Feature/optimalcontrol

[S-Dafarra]

First skeleton of a optimal control library (very skinny for the time being).

This branch is on top of improvements/spline, for which a PR is already opened #398

[traversaro]

If we plan to keep this library "unstable" (i.e. have the API of the library to change without deprecations) we should add this classes to iDynTreeExperimental, as in

idyntree/src/estimation/include/iDynTree/Estimation/BerdySparseMAPSolver.h

Line 39 in a982b71

* \ingroup iDynTreeExperimental

.
Check also the codacy comment: #399 (comment) .

[S-Dafarra]

Regarding the Codacy comment #399 (comment) I don't understand completely why it complains. Isn't it fine to use the operator ++ with an iterator?

Regarding the iDynTreeExperimental tag, is there a way to do it automatically for the whole directory?

[traversaro]

There are also a few warnings that may be worth checking: https://travis-ci.org/robotology/idyntree/jobs/312456782#L7303 .

[traversaro]

Fixed some CMake problems with the tests in 53db94c .

[traversaro]

I guess in the end you went for implementing your own integrator, we can remove this comments.

[traversaro]

Again a low priority comment, but we should follow the same scheme of the tests of the other components and also enabled Valgrind-based tests.

[traversaro]

Regarding the iDynTreeExperimental tag, is there a way to do it automatically for the whole directory?

Unfortunately no, but the basic idea is that every class belongs to a doxygen group, so eventually that part could be switched to a \ingroup iDynTreeOptimalControl.

[traversaro]

Regarding the Codacy comment #399 (comment) I don't understand completely why it complains. Isn't it fine to use the operator ++ with an iterator?**

For objects that are not integers or similarly simple objects, it is better (even is typically the difference is negligible) to use the prefix ++ in place of the postfix ++, see http://www.idinews.com/peeves/prefixIncr.html .

[traversaro]

The only failure is related to robotology/yarp#1486 , merging. The rest of open issues are stored in #400 .