Hi Konrad,

While I'm sure the developers of the motion_planners stack will take
steps to remove PR2 dependencies, I thought I should point to our
stack version policy, which may clarify things:

http://www.ros.org/wiki/StackVersionPolicy

motion_planners is 0.2.4, which is described as
"research/experimental... some review has occurred, and we are
migrating this stack to more stable APIs". It will not be stabilized
for Diamondback, but you may be interested in this ticket about
trajectory_msgs:

https://code.ros.org/trac/wg-ros-pkg/ticket/4675

Also, motion_planners is distributed in the 'pr2all' variant of ROS
which is described as "ROS plus PR2 and bleeding edge
research/experimental stacks." Only stacks in the stable 'base'
variant are expected to be robot generic.

 - Ken

On Sat, Oct 23, 2010 at 7:06 AM, Konrad Banachowicz <konradb3@gmail.com> wrote:
> Hi all,
> I currently run robots in our lab on ROS.
> With mobile everything goes fine, but when i started with manipulators.
> I found that in ROS is some strange dependencies, in manipulation stacks is
> many dependencies on PR2 specyfic packages. For example
> pr2_controllers_msgs/JointTrajectoryActionGoal is used by many of the common
> stacks. trajectory_msgs is also part of one of PR2 stacks. This kind of
> dependencies should be strictly forbiden.
> There are some PR2 specyfic packages residing in common stacks, the best
> example i think is interpolated_ik_motion_planner which reside in
> motion_planners stack. And documentation do not mention about PR2 specyfic
> of this package.
> Pozdrawiam
> Konrad Banachowicz
>
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>