Improve VBO rendering efficiency

[SpexGuy]

by removing unnecessarily repeated driver calls

[daid]

How is this more efficient?

[SpexGuy]

The major efficiency change is on line -210/+211, where the call to glBindBuffer(GL_ARRAY_BUFFER, 0) is moved out of the loop. This removes just under half the calls to bind to GL_ARRAY_BUFFER.
For clarity, the calls to glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, ...) have been moved out of the loop. But since the loop only executes once in this case, this is not an efficiency improvement.
The calls to glEnableClientState(GL_NORMAL_ARRAY) have been moved to the top, so as to be in the same scope as glDisableClientState(GL_NORMAL_ARRAY). Again this doesn't improve efficiency, but I would argue that it does improve readability, and make the symmetry of the code more obvious.