[ros-users] ROS Diamondback Beta 2

Ken Conley kwc at willowgarage.com
Wed Feb 9 17:35:28 UTC 2011


On Wed, Feb 9, 2011 at 2:52 AM, Felix Endres
<endres at informatik.uni-freiburg.de> wrote:
> Hi Ken,
>
> installed diamondback today via apt-get install ros-diamondback-all and
> noticed that it takes a lot of disk space. Examining the package cache
> I found that the new knowrob package had 1.2 Gig as .deb and 2.1G when
> installed, most of it in the subfolder 3rdparty.

Thanks for the report.  As Moritz's response notes, this is generally
a stack-specific issue and best addressed with the maintainer.  I
would *strongly* advise against ever installing ros-diamondback-all.
You are bringing in a bunch of stacks, both stable and experimental,
and there may be unintended side-effects as they will also bring in
thirdparty debs.  Your apache server might get changed, etc...  I
would recommend staying with the advised entry points (e.g. desktop,
desktop-full) and adding stacks from there.  'all' is only meant to be
used as a backdoor for other purposes.

As a side node: the diamondback debs are ~50% smaller than cturtle
and, if you use the advised entry-points, your install will be much
smaller as they pull in far fewer stacks than before.  We are now also
doing a better job with incremental updates, so install updates should
be much smaller.

> There seem to be two issues that can be easily addressed:
> - The package contains .svn folders in the lower levels of the fs-tree
> - it contains identical data both compressed and uncompressed
>
> Removing all .zip and .svn/ freed only about ~400M but is a quick
> improvement.
> Maybe one could remove the uncompressed data instead of the zip-files
> but I have no insight here.
>
> Maybe it's a worthwhile idea to generally exclude */.svn/* during the
> packaging?

We exclude the .svn directories in the creation of the source tarball
for a stack, but this doesn't apply to any subversion resources that a
stack's build may access.  We've recently added a prohibition on using
SVN checkouts during a build in released stacks, so this has the
unintended side-effect of accomplishing what you propose above.

thanks,
Ken

>
> Regards,
> Felix
>
>
> Am 07.02.2011 19:57, schrieb Ken Conley:
>> http://www.ros.org/news/2011/02/ros-diamondback-beta-2-release.html
>>
>> Thanks everyone for the helpful reports.  We've pushed a lot of stack
>> updates and are getting closer to our release candidate.  We
>> appreciate your continued help with integration, including the updates
>> to the installation instructions on the various Linux distributions.
>>
>> FYI: Here are some remaining OS X integration tickets based on
>> Nicholas Butko's reports, patches welcome:
>>
>> https://code.ros.org/trac/ros-pkg/ticket/4764
>> https://code.ros.org/trac/ros-pkg/ticket/4765
>> https://code.ros.org/trac/ros-pkg/ticket/4766
>>
>> We still have some fixes that have not been released yet, but will be
>> included in the next update.
>>
>>   - Ken
>> _______________________________________________
>> ros-users mailing list
>> ros-users at code.ros.org
>> https://code.ros.org/mailman/listinfo/ros-users
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>



More information about the ros-users mailing list