has expected values, other test seem to use it to read the internal state/last state. That is not true...

I'm sorry, I haven't read your statement properly. I agree :see_no_evil:

I discussed this with some colleagues today, and the common opinion is: it depends :laughing: - Using a time-discrete controller, you are already loosing the assumption of an "ideal robot"...

> I have used this controller on a 4 wheeled robot. I am able to move it forward, backward, rotate in left and right direction(Yaw), but i am UNABLE TO...

We have a demo now in the demos, see https://github.com/ros-controls/ros2_control_demos/pull/318 Is there anything more inside this PR, or can we close it?