First commit

runtime-linux/antlr4-runtime/atn/ATN.cpp

@@ -0,0 +1,209 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/LL1Analyzer.h"
+#include "Token.h"
+#include "atn/RuleTransition.h"
+#include "misc/IntervalSet.h"
+#include "RuleContext.h"
+#include "atn/DecisionState.h"
+#include "Recognizer.h"
+#include "atn/ATNType.h"
+#include "Exceptions.h"
+#include "support/CPPUtils.h"
+
+#include "atn/ATN.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlrcpp;
+
+ATN::ATN() : ATN(ATNType::LEXER, 0) {
+}
+
+ATN::ATN(ATN &&other) {
+  // All source vectors are implicitly cleared by the moves.
+  states = std::move(other.states);
+  decisionToState = std::move(other.decisionToState);
+  ruleToStartState = std::move(other.ruleToStartState);
+  ruleToStopState = std::move(other.ruleToStopState);
+  grammarType = std::move(other.grammarType);
+  maxTokenType = std::move(other.maxTokenType);
+  ruleToTokenType = std::move(other.ruleToTokenType);
+  lexerActions = std::move(other.lexerActions);
+  modeToStartState = std::move(other.modeToStartState);
+}
+
+ATN::ATN(ATNType grammarType_, size_t maxTokenType_) : grammarType(grammarType_), maxTokenType(maxTokenType_) {
+}
+
+ATN::~ATN() {
+  for (ATNState *state : states) {
+    delete state;
+  }
+}
+
+/**
+ * Required to be defined (even though not used) as we have an explicit move assignment operator.
+ */
+ATN& ATN::operator = (ATN &other) NOEXCEPT {
+  states = other.states;
+  decisionToState = other.decisionToState;
+  ruleToStartState = other.ruleToStartState;
+  ruleToStopState = other.ruleToStopState;
+  grammarType = other.grammarType;
+  maxTokenType = other.maxTokenType;
+  ruleToTokenType = other.ruleToTokenType;
+  lexerActions = other.lexerActions;
+  modeToStartState = other.modeToStartState;
+
+  return *this;
+}
+
+/**
+ * Explicit move assignment operator to make this the preferred assignment. With implicit copy/move assignment
+ * operators it seems the copy operator is preferred causing trouble when releasing the allocated ATNState instances.
+ */
+ATN& ATN::operator = (ATN &&other) NOEXCEPT {
+  // All source vectors are implicitly cleared by the moves.
+  states = std::move(other.states);
+  decisionToState = std::move(other.decisionToState);
+  ruleToStartState = std::move(other.ruleToStartState);
+  ruleToStopState = std::move(other.ruleToStopState);
+  grammarType = std::move(other.grammarType);
+  maxTokenType = std::move(other.maxTokenType);
+  ruleToTokenType = std::move(other.ruleToTokenType);
+  lexerActions = std::move(other.lexerActions);
+  modeToStartState = std::move(other.modeToStartState);
+
+  return *this;
+}
+
+misc::IntervalSet ATN::nextTokens(ATNState *s, RuleContext *ctx) const {
+  LL1Analyzer analyzer(*this);
+  return analyzer.LOOK(s, ctx);
+
+}
+
+misc::IntervalSet const& ATN::nextTokens(ATNState *s) const {
+  if (!s->_nextTokenUpdated) {
+    std::unique_lock<std::mutex> lock { _mutex };
+    if (!s->_nextTokenUpdated) {
+      s->_nextTokenWithinRule = nextTokens(s, nullptr);
+      s->_nextTokenUpdated = true;
+    }
+  }
+  return s->_nextTokenWithinRule;
+}
+
+void ATN::addState(ATNState *state) {
+  if (state != nullptr) {
+    //state->atn = this;
+    state->stateNumber = static_cast<int>(states.size());
+  }
+
+  states.push_back(state);
+}
+
+void ATN::removeState(ATNState *state) {
+  delete states.at(state->stateNumber);// just free mem, don't shift states in list
+  states.at(state->stateNumber) = nullptr;
+}
+
+int ATN::defineDecisionState(DecisionState *s) {
+  decisionToState.push_back(s);
+  s->decision = static_cast<int>(decisionToState.size() - 1);
+  return s->decision;
+}
+
+DecisionState *ATN::getDecisionState(size_t decision) const {
+  if (!decisionToState.empty()) {
+    return decisionToState[decision];
+  }
+  return nullptr;
+}
+
+size_t ATN::getNumberOfDecisions() const {
+  return decisionToState.size();
+}
+
+misc::IntervalSet ATN::getExpectedTokens(size_t stateNumber, RuleContext *context) const {
+  if (stateNumber == ATNState::INVALID_STATE_NUMBER || stateNumber >= states.size()) {
+    throw IllegalArgumentException("Invalid state number.");
+  }
+
+  RuleContext *ctx = context;
+  ATNState *s = states.at(stateNumber);
+  misc::IntervalSet following = nextTokens(s);
+  if (!following.contains(Token::EPSILON)) {
+    return following;
+  }
+
+  misc::IntervalSet expected;
+  expected.addAll(following);
+  expected.remove(Token::EPSILON);
+  while (ctx && ctx->invokingState != ATNState::INVALID_STATE_NUMBER && following.contains(Token::EPSILON)) {
+    ATNState *invokingState = states.at(ctx->invokingState);
+    RuleTransition *rt = static_cast<RuleTransition*>(invokingState->transitions[0]);
+    following = nextTokens(rt->followState);
+    expected.addAll(following);
+    expected.remove(Token::EPSILON);
+
+    if (ctx->parent == nullptr) {
+      break;
+    }
+    ctx = static_cast<RuleContext *>(ctx->parent);
+  }
+
+  if (following.contains(Token::EPSILON)) {
+    expected.add(Token::EOF);
+  }
+
+  return expected;
+}
+
+std::string ATN::toString() const {
+  std::stringstream ss;
+  std::string type;
+  switch (grammarType) {
+    case ATNType::LEXER:
+      type = "LEXER ";
+      break;
+
+    case ATNType::PARSER:
+      type = "PARSER ";
+      break;
+
+    default:
+      break;
+  }
+  ss << "(" << type << "ATN " << std::hex << this << std::dec << ") maxTokenType: " << maxTokenType << std::endl;
+  ss << "states (" << states.size() << ") {" << std::endl;
+
+  size_t index = 0;
+  for (auto state : states) {
+    if (state == nullptr) {
+      ss << "  " << index++ << ": nul" << std::endl;
+    } else {
+      std::string text = state->toString();
+      ss << "  " << index++ << ": " << indent(text, "  ", false) << std::endl;
+    }
+  }
+
+  index = 0;
+  for (auto state : decisionToState) {
+    if (state == nullptr) {
+      ss << "  " << index++ << ": nul" << std::endl;
+    } else {
+      std::string text = state->toString();
+      ss << "  " << index++ << ": " << indent(text, "  ", false) << std::endl;
+    }
+  }
+
+  ss << "}";
+
+  return ss.str();
+}
+

