This was taken out to make way for the new unified processing algorithm. It's in the works and should be done by 1.0, I believe. When there's something more concrete in the spec for it we should implement it.

• Add two methods for the processing model. One which processes cues and regions, and one which just processes cues without their regions. This is easy to do, will give us more granularity in our API, and will support the VTTRegion pref on Firefox.