[ros-users] dependency problems with current unstable

Ken Conley kwc at willowgarage.com
Mon Nov 7 16:53:49 UTC 2011


Background: unstable is currently broken, though that was part of the
intention.  We tried to keep unstable 'stable' all of last cycle,
which led to landing a lot of changes far too late.  Based on feedback
that unstable was not heavily used, we're instead 'breaking early,
breaking often' in order to build Fuerte from the bottom up, rather
than retrofit it in-place (how Electric was done).

Geometry 1.7 should hopefully come out this week. It was stalled by
issues found in our Oneiric update -- how to integrate with Eigen,
being one of those issues.

As for the value in releasing in unstable right now, it depends on
what your goal is.  Getting the release in now means that when things
are uncorked, they will be taken up immediately.  It also lets you use
the release tools to checkpoint your progress.  But any releases right
now will be w/o the benefit of prerelease checks if you depend on a
currently broken stack.

Vincent recommended last week setting up a stack maintainers mailing
list, which sounds like a good idea to me, and may in the future
answer your question of "How do I tell when it's time to try again?"

 - Ken


On Mon, Nov 7, 2011 at 8:25 AM, Jack O'Quin <jack.oquin at gmail.com> wrote:
> I am preparing to release camera_drivers with the new unary camera1394
> stack to ROS unstable. Running the pre-release tests shows a
> dependency problem with eigen:
>
>  http://build.willowgarage.com/view/pre-release/job/prerelease_unstable_camera1394_lucid_amd64/lastBuild/console
>
> As best I can tell, geometry 1.7.0 (which removes the eigen
> dependency) is blocked waiting on something else, leaving geometry
> 1.6.0 in unstable, which *does* depend on eigen.
>
> Given those problems, is it counter-productive for me to release the
> camera stacks? There is no point in bogging down the build systems
> with jobs that are expected to fail. How do I tell when it's time to
> try again?
> --
>  joq
> _______________________________________________
> 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