org.antlr.v4.runtime.atn.ATNConfig Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of antlr4-runtime Show documentation
Show all versions of antlr4-runtime Show documentation
The ANTLR 4 Runtime (Optimized)
/*
* Copyright (c) 2012 The ANTLR Project. All rights reserved.
* Use of this file is governed by the BSD-3-Clause license that
* can be found in the LICENSE.txt file in the project root.
*/
package org.antlr.v4.runtime.atn;
import org.antlr.v4.runtime.Recognizer;
import org.antlr.v4.runtime.misc.MurmurHash;
import org.antlr.v4.runtime.misc.NotNull;
import org.antlr.v4.runtime.misc.Nullable;
import org.antlr.v4.runtime.misc.ObjectEqualityComparator;
import java.util.ArrayDeque;
import java.util.Deque;
import java.util.IdentityHashMap;
import java.util.Map;
/**
* Represents a location with context in an ATN. The location is identified by
* the following values:
*
*
* - The current ATN state
* - The predicted alternative
* - The semantic context which must be true for this configuration to be
* enabled
* - The syntactic context, which is represented as a graph-structured stack
* whose path(s) lead to the root of the rule invocations leading to this
* state.
*
*
* In addition to these values, {@link ATNConfig} stores several properties
* about paths taken to get to the location which were added over time to help
* with performance, correctness, and/or debugging.
*
*
* - {@link #getReachesIntoOuterContext()}: Used to ensure semantic predicates
* are not evaluated in the wrong context.
* - {@link #hasPassedThroughNonGreedyDecision()}: Used for enabling
* first-match-wins instead of longest-match-wins after crossing a non-greedy
* decision.
* - {@link #getLexerActionExecutor()}: Used for tracking the lexer action(s)
* to execute should this instance be selected during lexing.
* - {@link #isPrecedenceFilterSuppressed()}: A state variable for one of the
* dynamic disambiguation strategies employed by
* {@link ParserATNSimulator#applyPrecedenceFilter}.
*
*
* Due to the use of a graph-structured stack, a single {@link ATNConfig} is
* capable of representing many individual ATN configurations which reached the
* same location in an ATN by following different paths.
*
* PERF: To conserve memory, {@link ATNConfig} is split into several
* different concrete types. {@link ATNConfig} itself stores the minimum amount
* of information typically used to define an {@link ATNConfig} instance.
* Various derived types provide additional storage space for cases where a
* non-default value is used for some of the object properties. The
* {@link ATNConfig#create} and {@link ATNConfig#transform} methods
* automatically select the smallest concrete type capable of representing the
* unique information for any given {@link ATNConfig}.
*/
public class ATNConfig {
/**
* This field stores the bit mask for implementing the
* {@link #isPrecedenceFilterSuppressed} property as a bit within the
* existing {@link #altAndOuterContextDepth} field.
*/
private static final int SUPPRESS_PRECEDENCE_FILTER = 0x80000000;
/** The ATN state associated with this configuration */
@NotNull
private final ATNState state;
/**
* This is a bit-field currently containing the following values.
*
*
* - 0x00FFFFFF: Alternative
* - 0x7F000000: Outer context depth
* - 0x80000000: Suppress precedence filter
*
*/
private int altAndOuterContextDepth;
/** The stack of invoking states leading to the rule/states associated
* with this config. We track only those contexts pushed during
* execution of the ATN simulator.