runtime-linux/antlr4-runtime/atn/ATN.h

@@ -0,0 +1,112 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "RuleContext.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC ATN {
+  public:
+    static const size_t INVALID_ALT_NUMBER = 0;
+
+    /// Used for runtime deserialization of ATNs from strings.
+    ATN();
+    ATN(ATN &&other);
+    ATN(ATNType grammarType, size_t maxTokenType);
+    virtual ~ATN();
+
+    std::vector<ATNState *> states;
+
+    /// Each subrule/rule is a decision point and we must track them so we
+    /// can go back later and build DFA predictors for them.  This includes
+    /// all the rules, subrules, optional blocks, ()+, ()* etc...
+    std::vector<DecisionState *> decisionToState;
+
+    /// Maps from rule index to starting state number.
+    std::vector<RuleStartState *> ruleToStartState;
+
+    /// Maps from rule index to stop state number.
+    std::vector<RuleStopState *> ruleToStopState;
+
+    /// The type of the ATN.
+    ATNType grammarType;
+
+    /// The maximum value for any symbol recognized by a transition in the ATN.
+    size_t maxTokenType;
+
+    /// <summary>
+    /// For lexer ATNs, this maps the rule index to the resulting token type.
+    /// For parser ATNs, this maps the rule index to the generated bypass token
+    /// type if the
+    /// <seealso cref="ATNDeserializationOptions#isGenerateRuleBypassTransitions"/>
+    /// deserialization option was specified; otherwise, this is {@code null}.
+    /// </summary>
+    std::vector<size_t> ruleToTokenType;
+
+    /// For lexer ATNs, this is an array of {@link LexerAction} objects which may
+    /// be referenced by action transitions in the ATN.
+    std::vector<Ref<LexerAction>> lexerActions;
+
+    std::vector<TokensStartState *> modeToStartState;
+
+    ATN& operator = (ATN &other) NOEXCEPT;
+    ATN& operator = (ATN &&other) NOEXCEPT;
+
+    /// <summary>
+    /// Compute the set of valid tokens that can occur starting in state {@code s}.
+    ///  If {@code ctx} is null, the set of tokens will not include what can follow
+    ///  the rule surrounding {@code s}. In other words, the set will be
+    ///  restricted to tokens reachable staying within {@code s}'s rule.
+    /// </summary>
+    virtual misc::IntervalSet nextTokens(ATNState *s, RuleContext *ctx) const;
+
+    /// <summary>
+    /// Compute the set of valid tokens that can occur starting in {@code s} and
+    /// staying in same rule. <seealso cref="Token#EPSILON"/> is in set if we reach end of
+    /// rule.
+    /// </summary>
+    virtual misc::IntervalSet const& nextTokens(ATNState *s) const;
+
+    virtual void addState(ATNState *state);
+
+    virtual void removeState(ATNState *state);
+
+    virtual int defineDecisionState(DecisionState *s);
+
+    virtual DecisionState *getDecisionState(size_t decision) const;
+
+    virtual size_t getNumberOfDecisions() const;
+
+    /// <summary>
+    /// Computes the set of input symbols which could follow ATN state number
+    /// {@code stateNumber} in the specified full {@code context}. This method
+    /// considers the complete parser context, but does not evaluate semantic
+    /// predicates (i.e. all predicates encountered during the calculation are
+    /// assumed true). If a path in the ATN exists from the starting state to the
+    /// <seealso cref="RuleStopState"/> of the outermost context without matching any
+    /// symbols, <seealso cref="Token#EOF"/> is added to the returned set.
+    /// <p/>
+    /// If {@code context} is {@code null}, it is treated as
+    /// <seealso cref="ParserRuleContext#EMPTY"/>.
+    /// </summary>
+    /// <param name="stateNumber"> the ATN state number </param>
+    /// <param name="context"> the full parse context </param>
+    /// <returns> The set of potentially valid input symbols which could follow the
+    /// specified state in the specified context. </returns>
+    /// <exception cref="IllegalArgumentException"> if the ATN does not contain a state with
+    /// number {@code stateNumber} </exception>
+    virtual misc::IntervalSet getExpectedTokens(size_t stateNumber, RuleContext *context) const;
+
+    std::string toString() const;
+
+  private:
+    mutable std::mutex _mutex;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ATNConfig.cpp

