[Rock-dev] Imminent move to eigen3 *Please Read*

[Jakob Schwendner]

After some discussion we have decided to delay the move to eigen3 and 
sync it with our new 3-tier development process (master/next/stable). 
This will probably happen some time this week.

cheers,

Jakob

On 05/16/2011 09:50 AM, Jakob Schwendner wrote:
> Hi,
>
> please read, as this is quite likely to affect you.
>
> as you might now, we are using eigen2 quite extensively in rock. Through
> the base/types package, which has a dependency to eigen, almost all
> packages in rock have an eigen dependency.
>
> Since eigen3 (http://eigen.tuxfamily.org/index.php?title=Main_Page) is
> out for a while now, and we have some problems with eigen2 and 32 bit
> machines, we've decided to move to eigen3. I've done the conversion for
> the rock modules as well as the imoby packages.
>
> *I would like to push these changes today around noon.*
> This has a good chance of breaking your code. Let me know if you have a
> problem with me pushing today (you can always hold off updating of course).
>
> Here are the things that you might need to do to move to eigen3:
> - eigen3 is not available as a package, and is extracted into
> external/eigen3 through the autoproj rock.base configuration.
> - to make future compatbility with packages easier, the module is
> referred to as eigen3 and not external/eigen3 in the manifests
> - base/types is already pulling in eigen3, so if you have a dependency
> on it, you don't need to update your makefiles. If you use base/types
> and also explicitly include eigen2, remove the FindEigen2 part.
> - if you include eigen2 without including base, the best way is to use
> pkg-config (and an eigen3 entry in the manifest of your package) to find
> eigen3 (this avoids copying FindEigen3 cmake scripts).
> - there is a page with all the things that have changed:
> http://eigen.tuxfamily.org/dox/Eigen2ToEigen3.html
>
> Common things I had to convert and how I did it:
> - Transform3d is now Affine3d.
> - begin() and end() are now head() and tail().
> - the include is "Eigen/Core.hpp" not "eigen2/Eigen/Core.hpp". The
> latter only worked because of the standard installation directory for
> the includes.
> - It seems that eigen2 was including<iostream>  and eigen3 is not. I had
> to add a number of iostream includes because of this.
>
> The big change for us is that now in base/types/eigen.h we have
> additional to the base::Vector3d etc types base::Quaterniond and
> base::Transform3d, which do not need to be aligned and can be used in
> orogen types. This does not mean that now all the code will work on 32
> bit, but that it can be made to work. Escpecially envire is still a bit
> flaky on the 32 bit side as far as I know.
>
> Let me know if you have comments or questions.
>
> Jakob
>
> _______________________________________________
> Rock-dev mailing list
> Rock-dev at dfki.de
> http://www.dfki.de/mailman/cgi-bin/listinfo/rock-dev


-- 
Jakob Schwendner, M.Sc.
Researcher

DFKI Bremen
Robotics Innovation Center
Robert-Hooke-Straße 5
28359 Bremen, Germany

Phone: +49 (0)421 17845-4120
Fax:   +49 (0)421 17845-4150
E-Mail: jakob.schwendner at dfki.de

Weitere Informationen: http://www.dfki.de/robotik
-----------------------------------------------------------------------
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Straße 122, D-67663 Kaiserslautern
Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
(Vorsitzender) Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313
Sitz der Gesellschaft: Kaiserslautern (HRB 2313)
USt-Id.Nr.:    DE 148646973
Steuernummer:  19/673/0060/3
-----------------------------------------------------------------------