*/
@NotNull
private PredictionContext context;
protected ATNConfig(@NotNull ATNState state,
int alt,
@NotNull PredictionContext context)
{
assert (alt & 0xFFFFFF) == alt;
this.state = state;
this.altAndOuterContextDepth = alt;
this.context = context;
}
protected ATNConfig(@NotNull ATNConfig c, @NotNull ATNState state, @NotNull PredictionContext context)
{
this.state = state;
this.altAndOuterContextDepth = c.altAndOuterContextDepth;
this.context = context;
}
public static ATNConfig create(@NotNull ATNState state, int alt, @Nullable PredictionContext context) {
return create(state, alt, context, SemanticContext.NONE, null);
}
public static ATNConfig create(@NotNull ATNState state, int alt, @Nullable PredictionContext context, @NotNull SemanticContext semanticContext) {
return create(state, alt, context, semanticContext, null);
}
public static ATNConfig create(@NotNull ATNState state, int alt, @Nullable PredictionContext context, @NotNull SemanticContext semanticContext, LexerActionExecutor lexerActionExecutor) {
if (semanticContext != SemanticContext.NONE) {
if (lexerActionExecutor != null) {
return new ActionSemanticContextATNConfig(lexerActionExecutor, semanticContext, state, alt, context, false);
}
else {
return new SemanticContextATNConfig(semanticContext, state, alt, context);
}
}
else if (lexerActionExecutor != null) {
return new ActionATNConfig(lexerActionExecutor, state, alt, context, false);
}
else {
return new ATNConfig(state, alt, context);
}
}
/** Gets the ATN state associated with this configuration */
@NotNull
public final ATNState getState() {
return state;
}
/** What alt (or lexer rule) is predicted by this configuration */
public final int getAlt() {
return altAndOuterContextDepth & 0x00FFFFFF;
}
@NotNull
public final PredictionContext getContext() {
return context;
}
public void setContext(@NotNull PredictionContext context) {
this.context = context;
}
public final boolean getReachesIntoOuterContext() {
return getOuterContextDepth() != 0;
}
/**
* We cannot execute predicates dependent upon local context unless
* we know for sure we are in the correct context. Because there is
* no way to do this efficiently, we simply cannot evaluate
* dependent predicates unless we are in the rule that initially
* invokes the ATN simulator.
*
*
* closure() tracks the depth of how far we dip into the outer context:
* depth > 0. Note that it may not be totally accurate depth since I
* don't ever decrement. TODO: make it a boolean then
*/
public final int getOuterContextDepth() {
return (altAndOuterContextDepth >>> 24) & 0x7F;
}
public void setOuterContextDepth(int outerContextDepth) {
assert outerContextDepth >= 0;
// saturate at 0x7F - everything but zero/positive is only used for debug information anyway
outerContextDepth = Math.min(outerContextDepth, 0x7F);
this.altAndOuterContextDepth = (outerContextDepth << 24) | (altAndOuterContextDepth & ~0x7F000000);
}
@Nullable
public LexerActionExecutor getLexerActionExecutor() {
return null;
}
@NotNull
public SemanticContext getSemanticContext() {
return SemanticContext.NONE;
}
public boolean hasPassedThroughNonGreedyDecision() {
return false;
}
@Override
public final ATNConfig clone() {
return transform(this.getState(), false);
}
public final ATNConfig transform(@NotNull ATNState state, boolean checkNonGreedy) {
return transform(state, this.context, this.getSemanticContext(), checkNonGreedy, this.getLexerActionExecutor());
}
public final ATNConfig transform(@NotNull ATNState state, @NotNull SemanticContext semanticContext, boolean checkNonGreedy) {
return transform(state, this.context, semanticContext, checkNonGreedy, this.getLexerActionExecutor());
}
public final ATNConfig transform(@NotNull ATNState state, @Nullable PredictionContext context, boolean checkNonGreedy) {
return transform(state, context, this.getSemanticContext(), checkNonGreedy, this.getLexerActionExecutor());
}
public final ATNConfig transform(@NotNull ATNState state, LexerActionExecutor lexerActionExecutor, boolean checkNonGreedy) {
return transform(state, context, this.getSemanticContext(), checkNonGreedy, lexerActionExecutor);
}
private ATNConfig transform(@NotNull ATNState state, @Nullable PredictionContext context, @NotNull SemanticContext semanticContext, boolean checkNonGreedy, LexerActionExecutor lexerActionExecutor) {
boolean passedThroughNonGreedy = checkNonGreedy && checkNonGreedyDecision(this, state);
if (semanticContext != SemanticContext.NONE) {
if (lexerActionExecutor != null || passedThroughNonGreedy) {
return new ActionSemanticContextATNConfig(lexerActionExecutor, semanticContext, this, state, context, passedThroughNonGreedy);
}
else {
return new SemanticContextATNConfig(semanticContext, this, state, context);
}
}
else if (lexerActionExecutor != null || passedThroughNonGreedy) {
return new ActionATNConfig(lexerActionExecutor, this, state, context, passedThroughNonGreedy);
}
else {
return new ATNConfig(this, state, context);
}
}
private static boolean checkNonGreedyDecision(ATNConfig source, ATNState target) {
return source.hasPassedThroughNonGreedyDecision()
|| target instanceof DecisionState && ((DecisionState)target).nonGreedy;
}
public ATNConfig appendContext(int context, PredictionContextCache contextCache) {
PredictionContext appendedContext = getContext().appendContext(context, contextCache);
ATNConfig result = transform(getState(), appendedContext, false);
return result;
}
public ATNConfig appendContext(PredictionContext context, PredictionContextCache contextCache) {
PredictionContext appendedContext = getContext().appendContext(context, contextCache);
ATNConfig result = transform(getState(), appendedContext, false);
return result;
}
/**
* Determines if this {@link ATNConfig} fully contains another
* {@link ATNConfig}.