@@ -0,0 +1,113 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "atn/PredictionContext.h"
+#include "SemanticContext.h"
+
+#include "atn/ATNConfig.h"
+
+using namespace antlr4::atn;
+
+const size_t ATNConfig::SUPPRESS_PRECEDENCE_FILTER = 0x40000000;
+
+ATNConfig::ATNConfig(ATNState *state_, size_t alt_, Ref<PredictionContext> const& context_)
+  : ATNConfig(state_, alt_, context_, SemanticContext::NONE) {
+}
+
+ATNConfig::ATNConfig(ATNState *state_, size_t alt_, Ref<PredictionContext> const& context_, Ref<SemanticContext> const& semanticContext_)
+  : state(state_), alt(alt_), context(context_), semanticContext(semanticContext_) {
+  reachesIntoOuterContext = 0;
+}
+
+ATNConfig::ATNConfig(Ref<ATNConfig> const& c) : ATNConfig(c, c->state, c->context, c->semanticContext) {
+}
+
+ATNConfig::ATNConfig(Ref<ATNConfig> const& c, ATNState *state_) : ATNConfig(c, state_, c->context, c->semanticContext) {
+}
+
+ATNConfig::ATNConfig(Ref<ATNConfig> const& c, ATNState *state, Ref<SemanticContext> const& semanticContext)
+  : ATNConfig(c, state, c->context, semanticContext) {
+}
+
+ATNConfig::ATNConfig(Ref<ATNConfig> const& c, Ref<SemanticContext> const& semanticContext)
+  : ATNConfig(c, c->state, c->context, semanticContext) {
+}
+
+ATNConfig::ATNConfig(Ref<ATNConfig> const& c, ATNState *state, Ref<PredictionContext> const& context)
+  : ATNConfig(c, state, context, c->semanticContext) {
+}
+
+ATNConfig::ATNConfig(Ref<ATNConfig> const& c, ATNState *state, Ref<PredictionContext> const& context,
+                     Ref<SemanticContext> const& semanticContext)
+  : state(state), alt(c->alt), context(context), reachesIntoOuterContext(c->reachesIntoOuterContext),
+    semanticContext(semanticContext) {
+}
+
+ATNConfig::~ATNConfig() {
+}
+
+size_t ATNConfig::hashCode() const {
+  size_t hashCode = misc::MurmurHash::initialize(7);
+  hashCode = misc::MurmurHash::update(hashCode, state->stateNumber);
+  hashCode = misc::MurmurHash::update(hashCode, alt);
+  hashCode = misc::MurmurHash::update(hashCode, context);
+  hashCode = misc::MurmurHash::update(hashCode, semanticContext);
+  hashCode = misc::MurmurHash::finish(hashCode, 4);
+  return hashCode;
+}
+
+size_t ATNConfig::getOuterContextDepth() const {
+  return reachesIntoOuterContext & ~SUPPRESS_PRECEDENCE_FILTER;
+}
+
+bool ATNConfig::isPrecedenceFilterSuppressed() const {
+  return (reachesIntoOuterContext & SUPPRESS_PRECEDENCE_FILTER) != 0;
+}
+
+void ATNConfig::setPrecedenceFilterSuppressed(bool value) {
+  if (value) {
+    reachesIntoOuterContext |= SUPPRESS_PRECEDENCE_FILTER;
+  } else {
+    reachesIntoOuterContext &= ~SUPPRESS_PRECEDENCE_FILTER;
+  }
+}
+
+bool ATNConfig::operator == (const ATNConfig &other) const {
+  return state->stateNumber == other.state->stateNumber && alt == other.alt &&
+    ((context == other.context) || (*context == *other.context)) &&
+    *semanticContext == *other.semanticContext &&
+    isPrecedenceFilterSuppressed() == other.isPrecedenceFilterSuppressed();
+}
+
+bool ATNConfig::operator != (const ATNConfig &other) const {
+  return !operator==(other);
+}
+
+std::string ATNConfig::toString() {
+  return toString(true);
+}
+
+std::string ATNConfig::toString(bool showAlt) {
+  std::stringstream ss;
+  ss << "(";
+
+  ss << state->toString();
+  if (showAlt) {
+    ss << "," << alt;
+  }
+  if (context) {
+    ss << ",[" << context->toString() << "]";
+  }
+  if (semanticContext != nullptr && semanticContext != SemanticContext::NONE) {
+    ss << "," << semanticContext.get();
+  }
+  if (getOuterContextDepth() > 0) {
+    ss << ",up=" << getOuterContextDepth();
+  }
+  ss << ')';
+
+  return ss.str();
+}

