Silvio Traversaro

Any palette is fine for me, the proposed meshcat one is great!

Probably something changed in all the removal of semantics classes in https://github.com/robotology/idyntree/pull/704 . In particular, now all the `getLinearVec3` should all return the same GeomVector3 type, see https://github.com/robotology/idyntree/blob/17d5a7db7f7190944aa6d13f96daaf1d4615946c/src/core/include/iDynTree/Core/GeomVector3.h#L38 , so...

This is indeed tricky and confusing for the users, but it should be easy to capture in a unit test. I did not remember reading this issue at the time...

Related issue: https://github.com/robotology/idyntree/issues/782 .

In any case, there is a specific reason for which you installed pybind11 via pip? In https://github.com/robotology/idyntree/issues/782, I think we can aim to support pybind11 installed via apt or the...

@GiulioRomualdi is this issue still present?

> Actually no, but I installed pybind11 from apt Ack, updated the title.

I think the problem is that that test is not deterministic, and for some configuration (really few) the threshold of consistency with COM is too small.

This happened again in https://github.com/robotology/idyntree/runs/860100190?check_suite_focus=true .

I imagine three main use cases: * You know the total robot mass, and you want to compute 3D inertia and center of mass for each link * You know...