The context cache maps all PredictionContext objects that are equals()
to a single cached copy. This cache is shared across all contexts
in all ATNConfigs in all DFA states. We rebuild each ATNConfigSet
to use only cached nodes/graphs in addDFAState(). We don't want to
fill this during closure() since there are lots of contexts that
pop up but are not used ever again. It also greatly slows down closure().
Clear the DFA cache used by the current instance. Since the DFA cache may
be shared by multiple ATN simulators, this method may affect the
performance (but not accuracy) of other parsers which are being used
concurrently.