runtime-linux/antlr4-runtime/atn/ATNConfig.h

@@ -0,0 +1,149 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// A tuple: (ATN state, predicted alt, syntactic, semantic context).
+  ///  The syntactic context is a graph-structured stack node whose
+  ///  path(s) to the root is the rule invocation(s)
+  ///  chain used to arrive at the state.  The semantic context is
+  ///  the tree of semantic predicates encountered before reaching
+  ///  an ATN state.
+  /// </summary>
+  class ANTLR4CPP_PUBLIC ATNConfig {
+  public:
+    struct Hasher
+    {
+      size_t operator()(ATNConfig const& k) const {
+        return k.hashCode();
+      }
+    };
+
+    struct Comparer {
+      bool operator()(ATNConfig const& lhs, ATNConfig const& rhs) const {
+        return (&lhs == &rhs) || (lhs == rhs);
+      }
+    };
+
+
+    using Set = std::unordered_set<Ref<ATNConfig>, Hasher, Comparer>;
+
+    /// The ATN state associated with this configuration.
+    ATNState * state;
+
+    /// What alt (or lexer rule) is predicted by this configuration.
+    const size_t alt;
+
+    /// 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.
+    ///
+    /// Can be shared between multiple ANTConfig instances.
+    Ref<PredictionContext> context;
+
+    /**
+     * 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.
+     *
+     * <p>
+     * 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. TO_DO: make it a boolean then</p>
+     *
+     * <p>
+     * For memory efficiency, the {@link #isPrecedenceFilterSuppressed} method
+     * is also backed by this field. Since the field is publicly accessible, the
+     * highest bit which would not cause the value to become negative is used to
+     * store this field. This choice minimizes the risk that code which only
+     * compares this value to 0 would be affected by the new purpose of the
+     * flag. It also ensures the performance of the existing {@link ATNConfig}
+     * constructors as well as certain operations like
+     * {@link ATNConfigSet#add(ATNConfig, DoubleKeyMap)} method are
+     * <em>completely</em> unaffected by the change.</p>
+     */
+    size_t reachesIntoOuterContext;
+
+    /// Can be shared between multiple ATNConfig instances.
+    Ref<SemanticContext> semanticContext;
+
+    ATNConfig(ATNState *state, size_t alt, Ref<PredictionContext> const& context);
+    ATNConfig(ATNState *state, size_t alt, Ref<PredictionContext> const& context, Ref<SemanticContext> const& semanticContext);
+
+    ATNConfig(Ref<ATNConfig> const& c); // dup
+    ATNConfig(Ref<ATNConfig> const& c, ATNState *state);
+    ATNConfig(Ref<ATNConfig> const& c, ATNState *state, Ref<SemanticContext> const& semanticContext);
+    ATNConfig(Ref<ATNConfig> const& c, Ref<SemanticContext> const& semanticContext);
+    ATNConfig(Ref<ATNConfig> const& c, ATNState *state, Ref<PredictionContext> const& context);
+    ATNConfig(Ref<ATNConfig> const& c, ATNState *state, Ref<PredictionContext> const& context, Ref<SemanticContext> const& semanticContext);
+
+    ATNConfig(ATNConfig const&) = default;
+    virtual ~ATNConfig();
+    ATNConfig& operator=(ATNConfig const&) = default;
+
+    virtual size_t hashCode() const;
+
+    /**
+     * This method gets the value of the {@link #reachesIntoOuterContext} field
+     * as it existed prior to the introduction of the
+     * {@link #isPrecedenceFilterSuppressed} method.
+     */
+    size_t getOuterContextDepth() const ;
+    bool isPrecedenceFilterSuppressed() const;
+    void setPrecedenceFilterSuppressed(bool value);
+
+    /// 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.
+    bool operator == (const ATNConfig &other) const;
+    bool operator != (const ATNConfig &other) const;
+
+    virtual std::string toString();
+    std::string toString(bool showAlt);
+
+  private:
+    /**
+     * This field stores the bit mask for implementing the
+     * {@link #isPrecedenceFilterSuppressed} property as a bit within the
+     * existing {@link #reachesIntoOuterContext} field.
+     */
+    static const size_t SUPPRESS_PRECEDENCE_FILTER;
+  };
+
+} // namespace atn
+} // namespace antlr4
+
+
+// Hash function for ATNConfig.
+
+namespace std {
+  using antlr4::atn::ATNConfig;
+
+  template <> struct hash<ATNConfig>
+  {
+    size_t operator() (const ATNConfig &x) const
+    {
+      return x.hashCode();
+    }
+  };
+
+  template <> struct hash<std::vector<Ref<ATNConfig>>>
+  {
+    size_t operator() (const std::vector<Ref<ATNConfig>> &vector) const
+    {
+      std::size_t seed = 0;
+      for (auto &config : vector) {
+        seed ^= config->hashCode() + 0x9e3779b9 + (seed << 6) + (seed >> 2);
+      }
+      return seed;
+    }
+  };
+}

