ROS2: Cannot set/publish JointState with user-specified values

[kschwan]

ROS2: Cannot set/publish JointState with user-specified values

[kschwan]

It turns out I hadn't noticed that I was getting this error from rqt_publisher

[ERROR] [1663763988.089399671] [get_message_class]: Malformed message_type: sequence<string>
Traceback (most recent call last):
  File "/opt/ros/humble/lib/python3.10/site-packages/rqt_publisher/publisher.py", line 161, in change_publisher
    new_text = handler(self._publishers[publisher_id], topic_name, new_value)
  File "/opt/ros/humble/lib/python3.10/site-packages/rqt_publisher/publisher.py", line 244, in _change_publisher_expression
    success, _ = self._evaluate_expression(expression, slot_type)
  File "/opt/ros/humble/lib/python3.10/site-packages/rqt_publisher/publisher.py", line 329, in _evaluate_expression
    if successful_eval and isinstance(value, slot_type):
TypeError: isinstance() arg 2 must be a type, a tuple of types, or a union

[gbiggs]

Could you try Rolling and see if the same bug is present there as well?

[kschwan]

Could you try Rolling and see if the same bug is present there as well?

Sure! Thanks for the suggestion.

If I type in a list in the position expression, I get the error shown below. If I type in an array.array (fx array('d', [123])), it works. The sequence<double> type is a bit misleading though.

Publisher._evaluate_expression(): failed to evaluate expression: "[123]" as Python type "<class 'array.array'>

[gbiggs]

Are you happy with that solution?

[kschwan]

Are you happy with that solution?

Yes, I'm fine using rolling and typing in an array.array, thanks! It would be nice if typing in a list would work too though -- that's how it was on ROS1.

[clalancette]

Yes, I'm fine using rolling and typing in an array.array, thanks! It would be nice if typing in a list would work too though -- that's how it was on ROS1.

Yeah, I agree, that would be much nicer. I did some work on this earlier this year, but I never got it completed. Hopefully I can find some time to complete that.

That said, this is similar in spirit to #23. @gbiggs I'm not sure if you want to close #23 (since technically that one works now) and leave this one open, or vice-versa. Either is fine with me.

[gbiggs]

I'll close this one, since the other looks like it still has proposed changes that haven't been PR'd.