On 02/13/2011 08:24 PM, brian Corwin wrote:
> As another interesting thing, I installed diamondback, and tried to build NI and it crashed on making one of the PCL
> packages.
>
> I overlayed the unstable perception pcl and addons and it segfaulted again on
>
>    /home/bchinn/ni_diamond/ros_pkg/stacks/perception_pcl/pcl/include/pcl/pcl_base.h:216:   instantiated from here
>    /usr/include/c++/4.4/bits/stl_vector.h:435: internal compiler error: Segmentation fault
>    Please submit a full bug report,
>    with preprocessed source if appropriate.
>    See <file:///usr/share/doc/gcc-4.4/README.Bugs> for instructions.

I'm really lost here. :)

Do you have any other systems that you could try this on?

As far as I know, random internal compiler errors like this indicate bad/faulty hardware. Can you compile anything else 
for that matter? Is this a new machine or an older development machine?

Cheers,
Radu.
--
http://pointclouds.org
>
>
>
> Doh
> brian
>
>
> On Sun, Feb 13, 2011 at 6:44 PM, brian Corwin <brian.k.corwin@gmail.com <mailto:brian.k.corwin@gmail.com>> wrote:
>
>     Hi Radu,
>     I know what mercurial is and have used it in the past.  I was just confused as to why it issued the error at the
>     beginning of my make.
>
>     I am running a pentium 4 2.6ghz running ubuntu 10.10.  What version information would be useful?
>
>     I am downloading diamondback as I type this.  Hopefully that will work!
>
>     Thanks much
>     Brian
>
>
>
>     On Sun, Feb 13, 2011 at 6:40 PM, Radu Bogdan Rusu <rusu@willowgarage.com <mailto:rusu@willowgarage.com>> wrote:
>
>
>         On 02/13/2011 06:27 PM, brian Corwin wrote:
>
>             I have encountered a few things I have not seen before.  First, when I run the rosmake ~/ni --rosdep  I see
>             this at the
>             beginning
>
>             [ rosmake ] Packages requested are: ['ni']
>             [ rosmake ] Logging to directory/home/bchinn/.ros/rosmake/rosmake_output-20110213-181938
>             [ rosmake ] Expanded args ['ni'] to:
>             ['openni_pcl', 'ps_engine', 'pointcloud_to_laserscan', 'openni_pointer', 'openni_camera', 'nite', 'openni',
>             'openni_tracker', 'disparity_view', 'openni_scene']
>             [ rosmake ] Generating Install Script using rosdep then executing. This may take a minute, you will be
>             prompted for
>             permissions. . .
>             Failed to find rosdep mercurial for package openni_pcl on OS:ubuntu version:10.10
>             Failed to find rosdep mercurial for package pointcloud_to_laserscan on OS:ubuntu version:10.10
>             Failed to find rosdep mercurial for package openni_camera on OS:ubuntu version:10.10
>             Failed to find rosdep mercurial for package openni_tracker on OS:ubuntu version:10.10
>             Failed to find rosdep mercurial for package openni_scene on OS:ubuntu version:10.10
>             WARNING: Rosdeps [u'mercurial', u'mercurial', u'mercurial', u'mercurial', u'mercurial'] could not be resolved
>             executing this script:
>
>
>             What does the Failed to find rosdep mercurial warning mean?  It continues to build but that seems like a bad
>             thing.
>
>
>         Mercurial is a DVCS that we use. Not sure why rosmake would issue those warnings -- my colleagues might be able
>         to provide more information on that.
>
>         As always, in order to help you better, it's good to provide some information regarding your operating system,
>         compiler, versions, etc.
>
>
>             Second,  As NI continues to make, I run into a segfault while trying to build pcl_ros.
>
>                /opt/ros/cturtle/stacks/driver_common/dynamic_reconfigure/include/dynamic_reconfigure/server.h:73:
>             internal compiler
>             error: Segmentation fault
>
>
>         An internal compiler error is a serious one, and I cannot think of a reason why compiling dynamic_reconfigure
>         would give you that.
>
>         At this point, my recommendation for path of least pain is to install unstable or diamondback, and see if that
>         works out.
>
>
>         Cheers,
>         Radu.
>         --
>         http://pointclouds.org
>
>
>