Update to Manifolds v0.10 SE(n, vectors=HybridTangentRepresentation())

[Affie]

Tests on master are not passing even without these changes.

[Affie]

I'm merging this and will sort out the current broken master branch in a separate PR

[dehann]

I haven't dug into the MJL 0.10 change to the Hybrid vectors. We probably need an explicit test or the 0.10 changes.

[Affie]

For safety, all SE(n) manifolds were changed to the old behaviour (adding vectors=HybridTangentRepresentation()) so nothing changes as the old SE(n) is used throughout. The tests were passing locally in RoME and IIF, but AMP master is broken unrelated to this change.

We probably don't need vectors=HybridTangentRepresentation() everywhere, but it is a different group so has to be changed everywhere for dispatch.
As we discussed yesterday, I think the way forward is to create a "Motion(n)" (or similar name) group even if it's just an alias for SpecialEuclidean(n; vectors=HybridTangentRepresentation()) to start off with to help with this kind of problem.
We will have to be careful with depending on manifold properties/structures that is not standard as this will mean not all manifolds are supported.