As usual, it's up to caller to cache the value if it's not used right away.

Had to use Matrix4's internal matrix to avoid circular cross-dependency between Vector3 and Matrix4 (added extra internal matrix m3 for this not used anywhere in Matrix4 itself to minimize chances of messing up something).

Added Vector3.absoluteToRelative and Vector3.relativeToAbsolute. Thanks to @G3z for the initial suggestion. Thanks to @em for the naming hint. Fixes #1043.

Matrix4.lookAt(): y doesn't need to be normalized because it's a cross product of 2 normalized vectors

Also trying to make a bit nicer UVs for ExtrudeGeometry / TextGeometry.

Default computation of morph normals for all models with morphs was taking unnecessary time and memory.

Also made it less resource intensive to recompute morph normals.

It seems to work, though not sure yet how to handle some things. Also for the moment it's implemented just for Lambert material.

As per discussion in #1170

Added a silly check on Object3D.add(). It's a quick check and it avoids enless recursivity making the browser silently hang.

Not sure about this hardcoded epsilon thing for "isZero", also related strict equality check. Maybe these should be configurable somehow.

Updated Closure compiler. Added more verbose compile warnings. Fixed all compile warnings. Started to fix runtime warnings in examples.

For verbose compile warnings to be practical I had to add externs and disable some warnings manually (there were tons of complains about "global this" and "cannot instantiate non-constructor"). Other solution would be to annotate explicitly constructors all over the code base.

(prompted by John Carmack's praise of static code analysis http://altdevblogaday.com/2011/12/24/static-code-analysis/)

Work in progress. For the moment works just with CTMLoader. Still not sure where some pieces of code would go.

Upsides:
	- initialization time cut down significantly (1.8 seconds => 0.7 seconds for CTM example)
	- much smaller GL buffers thanks to fully used indexing (Walt has 6x less vertices)
	- consequently much smaller memory footprint all around, both for main and GPU memories,
	  which means less swapping and better vertex caching

Downsides:
	- can't have anymore sharp per-face effects like flat shading (at least if mesh is not constructed in a way where each face has own vertices)
	- can't have multiple materials per mesh
	- no more easy manipulation of anything (for the moment actually no runtime manipulation at all, typed arrays are just thrown away, dynamic geometries code path not yet implemented)

Is not used so much anymore so we can have it with a less obfuscating syntax.

"Fixed typos in Matrix4's transpose() and getInverse()"

https://github.com/AddictArts/three.js/commit/62d2560b3c7b9868fc25f628c4022448aaff6b00

Also some bugfixes in ShadowMapPlugin and SpritePlugin.

This is still pretty rough, will need to clean it more.