runtime-linux/antlr4-runtime/atn/ATNConfigSet.cpp

@@ -0,0 +1,228 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/PredictionContext.h"
+#include "atn/ATNConfig.h"
+#include "atn/ATNSimulator.h"
+#include "Exceptions.h"
+#include "atn/SemanticContext.h"
+#include "support/Arrays.h"
+
+#include "atn/ATNConfigSet.h"
+
+using namespace antlr4::atn;
+using namespace antlrcpp;
+
+ATNConfigSet::ATNConfigSet(bool fullCtx) : fullCtx(fullCtx) {
+  InitializeInstanceFields();
+}
+
+ATNConfigSet::ATNConfigSet(const Ref<ATNConfigSet> &old) : ATNConfigSet(old->fullCtx) {
+  addAll(old);
+  uniqueAlt = old->uniqueAlt;
+  conflictingAlts = old->conflictingAlts;
+  hasSemanticContext = old->hasSemanticContext;
+  dipsIntoOuterContext = old->dipsIntoOuterContext;
+}
+
+ATNConfigSet::~ATNConfigSet() {
+}
+
+bool ATNConfigSet::add(const Ref<ATNConfig> &config) {
+  return add(config, nullptr);
+}
+
+bool ATNConfigSet::add(const Ref<ATNConfig> &config, PredictionContextMergeCache *mergeCache) {
+  if (_readonly) {
+    throw IllegalStateException("This set is readonly");
+  }
+  if (config->semanticContext != SemanticContext::NONE) {
+    hasSemanticContext = true;
+  }
+  if (config->getOuterContextDepth() > 0) {
+    dipsIntoOuterContext = true;
+  }
+
+  size_t hash = getHash(config.get());
+  ATNConfig *existing = _configLookup[hash];
+  if (existing == nullptr) {
+    _configLookup[hash] = config.get();
+    _cachedHashCode = 0;
+    configs.push_back(config); // track order here
+
+    return true;
+  }
+
+  // a previous (s,i,pi,_), merge with it and save result
+  bool rootIsWildcard = !fullCtx;
+  Ref<PredictionContext> merged = PredictionContext::merge(existing->context, config->context, rootIsWildcard, mergeCache);
+  // no need to check for existing.context, config.context in cache
+  // since only way to create new graphs is "call rule" and here. We
+  // cache at both places.
+  existing->reachesIntoOuterContext = std::max(existing->reachesIntoOuterContext, config->reachesIntoOuterContext);
+
+  // make sure to preserve the precedence filter suppression during the merge
+  if (config->isPrecedenceFilterSuppressed()) {
+    existing->setPrecedenceFilterSuppressed(true);
+  }
+
+  existing->context = merged; // replace context; no need to alt mapping
+
+  return true;
+}
+
+bool ATNConfigSet::addAll(const Ref<ATNConfigSet> &other) {
+  for (auto &c : other->configs) {
+    add(c);
+  }
+  return false;
+}
+
+std::vector<ATNState*> ATNConfigSet::getStates() {
+  std::vector<ATNState*> states;
+  for (auto c : configs) {
+    states.push_back(c->state);
+  }
+  return states;
+}
+
+/**
+ * Gets the complete set of represented alternatives for the configuration
+ * set.
+ *
+ * @return the set of represented alternatives in this configuration set
+ *
+ * @since 4.3
+ */
+
+BitSet ATNConfigSet::getAlts() {
+  BitSet alts;
+  for (ATNConfig config : configs) {
+    alts.set(config.alt);
+  }
+  return alts;
+}