*
* An ATN configuration represents a position (including context) in an
* ATN during parsing. Since {@link ATNConfig} stores the context as a
* graph, a single {@link ATNConfig} instance is capable of representing
* many ATN configurations which are all in the same "location" but have
* different contexts. These {@link ATNConfig} instances are again merged
* when they are added to an {@link ATNConfigSet}. This method supports
* {@link ATNConfigSet#contains} by evaluating whether a particular
* {@link ATNConfig} contains all of the ATN configurations represented by
* another {@link ATNConfig}.
*
* An {@link ATNConfig} a contains another {@link ATNConfig}
* b if all of the following conditions are met:
*
*
* - The configurations are in the same state ({@link #getState()})
* - The configurations predict the same alternative
* ({@link #getAlt()})
* - The semantic context of a implies the semantic context of
* b (this method performs a weaker equality check)
* - Joining the prediction contexts of a and b results
* in the prediction context of a
*
*
* This method implements a conservative approximation of containment. As
* a result, when this method returns {code true} it is known that parsing
* from {@code subconfig} can only recognize a subset of the inputs which
* can be recognized starting at the current {@link ATNConfig}. However, due
* to the imprecise evaluation of implication for the semantic contexts, no
* assumptions can be made about the relationship between the configurations
* when this method returns {@code false}.
*
* @param subconfig The sub configuration.
* @return {@code true} if this configuration contains {@code subconfig};
* otherwise, {@code false}.
*/
public boolean contains(ATNConfig subconfig) {
if (this.getState().stateNumber != subconfig.getState().stateNumber
|| this.getAlt() != subconfig.getAlt()
|| !this.getSemanticContext().equals(subconfig.getSemanticContext())) {
return false;
}
Deque leftWorkList = new ArrayDeque();
Deque rightWorkList = new ArrayDeque();
leftWorkList.add(getContext());
rightWorkList.add(subconfig.getContext());
while (!leftWorkList.isEmpty()) {
PredictionContext left = leftWorkList.pop();
PredictionContext right = rightWorkList.pop();
if (left == right) {
return true;
}
if (left.size() < right.size()) {
return false;
}
if (right.isEmpty()) {
return left.hasEmpty();
} else {
for (int i = 0; i < right.size(); i++) {
int index = left.findReturnState(right.getReturnState(i));
if (index < 0) {
// assumes invokingStates has no duplicate entries
return false;
}
leftWorkList.push(left.getParent(index));
rightWorkList.push(right.getParent(i));
}
}
}
return false;
}
public final boolean isPrecedenceFilterSuppressed() {
return (altAndOuterContextDepth & SUPPRESS_PRECEDENCE_FILTER) != 0;
}
public final void setPrecedenceFilterSuppressed(boolean value) {
if (value) {
this.altAndOuterContextDepth |= SUPPRESS_PRECEDENCE_FILTER;
}
else {
this.altAndOuterContextDepth &= ~SUPPRESS_PRECEDENCE_FILTER;
}
}
/** An ATN configuration is equal to another if both have
* the same state, they predict the same alternative, and
* syntactic/semantic contexts are the same.
