Implement KinDynComputations::getCentroidalTotalMomentumJacobian() method

[GiulioRomualdi]

This PR introduces the getCentroidalTotalMomentumJacobian() function in KinDynComputations class.
I also wrote a simple test to spot some bugs in the method

Associated issue robotology/wb-toolbox#20

cc @traversaro @robotology/iit-dynamic-interaction-control

[GiulioRomualdi]

The test that I've developed is failing on macOS. Since I don't have an apple machine I don't know how to find the bug. Any hints?

[GiulioRomualdi]

f9df8a0 fixes the problems on macOS. The PR is now ready to be reviewed.

[traversaro]

A minor comment.

[traversaro]

Small note: the current devel branch is going to be eventually released as iDynTree 2 due to the breaking changes, but I guess it will not be released before August. If you aim for a quick release of this feature that is back-compatible, feel free to cherry-pick the changes to the master branches, so it can be released in a iDynTree 1.2 release even sooner. However if you don't have any urgency of having this change in a iDynTree released version, feel free to keep it against devel as it is.

[traversaro]

@Yeshasvitvs as I guess you were interested in this, feel free to check this PR as well!

[GiulioRomualdi]

Small note: the current devel branch is going to be eventually released as iDynTree 2 due to the breaking changes, but I guess it will not be released before August. If you aim for a quick release of this feature that is back-compatible, feel free to cherry-pick the changes to the master branches, so it can be released in a iDynTree 1.2 release even sooner. However if you don't have any urgency of having this change in a iDynTree released version, feel free to keep it against devel as it is.

For the time being, I don't need this feature. Probably @Giulero @gabrielenava or @CarlottaSartore may need it in the next future

[GiulioRomualdi]

@traversaro, regarding the CHANGELOG.md, shall I add the feature in [Unreleased Major] or [Unreleased]?

[traversaro]

@traversaro, regarding the CHANGELOG.md, shall I add the feature in [Unreleased Major] or [Unreleased]?

Use [Unreleased Major], thanks! [Unreleased] is used for the changes happening on the master branch, and to avoid conflicts anytime I forward merge changes from the master to the devel.

[traversaro]

Thanks a lot @GiulioRomualdi ! Let me know if you prefer to curate the commit yourself, or if it is ok that I merge with squash.

[GiulioRomualdi]

Merge with squash it's ok with me. Thank you!

[gabrielenava]

For the time being, I don't need this feature. Probably @Giulero @gabrielenava or @CarlottaSartore may need it in the next future

In the iRonCub team we use an homemade centroidal momentum matrix block which is actually overcomplicated for the task, but for the moment it works. Also at the moment we use iDynTree in master branch, but in case of necessity it is ok for us to migrate to devel if no big changes are required at the Simulink level.

So it is ok to wait for August for what concerns iRonCub team.

[traversaro]

Thanks @GiulioRomualdi !

[CarlottaSartore]

For the time being, I don't need this feature. Probably @Giulero @gabrielenava or @CarlottaSartore may need it in the next future

I do not see a problem for me to stay in devel rather than Master! Thank you @GiulioRomualdi