Indeed, debugger sounds like the next step. I had some recent hangs due to
alignment issues in rpc messaging on an arm, eventually started making
progress by attaching gdb (or gdbserver) to it while it was hanging and
getting a backtrace that way.
Good luck,
Daniel.
On 4 August 2010 09:16, Sebastian Haug <
sebhaug@gmail.com> wrote:
> Hi Daniel,
>
> thanks for the input!
> On Tue, Aug 3, 2010 at 4:31 PM, Daniel Stonier <d.stonier@gmail.com>
> wrote:
> > Just noticed that all of those failed open library calls are smoke and
> > mirrors....its searching in alot of places, but eventually finds and
> opens
> > all of the libraries successfully. So probably no need to worry about
> that.
> >
> > The last function call before the futex is a gettimeofday call. Can you
> > compile a normal program with that call and get it to run?
> Just did that. When I compile using g++ without ROS (or in my package
> but without the roscpp dependency in place) it works, when I compile
> it with the roscpp dependency it gets stuck within the futex after the
> first gettimeofday().
>
> The gettimeofday() call is not the one I try to execute in my code. My
> gettimeofday() call is located after an output to the console, which
> is not displayed, it gets stuck before it comes to echo something.
>
> Probably I will set up a debugger on the gumstix and try to figure out
> which in library and where the futex() and gettimeofday() calls are
> made.
>
> Thanks!
> Sebastian
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
--
Phone : +82-10-5400-3296 (010-5400-3296)
Home:
http://snorriheim.dnsdojo.com/
Yujin Robot:
http://www.yujinrobot.com/
Embedded Control Libraries:
http://snorriheim.dnsdojo.com/redmine/wiki/ecl
(text/plain)