It's a source install of ros-latest (revision 10036) installed a
couple weeks ago using rosmake -a -i. The install is shared among a
few machines if you think that might be relevant.
I don't think the issue is an incomplete or inconsistent build. My
experiments involve launching and running executables many times with
varying parameters to my algorithm and varying noise settings.
Everything runs fine for at least the first 100 trials, which I
wouldn't expect if something were inconsistently built. rosout is very
consistently dying somewhere around 100 to 300 trials in, though. Is
it possible that so many calls to roslaunch is somehow overflowing a
buffer somewhere?
If you have instructions for running roscore in gdb and you think the
output would be useful, I could try doing that.
-Mike
> The -11 exit code indicates that rosout segfaulted, which it shouldn't do.
>
> Are you using pre-built binaries or a source-based install? Any chance that you're in an incompletely- or inconsistently-built state?
>
> brian.
(text/plain)