Memory leak in StaticOptimization and in InverseDynamics (begin())

[chrisdembia]

I was playing around with cppcheck and it reports memory leaks here and here.

[aseth1]

The InverseDynamics Analysis should be removed or at least hidden. It only remains to serve RRA. RRA should be switched to using the InverseDynamicsSolver. I gave it a go awhile back but it wasn't as straightforward as one would expect. There were something scary there that I blocked out of my mind.

Anyhow the leaks look legit. Let's see- the model's original force set (call it O ) is cloned (call it C). The analyis' _forceSet (call it A) is assigned to a set of CoordinateActuators (also heap allocated). Then it appends clones of all non muscles from C to O. Doesn't that mean O has a double count of the non muscle actuators? In the end A has coordinate actuators but should have model actuators and the coordinate actuators? I truly hope this branch is never actually executed. @aymanhab any ideas what this is supposed to be doing?