Name collision with ur_driver pkg in ros-industrial/universal_robot

[gavanderhoorn]

As per the title:

the universal_robot repository already contains a package named ur_driver. Having both this repository (robot_movement_interface) and the universal_robot repository in the same workspace (fi to use the urdfs from the latter) results in errors.

Now you could just say "then don't have them in the same workspace", but I'm curious as to why the name was chosen in the first place, as it is clearly not the same package, and I assume the one from ros-industrial/universal_robot was known at the time the one in this repository was named?

[pablo-quilez]

Hi,

You are right. I just changed during the weekend the name from
ipa325_ur_driver to ur_driver without thinking too much about it. It is
perfectly fine to change it again. Do you have any name suggestion?

The dependency ur_description is the same package from the universal_robot
repository. I copied here into dependencies for two reasons:

• To make it work out of the work in a matter of 30 seconds (git clone,
  catkin_make, roslaunch...)
• Because I know this version works and I don't have the time to test it.

Regards,
Pablo

2015-11-30 18:30 GMT+01:00 G.A. vd. Hoorn notifications@github.com:

As per the title:

the universal_robot repository already contains a package named ur_driver.
Having both this repository (robot_movement_interface) and the
universal_robot repository in the same workspace (fi to use the urdfs
from the latter) results in errors.

Now you could just say "then don't have them in the same workspace", but
I'm curious as to why the name was chosen in the first place, as it is
clearly not the same package, and I assume the one from
ros-industrial/universal_robot was known at the time the one in this
repository was named?

—
Reply to this email directly or view it on GitHub
#1.

[gavanderhoorn]

Do you have any name suggestion?

Well, anything but ur_driver. Perhaps adding a prefix would be an idea. What about rmi_ur_driver?

The dependency ur_description is the same package from the universal_robot repository.

I understand, but it seems to be a rather old version. I can imagine users becoming rather confused, as there are now two versions of ur_description (with the same package meta-data, similarly named files but with different content), and it would also mean that the burden of maintaining this copy now falls upon you.

I copied here into dependencies for two reasons:

• To make it work out of the work in a matter of 30 seconds (git clone, catkin_make, roslaunch...)

Wouldn't rosdep and run_depends solve that? Or even wstool with a suitable rosinstall file?

• Because I know this version works and I don't have the time to test it.

Yes, that is always difficult. Perhaps a rosinstall with the version keys set to a particuar commit could help with this.

[pablo-quilez]

rmi_ur_driver makes sense. I can change it probably on Friday. The
dependency is a little bit trickier and I guess until next year I won't
have time to change it and test it. I'm doing this mostly in my private
time. It would be nice if someone could take a look about it.

2015-11-30 20:56 GMT+01:00 G.A. vd. Hoorn notifications@github.com:

Do you have any name suggestion?

Well, anything but ur_driver. Perhaps adding a prefix would be an idea.
What about rmi_ur_driver?

The dependency ur_description is the same package from the universal_robot
repository.

I understand, but it seems to be a rather old version. I can imagine users
becoming rather confused, as there are now two versions of ur_description
(with the same package meta-data, similarly named files but with different
content), and it would also mean that the burden of maintaining this copy
now falls upon you.

I copied here into dependencies for two reasons:

• To make it work out of the work in a matter of 30 seconds (git
  clone, catkin_make, roslaunch...)

Wouldn't rosdep and run_depends solve that? Or even wstool with a
suitable rosinstall file?

• Because I know this version works and I don't have the time to test
  it.

Yes, that is always difficult. Perhaps a rosinstall with the version keys
set to a particuar commit could help with this.

—
Reply to this email directly or view it on GitHub
#1 (comment)
.