above their hip joints. First on the IK side this prevents parent joints
from being included in the accumulators in AnimInverseKinematics. Second
in AnimClip the boneLengthScale now takes into account translation and
scale on these extra parent joints.
When used in combination with the elbow pole vector, this was causing the hand to never properly reach its rotation target.
It was also having side effects causing the elbow to never straighten properly to reach its position target.
There is possibly a bug in how this tip constraint is enforced and I've verified that there are no ill effects in removing it.
With the addition of the hips IK target, the hips offset shifting code is no longer necessary.
This PR should not effect any behavior, but it removes this unused code from the animation system.
The pole vector constraint calculation within the IK system would sometimes compute the incorrect rotations.
This would be visible as an instantaneous snap of the elbow joint as the bicep was curled.
When applying pole vector constraints, there needs to be two methods of determining the current orientation of the elbow joint.
One for when the arm/elbow joint is bent, and one for when the arm/elbow is straight.
Previously, the way we would switch between these two solutions could cause a large rotation delta to accur between very small angles.
Now we use the more accurate method (1) more often, and we smoothly blend between the solutions as the joint gets straighter.
* AnimInverseKinematics: debug draw for secondary targets
* AnimInverseKienmatics: better clean up of ik target debug draw
* GeometryUtil: added findPlaneFromPoints()
* ViveControllerManager: external dependency on eigen
* ViveControllerManager: record history of left/right hand controllers
* ViveControllerManager: use history to determine user shoulder location for better calibration
* ViveControllerManager: pass defaultToReferenceMat by const ref to calibrate functions.
* CMake: added external depenency to eigen linear algebra library.
Also, A change in how the bone name to bone index lookup occurs exposed a bug
in Rig::computeAvatarBoundingCapsule(), basically it was not actually preforming IK,
and the ik targets were in the wrong coordinate frame. So when IK was actually
performed it would give bad results. This bug is now fixed.
