Mirko Ferrati

Hey @traversaro ! Previously our snapcraft was replacing the needed paths directly in the source code using sed during the snap creation process. When moving to Fortress I realized most...

Indeed I just remembered I could not remove these 2 lines https://github.com/canonical/gazebo_snap/blob/9dbe05ede2530ef0ea43b69b1581a1d0d64b871e/scripts/override_pull.sh#L20-L21 I'll update my PR as soon as the new mechanism works, it will clean up a lot of...

The T_controller was supposed to be used in position (and velocity and torque?) modes since joint PIDs were giving us problems due to a position step too little to generated...

This error is not related to the IMU, but to any analog sensor. If you disable the IMU in the URDF, you will see that the simulation blocks trying to...

Ok, I just found out the single line that produces deadlock: https://github.com/robotology/yarp/blob/master/src/libYARP_dev/src/modules/AnalogWrapper/AnalogWrapper.cpp#L469 If you comment it out, recompile yarp, recompile GazeboYarpPlugin, everything will work (except that you do not get...

@traversaro Inside `analogWrapper::close()` there is a call to `Stamp::Update()`, which in turn enters `yarp::os::Time` and deadlocks, I don't know it the reason is `delay()` or some different function ( I...

Yes, that one. I am testing my changes in a larger URDF model, there may be problems even with the ControlBoards, but they seem to raise only if another AnalogWrapper...

No no no, `close()` calls `join()` on the thread `run()`. But if `run()` is blocked waiting for a clock do unblock some yarp::os::Time primitives, the `close()` is blocked too. Anyway,...

Update: I have the ServerInertial closing fine, but I had to put an hack in its close function in order to tell our IMUDriver to stop reading from Gazebo. How...

Branch issue181 works as expected. I don't think we can merge the hack in master, but as soon as someone knows how to do the fix in the yarp way,...