+
+std::vector<Ref<SemanticContext>> ATNConfigSet::getPredicates() {
+  std::vector<Ref<SemanticContext>> preds;
+  for (auto c : configs) {
+    if (c->semanticContext != SemanticContext::NONE) {
+      preds.push_back(c->semanticContext);
+    }
+  }
+  return preds;
+}
+
+Ref<ATNConfig> ATNConfigSet::get(size_t i) const {
+  return configs[i];
+}
+
+void ATNConfigSet::optimizeConfigs(ATNSimulator *interpreter) {
+  if (_readonly) {
+    throw IllegalStateException("This set is readonly");
+  }
+  if (_configLookup.empty())
+    return;
+
+  for (auto &config : configs) {
+    config->context = interpreter->getCachedContext(config->context);
+  }
+}
+
+bool ATNConfigSet::operator == (const ATNConfigSet &other) {
+  if (&other == this) {
+    return true;
+  }
+
+  if (configs.size() != other.configs.size())
+    return false;
+
+  if (fullCtx != other.fullCtx || uniqueAlt != other.uniqueAlt ||
+      conflictingAlts != other.conflictingAlts || hasSemanticContext != other.hasSemanticContext ||
+      dipsIntoOuterContext != other.dipsIntoOuterContext) // includes stack context
+    return false;
+
+  return Arrays::equals(configs, other.configs);
+}
+
+size_t ATNConfigSet::hashCode() {
+  if (!isReadonly() || _cachedHashCode == 0) {
+    _cachedHashCode = 1;
+    for (auto &i : configs) {
+      _cachedHashCode = 31 * _cachedHashCode + i->hashCode(); // Same as Java's list hashCode impl.
+    }
+  }
+
+  return _cachedHashCode;
+}
+
+size_t ATNConfigSet::size() {
+  return configs.size();
+}
+
+bool ATNConfigSet::isEmpty() {
+  return configs.empty();
+}
+
+void ATNConfigSet::clear() {
+  if (_readonly) {
+    throw IllegalStateException("This set is readonly");
+  }
+  configs.clear();
+  _cachedHashCode = 0;
+  _configLookup.clear();
+}
+
+bool ATNConfigSet::isReadonly() {
+  return _readonly;
+}
+
+void ATNConfigSet::setReadonly(bool readonly) {
+  _readonly = readonly;
+  _configLookup.clear();
+}
+
+std::string ATNConfigSet::toString() {
+  std::stringstream ss;
+  ss << "[";
+  for (size_t i = 0; i < configs.size(); i++) {
+    ss << configs[i]->toString();
+  }
+  ss << "]";
+
+  if (hasSemanticContext) {
+    ss << ",hasSemanticContext = " <<  hasSemanticContext;
+  }
+  if (uniqueAlt != ATN::INVALID_ALT_NUMBER) {
+    ss << ",uniqueAlt = " << uniqueAlt;
+  }
+
+  if (conflictingAlts.size() > 0) {
+    ss << ",conflictingAlts = ";
+    ss << conflictingAlts.toString();
+  }
+
+  if (dipsIntoOuterContext) {
+    ss << ", dipsIntoOuterContext";
+  }
+  return ss.str();
+}
+
+size_t ATNConfigSet::getHash(ATNConfig *c) {
+  size_t hashCode = 7;
+  hashCode = 31 * hashCode + c->state->stateNumber;
+  hashCode = 31 * hashCode + c->alt;
+  hashCode = 31 * hashCode + c->semanticContext->hashCode();
+  return hashCode;
+}
+
+void ATNConfigSet::InitializeInstanceFields() {
+  uniqueAlt = 0;
+  hasSemanticContext = false;
+  dipsIntoOuterContext = false;
+
+  _readonly = false;
+  _cachedHashCode = 0;
+}