*/
@Override
public boolean equals(Object o) {
if (!(o instanceof ATNConfig)) {
return false;
}
return this.equals((ATNConfig)o);
}
public boolean equals(ATNConfig other) {
if (this == other) {
return true;
}
else if (other == null) {
return false;
}
return this.getState().stateNumber==other.getState().stateNumber
&& this.getAlt()==other.getAlt()
&& this.getReachesIntoOuterContext() == other.getReachesIntoOuterContext()
&& this.getContext().equals(other.getContext())
&& this.getSemanticContext().equals(other.getSemanticContext())
&& this.isPrecedenceFilterSuppressed() == other.isPrecedenceFilterSuppressed()
&& this.hasPassedThroughNonGreedyDecision() == other.hasPassedThroughNonGreedyDecision()
&& ObjectEqualityComparator.INSTANCE.equals(this.getLexerActionExecutor(), other.getLexerActionExecutor());
}
@Override
public int hashCode() {
int hashCode = MurmurHash.initialize(7);
hashCode = MurmurHash.update(hashCode, getState().stateNumber);
hashCode = MurmurHash.update(hashCode, getAlt());
hashCode = MurmurHash.update(hashCode, getReachesIntoOuterContext() ? 1 : 0);
hashCode = MurmurHash.update(hashCode, getContext());
hashCode = MurmurHash.update(hashCode, getSemanticContext());
hashCode = MurmurHash.update(hashCode, hasPassedThroughNonGreedyDecision() ? 1 : 0);
hashCode = MurmurHash.update(hashCode, getLexerActionExecutor());
hashCode = MurmurHash.finish(hashCode, 7);
return hashCode;
}
/**
* Returns a graphical representation of the current {@link ATNConfig} in
* Graphviz format. The graph can be stored to a .dot file
* and then rendered to an image using Graphviz.
*
* @return A Graphviz graph representing the current {@link ATNConfig}.
*
* @see http://www.graphviz.org/
*/
public String toDotString() {
StringBuilder builder = new StringBuilder();
builder.append("digraph G {\n");
builder.append("rankdir=LR;\n");
Map visited = new IdentityHashMap();
Deque workList = new ArrayDeque();
workList.add(getContext());
visited.put(getContext(), getContext());
while (!workList.isEmpty()) {
PredictionContext current = workList.pop();
for (int i = 0; i < current.size(); i++) {
builder.append(" s").append(System.identityHashCode(current));
builder.append("->");
builder.append("s").append(System.identityHashCode(current.getParent(i)));
builder.append("[label=\"").append(current.getReturnState(i)).append("\"];\n");
if (visited.put(current.getParent(i), current.getParent(i)) == null) {
workList.push(current.getParent(i));
}
}
}
builder.append("}\n");
return builder.toString();
}
@Override
public String toString() {
return toString(null, true, false);
}
public String toString(@Nullable Recognizer recog, boolean showAlt) {
return toString(recog, showAlt, true);
}
public String toString(@Nullable Recognizer recog, boolean showAlt, boolean showContext) {
StringBuilder buf = new StringBuilder();
// if ( state.ruleIndex>=0 ) {
// if ( recog!=null ) buf.append(recog.getRuleNames()[state.ruleIndex]+":");
// else buf.append(state.ruleIndex+":");
// }
String[] contexts;
if (showContext) {
contexts = getContext().toStrings(recog, this.getState().stateNumber);
}
else {
contexts = new String[] { "?" };
}
boolean first = true;
for (String contextDesc : contexts) {
if ( first ) {
first = false;
}
else {
buf.append(", ");
}
buf.append('(');
buf.append(getState());
if ( showAlt ) {
buf.append(",");
buf.append(getAlt());
}
if ( getContext()!=null ) {
buf.append(",");
buf.append(contextDesc);
}
if ( getSemanticContext()!=null && getSemanticContext() != SemanticContext.NONE ) {
buf.append(",");
buf.append(getSemanticContext());
}
if ( getReachesIntoOuterContext() ) {
buf.append(",up=").append(getOuterContextDepth());
}
buf.append(')');
}
return buf.toString();
}
/**
* This class was derived from {@link ATNConfig} purely as a memory
* optimization. It allows for the creation of an {@link ATNConfig} with a
* non-default semantic context.
