On Thu, Jan 13, 2011 at 1:41 PM, Michael Krainin
<mkrainin@cs.washington.edu> wrote:
> That seems to have done the trick. Thanks!
> -Mike

I've tweaked the patch somewhat and commit to the ROS 1.2 branch:

https://code.ros.org/trac/ros/changeset/12843

 - Ken

> Author: Ken Conley
> Date: 2011-01-12 22:31 -800
> To: User discussions
> Subject: Re: [ros-users] socket-related error
> Thanks for sending that along. It appears that a bug I had in cturtle
> is preventing me from seeing the relevant message, but I have some
> hypotheses.
>
> I tested the attached patch on my diamondback branch, it should work
> in cturtle as well. It changes the very end of
> rosib/src/roslib/xmlrpc.py. This logic should have the serve_forever
> re-enter on interrupted calls. Because the cturtle code has the
> (non)-logging bug, I also threw in a commented out
> traceback.print_exc(). If you find that the patch isn't working, then
> uncommenting that traceback would hopefully reveal the true cause.
>
> If this patch works, let me know and I'll integrate it.
>
> cheers,
> Ken
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>