runtime-linux/antlr4-runtime/atn/ATNConfigSet.h

@@ -0,0 +1,110 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "support/BitSet.h"
+#include "atn/PredictionContext.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Specialized set that can track info about the set, with support for combining similar configurations using a
+  /// graph-structured stack.
+  class ANTLR4CPP_PUBLIC ATNConfigSet {
+  public:
+    /// Track the elements as they are added to the set; supports get(i)
+    std::vector<Ref<ATNConfig>> configs;
+
+    // TO_DO: these fields make me pretty uncomfortable but nice to pack up info together, saves recomputation
+    // TO_DO: can we track conflicts as they are added to save scanning configs later?
+    size_t uniqueAlt;
+
+    /** Currently this is only used when we detect SLL conflict; this does
+     *  not necessarily represent the ambiguous alternatives. In fact,
+     *  I should also point out that this seems to include predicated alternatives
+     *  that have predicates that evaluate to false. Computed in computeTargetState().
+     */
+    antlrcpp::BitSet conflictingAlts;
+
+    // Used in parser and lexer. In lexer, it indicates we hit a pred
+    // while computing a closure operation.  Don't make a DFA state from this.
+    bool hasSemanticContext;
+    bool dipsIntoOuterContext;
+
+    /// Indicates that this configuration set is part of a full context
+    /// LL prediction. It will be used to determine how to merge $. With SLL
+    /// it's a wildcard whereas it is not for LL context merge.
+    const bool fullCtx;
+
+    ATNConfigSet(bool fullCtx = true);
+    ATNConfigSet(const Ref<ATNConfigSet> &old);
+
+    virtual ~ATNConfigSet();
+
+    virtual bool add(const Ref<ATNConfig> &config);
+
+    /// <summary>
+    /// Adding a new config means merging contexts with existing configs for
+    /// {@code (s, i, pi, _)}, where {@code s} is the
+    /// <seealso cref="ATNConfig#state"/>, {@code i} is the <seealso cref="ATNConfig#alt"/>, and
+    /// {@code pi} is the <seealso cref="ATNConfig#semanticContext"/>. We use
+    /// {@code (s,i,pi)} as key.
+    /// <p/>
+    /// This method updates <seealso cref="#dipsIntoOuterContext"/> and
+    /// <seealso cref="#hasSemanticContext"/> when necessary.
+    /// </summary>
+    virtual bool add(const Ref<ATNConfig> &config, PredictionContextMergeCache *mergeCache);
+
+    virtual std::vector<ATNState *> getStates();
+
+    /**
+     * Gets the complete set of represented alternatives for the configuration
+     * set.
+     *
+     * @return the set of represented alternatives in this configuration set
+     *
+     * @since 4.3
+     */
+    antlrcpp::BitSet getAlts();
+    virtual std::vector<Ref<SemanticContext>> getPredicates();
+
+    virtual Ref<ATNConfig> get(size_t i) const;
+
+    virtual void optimizeConfigs(ATNSimulator *interpreter);
+
+    bool addAll(const Ref<ATNConfigSet> &other);
+
+    bool operator == (const ATNConfigSet &other);
+    virtual size_t hashCode();
+    virtual size_t size();
+    virtual bool isEmpty();
+    virtual void clear();
+    virtual bool isReadonly();
+    virtual void setReadonly(bool readonly);
+    virtual std::string toString();
+
+  protected:
+    /// Indicates that the set of configurations is read-only. Do not
+    /// allow any code to manipulate the set; DFA states will point at
+    /// the sets and they must not change. This does not protect the other
+    /// fields; in particular, conflictingAlts is set after
+    /// we've made this readonly.
+    bool _readonly;
+
+    virtual size_t getHash(ATNConfig *c); // Hash differs depending on set type.
+
+  private:
+    size_t _cachedHashCode;
+
+    /// All configs but hashed by (s, i, _, pi) not including context. Wiped out
+    /// when we go readonly as this set becomes a DFA state.
+    std::unordered_map<size_t, ATNConfig *> _configLookup;
+
+    void InitializeInstanceFields();
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ATNDeserializationOptions.cpp

@@ -0,0 +1,64 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ATNDeserializationOptions.h"
+
+using namespace antlr4::atn;
+
+ATNDeserializationOptions ATNDeserializationOptions::defaultOptions;
+
+ATNDeserializationOptions::ATNDeserializationOptions() {
+  InitializeInstanceFields();
+}
+
+ATNDeserializationOptions::ATNDeserializationOptions(ATNDeserializationOptions *options) : ATNDeserializationOptions() {
+  this->verifyATN = options->verifyATN;
+  this->generateRuleBypassTransitions = options->generateRuleBypassTransitions;
+}
+
+ATNDeserializationOptions::~ATNDeserializationOptions() {
+}
+
+const ATNDeserializationOptions& ATNDeserializationOptions::getDefaultOptions() {
+  return defaultOptions;
+}
+
+bool ATNDeserializationOptions::isReadOnly() {
+  return readOnly;
+}
+
+void ATNDeserializationOptions::makeReadOnly() {
+  readOnly = true;
+}
+
+bool ATNDeserializationOptions::isVerifyATN() {
+  return verifyATN;
+}
+
+void ATNDeserializationOptions::setVerifyATN(bool verify) {
+  throwIfReadOnly();
+  verifyATN = verify;
+}
+
+bool ATNDeserializationOptions::isGenerateRuleBypassTransitions() {
+  return generateRuleBypassTransitions;
+}
+
+void ATNDeserializationOptions::setGenerateRuleBypassTransitions(bool generate) {
+  throwIfReadOnly();
+  generateRuleBypassTransitions = generate;
+}
+
+void ATNDeserializationOptions::throwIfReadOnly() {
+  if (isReadOnly()) {
+    throw "The object is read only.";
+  }
+}
+
+void ATNDeserializationOptions::InitializeInstanceFields() {
+  readOnly = false;
+  verifyATN = true;
+  generateRuleBypassTransitions = false;
+}

