org.graalvm.shadowed.org.antlr.v4.runtime.atn

Class ATNSimulator

java.lang.Object

org.graalvm.shadowed.org.antlr.v4.runtime.atn.ATNSimulator

Direct Known Subclasses:

LexerATNSimulator, ParserATNSimulator

public abstract class ATNSimulator
extends Object

Field Summary

Fields

Modifier and Type	Field and Description
ATN	atn
static DFAState	ERROR Must distinguish between missing edge and edge we know leads nowhere
protected PredictionContextCache	sharedContextCache The context cache maps all PredictionContext objects that are equals() to a single cached copy.

Constructor Summary

Constructors

Constructor and Description
ATNSimulator(ATN atn, PredictionContextCache sharedContextCache)

Method Summary

Modifier and Type	Method and Description
void	clearDFA() Clear the DFA cache used by the current instance.
PredictionContext	getCachedContext(PredictionContext context)
PredictionContextCache	getSharedContextCache()
abstract void	reset()

Methods inherited from class java.lang.Object

clone, equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

ERROR

public static final DFAState ERROR

Must distinguish between missing edge and edge we know leads nowhere

atn

public final ATN atn

sharedContextCache

protected final PredictionContextCache sharedContextCache

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().

This cache makes a huge difference in memory and a little bit in speed. For the Java grammar on java.*, it dropped the memory requirements at the end from 25M to 16M. We don't store any of the full context graphs in the DFA because they are limited to local context only, but apparently there's a lot of repetition there as well. We optimize the config contexts before storing the config set in the DFA states by literally rebuilding them with cached subgraphs only.

I tried a cache for use during closure operations, that was whacked after each adaptivePredict(). It cost a little bit more time I think and doesn't save on the overall footprint so it's not worth the complexity.

Constructor Detail

ATNSimulator

public ATNSimulator(ATN atn,
                    PredictionContextCache sharedContextCache)

Method Detail

reset

public abstract void reset()

clearDFA

public void clearDFA()

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.

Throws:

UnsupportedOperationException - if the current instance does not support clearing the DFA.

Since:

4.3

getSharedContextCache

public PredictionContextCache getSharedContextCache()

getCachedContext

public PredictionContext getCachedContext(PredictionContext context)