This allows to do 'print solver' in the python interface.
The 'dispTask' method is kept for transition purpose.

Again, for homogeneization purposes.

Use the dedicated macro to this purpose.
The error came from the fact that the method was not declared
as const. The consequences were that the following python orders
were possible:

mytask = TaskInequality ('mytask')
mytask = Task ('mytask')

The second line should crash but doesn't, because the method getClassName
called is the one of the mother class.

There is meta_task_dyn_passing_point to help in the initialization
of this task

Define MetaTaskDynLimits to handle the joint limits in the
dynamic solver. The gain is arbitrarily set.

Modified function name and command to set second order kinematics in order to uniform it to Device::setSecondOrderIntegration

Added secondOrderKinematics option with drift Jdot*dq compensation in solver-kine and userfriendly changes/commentaries in solver-op-space.cpp

Added secondOrderKinematics option with drift Jdot*dq compensation in solver-kine and userfriendly changes/commentaries in solver-op-space.cpp

This reverts commit 8eb4d5ef.

This reverts commit d16b100c.

There were 2 issues:
- The task was moved two layers down (instead of one).
- If the task to move had already the lowest priority,
it used to create a segfault. Since in this case, the
operation 'down' is useless, a guard has been added.

Directly use the corresponding macro