*
* See the {@link ATNConfig} documentation for more information about
* conserving memory through the use of several concrete types.
*/
private static class SemanticContextATNConfig extends ATNConfig {
@NotNull
private final SemanticContext semanticContext;
public SemanticContextATNConfig(SemanticContext semanticContext, @NotNull ATNState state, int alt, @Nullable PredictionContext context) {
super(state, alt, context);
this.semanticContext = semanticContext;
}
public SemanticContextATNConfig(SemanticContext semanticContext, @NotNull ATNConfig c, @NotNull ATNState state, @Nullable PredictionContext context) {
super(c, state, context);
this.semanticContext = semanticContext;
}
@Override
public SemanticContext getSemanticContext() {
return semanticContext;
}
}
/**
* This class was derived from {@link ATNConfig} purely as a memory
* optimization. It allows for the creation of an {@link ATNConfig} with a
* lexer action.
*
* See the {@link ATNConfig} documentation for more information about
* conserving memory through the use of several concrete types.
*/
private static class ActionATNConfig extends ATNConfig {
private final LexerActionExecutor lexerActionExecutor;
private final boolean passedThroughNonGreedyDecision;
public ActionATNConfig(LexerActionExecutor lexerActionExecutor, @NotNull ATNState state, int alt, @Nullable PredictionContext context, boolean passedThroughNonGreedyDecision) {
super(state, alt, context);
this.lexerActionExecutor = lexerActionExecutor;
this.passedThroughNonGreedyDecision = passedThroughNonGreedyDecision;
}
protected ActionATNConfig(LexerActionExecutor lexerActionExecutor, @NotNull ATNConfig c, @NotNull ATNState state, @Nullable PredictionContext context, boolean passedThroughNonGreedyDecision) {
super(c, state, context);
if (c.getSemanticContext() != SemanticContext.NONE) {
throw new UnsupportedOperationException();
}
this.lexerActionExecutor = lexerActionExecutor;
this.passedThroughNonGreedyDecision = passedThroughNonGreedyDecision;
}
@Override
public LexerActionExecutor getLexerActionExecutor() {
return lexerActionExecutor;
}
@Override
public boolean hasPassedThroughNonGreedyDecision() {
return passedThroughNonGreedyDecision;
}
}
/**
* This class was derived from {@link SemanticContextATNConfig} purely as a memory
* optimization. It allows for the creation of an {@link ATNConfig} with
* both a lexer action and a non-default semantic context.
*
* See the {@link ATNConfig} documentation for more information about
* conserving memory through the use of several concrete types.
*/
private static class ActionSemanticContextATNConfig extends SemanticContextATNConfig {
private final LexerActionExecutor lexerActionExecutor;
private final boolean passedThroughNonGreedyDecision;
public ActionSemanticContextATNConfig(LexerActionExecutor lexerActionExecutor, @NotNull SemanticContext semanticContext, @NotNull ATNState state, int alt, @Nullable PredictionContext context, boolean passedThroughNonGreedyDecision) {
super(semanticContext, state, alt, context);
this.lexerActionExecutor = lexerActionExecutor;
this.passedThroughNonGreedyDecision = passedThroughNonGreedyDecision;
}
public ActionSemanticContextATNConfig(LexerActionExecutor lexerActionExecutor, @NotNull SemanticContext semanticContext, @NotNull ATNConfig c, @NotNull ATNState state, @Nullable PredictionContext context, boolean passedThroughNonGreedyDecision) {
super(semanticContext, c, state, context);
this.lexerActionExecutor = lexerActionExecutor;
this.passedThroughNonGreedyDecision = passedThroughNonGreedyDecision;
}
@Override
public LexerActionExecutor getLexerActionExecutor() {
return lexerActionExecutor;
}
@Override
public boolean hasPassedThroughNonGreedyDecision() {
return passedThroughNonGreedyDecision;
}
}
}