Maintains resources such as Shaders and GlyphCaches that are intended to be cached on a per-Screen basis, and provides methods that are called by BaseShaderGraphics to validate current state. The inner State class is used to encapsulate the current and previously validated state (such as texture bindings) so that the validation routines can avoid redundant state changes. There should be only one BaseShaderContext per Screen, however there may be one or more State instances per BaseShaderContext.

A note about State objects... The JOGL architecture creates a GLContext for each GLDrawable (one GLContext per GLDrawable, and one GLDrawable per onscreen window). Resources such as textures and shaders can be shared between those GLContext instances, but other state (texture bindings, scissor rect, etc) cannot be shared. Therefore we need to maintain one State instance per GLContext instance, which means there may be more than one State instance per BaseShaderContext. The currentState variable holds the current State instance corresponding to the current RenderTarget, and is revalidated as part of the updateRenderTarget() method. The ES2 backend will create a new State instance for each window, but the D3D backend is free to create a single State instance that can be shared for the entire Screen.