[ur_msgs] Fix domain constants in Analog.msg

[galou]

CURRENT and VOLTAGE should be inverted.

Tested with URSoftware 3.8.0.61336.

Signed-off-by: Gaël Écorchard gael.ecorchard@cvut.cz

[galou]

Cf. discussion in fmauch/universal_robot#8.

[fmauch]

Yes, this fixes a mistake I made earlier. I checked the docs again and this seems to be the correct order.

Edit: E.g. see the RTDE Guide for field analog_io_types

[fmauch]

lgtm

[gavanderhoorn]

@galou: seems your @cvut.cz address is not recognised by GH.

Would you want to fix this before a merge?

[galou]

@gavanderhoorn : what do you mean "is not recognised by Github"?

[gavanderhoorn]

I would expect to see your picture there, instead we see the default GH icon.

[galou]

Actually, the guide states {0=current[mA], 1=voltage[V]} for the field analog_io_types. What I did is launching ur_robot_driver as of commit 936ff8d27e and looked at the topic /joint_states which showed a domain value of 0 when the input type was set to Voltage on the pendant. I think a double-check on a real robot is necessary.

[gavanderhoorn]

looked at the topic /joint_states

io_states I guess?

[galou]

@gavanderhoorn I added the unrecognized email address to my GH settings.

[galou]

I actually meant /ur_hardware_interface/io_states.

[gavanderhoorn]

I won't be able to verify this with real hw for some time.

I do agree it should be verified though.

[galou]

I'm sorry, I shouldn't do two things at once...

My patch does indeed agree with the doc.

What I saw (last week) on the real HW is what is on the patched version, sorry for the misinformation.

[gavanderhoorn]

Approving after checking the Real-Time Data Exchange (RTDE) Guide document.

[gavanderhoorn]

Thanks @galou.

This will take some time to reach the repositories, as Melodic was just synced.

[gavanderhoorn]

It would be really good if we could be vocal about this.

The values in the reported IO states were correct, but the constants in the message def weren't.

However, both consumers as well as producers could be affected by this change, as they could both be using the constants and then run into issues now the values have been switched.

[fmauch]

It would be really good if we could be vocal about this.

The values in the reported IO states were correct, but the constants in the message def weren't.

However, both consumers as well as producers could be affected by this change, as they could both be using the constants and then run into issues now the values have been switched.

Verbal where? Inside the driver, this gets read as raw uint32_t and the value is copied directly inside the hardware interface.

Do you refer to other ROS nodes as producers and consumers?

[gavanderhoorn]

Do you refer to other ROS nodes as producers and consumers?

Yes.

[gavanderhoorn]

See Change of constant values in ur_msgs/Analog.