Joint move tool violates closed loop constraint

[kkevlar]

Moving non-pelvis qposes with the joint-move selection type allows violation of the closed-loop constraint of the leg because the raw qposes are changed, but none of the physics is solved or validated.

Potential solutions:

• Determine the closed-loop-join-relationship on a join-by-joint basis and fix joints manually when a transformation is performed
• Create a different version of the pdik solver which a applies a pd-controller to the joint with d->qrfc_applied and steps through physics while the solver is running

[kkevlar]

Another potential solution:

• Re-apply inverse kinematics or physics which to fix the closed loop

[kkevlar]

Less of an issue if the user performs position (green nodes) perturbations following the joint tuning. When the closed loop is broken, the nodes can be slightly dragged nearby their initial positions and the Ik solver will fix the closed loop constraint.