Some review work still required...

Corrected lighting system crash and VR variables not found...

Matching the same stereo rendering result given by Oculus PC SDK for
Oculus Rift CV1 is very difficult because hardware has changed a lot and
DK2 distortion shader and parameters don't fit on CV1. Some custom
parameters have been calculated to simulate kind of CV1 stereo
rendering. Further work is required on this point.

Simulator configuration is directly obtained from VR device parameters!

Oculus Rift CV1 default values seem not to be available...

...at least, directly, available if using directly audio module...

- Embedded VR distortion shader
- Ready to support multiple VR devices
- Fallback to VR Simulator if device not ready
- Support mono rendering over stereo rendering

-IN PROGRESS-

...to avoid dynamic conditions on for loop (lightsCount) on standard
shader, it seems GLSL 100 doesn't support that feature... on some GPUs
like RaspberryPi...

Stereo rendering has been moved again to Begin3dMode() and End3dMode(),
it has some limitations but makes more sense...

Still thinking about the best way to manage this...

This function needs to be redesigned or removed...

Now it's easier for the user! Just init Oculus device and get stereo
rendering!

Based on HMD parameters (IPD, ScreenSize, LesnsSeparation...)

Redesigned rlgl usage:
- rlViewport() must be called by user
- Internal projection/modelview matrices must be setup by user

Following XNA style, now we have InitGraphicsDevice(), replacing
InitDisplay() + InitGraphics()

It's more educational to go through new material system, so, I decide to
remove this function to avoid students confusion...

- Added support for RLGL_NO_STANDARD_SHADER
- Store framebuffer width and height as globals
- Reorganize rlglInit() function