TODO verify whole body trajectories

- Add new class to manage the txt file with 3 methods (read, add text to the end, delete and add new text). It is still work in progress..
- Fixed the path to OpenHRP2 (now the pos zmp and waist are created inside bin folder)
- Removed function dummy

remove an additionnal memeber with the same fuctionnality in OnLineFootTrajectoryGeneration.cpp
Fix the callMethod to handle this step height member

Increase the size of some buffer

use the same test in the SQP as for the herdt QPs algorithm

TODO use the dynamic filter on it

introduce the case of standing stil inside the PG. To be verified after full intergration inside jrl-walkgen

correct a bug in reference_t::frame_t::frame_t(const frame_t & F): where the copy where not good (old : x=x y=x!)

overload a function of OnLineFootTrajectoryGeneration to do so

adding debug code, and disable them commenting out "#define DEBUG"

prepare the online morisawa.

- The algorithm has been tested with the StackOfTask on OpenHrp3.1 on Ubuntu 12.04 precise.
- The Beam movement with morisawa's algorithm has been tested and approuved.
- correct the use of the HRP2 specialized inverse geometry
- Modify sligthly some tests

TODO
- Need to finish the implementation on the online pattern generator of Morisawa.
- Need to set the z position of the foot on morisawa to 0.

- finish the good computation for bspline without point in the middle, the other need checking and dynamic simulation to be sure.
- fix a bug in the interpolation of the COM/ZMP in herdt 2010.

Enable interpolation at the acceleration level for with the bspline.

get rid of metapod inside jrl-walkgen

test fail for the going down stairs 15cm, succeeded without dynamic filter

Implements a test doing interpolation for hands and feet placement on a stair + hand rail setup.
Uses TestHirukawa2007.cpp for metapod usage debug.

      TODO retest all the trajectories in simulation and in reality

fix bug in the com vertical trajectories