runtime-linux/antlr4-runtime/atn/ATNDeserializationOptions.h

@@ -0,0 +1,50 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "antlr4-common.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC ATNDeserializationOptions {
+  private:
+    static ATNDeserializationOptions defaultOptions;
+
+    bool readOnly;
+    bool verifyATN;
+    bool generateRuleBypassTransitions;
+
+  public:
+    ATNDeserializationOptions();
+    ATNDeserializationOptions(ATNDeserializationOptions *options);
+    ATNDeserializationOptions(ATNDeserializationOptions const&) = default;
+    virtual ~ATNDeserializationOptions();
+    ATNDeserializationOptions& operator=(ATNDeserializationOptions const&) = default;
+
+    static const ATNDeserializationOptions& getDefaultOptions();
+
+    bool isReadOnly();
+
+    void makeReadOnly();
+
+    bool isVerifyATN();
+
+    void setVerifyATN(bool verify);
+
+    bool isGenerateRuleBypassTransitions();
+
+    void setGenerateRuleBypassTransitions(bool generate);
+
+  protected:
+    virtual void throwIfReadOnly();
+
+  private:
+    void InitializeInstanceFields();
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ATNDeserializer.cpp

@@ -0,0 +1,758 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ATNDeserializationOptions.h"
+
+#include "atn/ATNType.h"
+#include "atn/ATNState.h"
+#include "atn/ATN.h"
+
+#include "atn/LoopEndState.h"
+#include "atn/DecisionState.h"
+#include "atn/RuleStartState.h"
+#include "atn/RuleStopState.h"
+#include "atn/TokensStartState.h"
+#include "atn/RuleTransition.h"
+#include "atn/EpsilonTransition.h"
+#include "atn/PlusLoopbackState.h"
+#include "atn/PlusBlockStartState.h"
+#include "atn/StarLoopbackState.h"
+#include "atn/BasicBlockStartState.h"
+#include "atn/BasicState.h"
+#include "atn/BlockEndState.h"
+#include "atn/StarLoopEntryState.h"
+
+#include "atn/AtomTransition.h"
+#include "atn/StarBlockStartState.h"
+#include "atn/RangeTransition.h"
+#include "atn/PredicateTransition.h"
+#include "atn/PrecedencePredicateTransition.h"
+#include "atn/ActionTransition.h"
+#include "atn/SetTransition.h"
+#include "atn/NotSetTransition.h"
+#include "atn/WildcardTransition.h"
+#include "Token.h"
+
+#include "misc/IntervalSet.h"
+#include "Exceptions.h"
+#include "support/CPPUtils.h"
+#include "support/StringUtils.h"
+
+#include "atn/LexerCustomAction.h"
+#include "atn/LexerChannelAction.h"
+#include "atn/LexerModeAction.h"
+#include "atn/LexerMoreAction.h"
+#include "atn/LexerPopModeAction.h"
+#include "atn/LexerPushModeAction.h"
+#include "atn/LexerSkipAction.h"
+#include "atn/LexerTypeAction.h"
+
+#include "atn/ATNDeserializer.h"
+
+#include <string>
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlrcpp;
+
+const size_t ATNDeserializer::SERIALIZED_VERSION = 3;
+
+namespace {
+
+uint32_t deserializeInt32(const std::vector<uint16_t>& data, size_t offset) {
+  return (uint32_t)data[offset] | ((uint32_t)data[offset + 1] << 16);
+}
+
+ssize_t readUnicodeInt(const std::vector<uint16_t>& data, int& p) {
+  return static_cast<ssize_t>(data[p++]);
+}
+
+ssize_t readUnicodeInt32(const std::vector<uint16_t>& data, int& p) {
+  auto result = deserializeInt32(data, p);
+  p += 2;
+  return static_cast<ssize_t>(result);
+}
+
+// We templatize this on the function type so the optimizer can inline
+// the 16- or 32-bit readUnicodeInt/readUnicodeInt32 as needed.
+template <typename F>
+void deserializeSets(
+  const std::vector<uint16_t>& data,
+  int& p,
+  std::vector<misc::IntervalSet>& sets,
+  F readUnicode) {
+  int nsets = data[p++];
+  for (int i = 0; i < nsets; i++) {
+    int nintervals = data[p++];
+    misc::IntervalSet set;
+
+    bool containsEof = data[p++] != 0;
+    if (containsEof) {
+      set.add(-1);
+    }
+
+    for (int j = 0; j < nintervals; j++) {
+      auto a = readUnicode(data, p);
+      auto b = readUnicode(data, p);
+      set.add(a, b);
+    }
+    sets.push_back(set);
+  }
+}
+
+}
+
+ATNDeserializer::ATNDeserializer(): ATNDeserializer(ATNDeserializationOptions::getDefaultOptions()) {
+}
+
+ATNDeserializer::ATNDeserializer(const ATNDeserializationOptions& dso): deserializationOptions(dso) {
+}
+
+ATNDeserializer::~ATNDeserializer() {
+}
+
+/**
+ * This value should never change. Updates following this version are
+ * reflected as change in the unique ID SERIALIZED_UUID.
+ */
+Guid ATNDeserializer::ADDED_PRECEDENCE_TRANSITIONS() {
+  return Guid("1DA0C57D-6C06-438A-9B27-10BCB3CE0F61");
+}
+
+Guid ATNDeserializer::ADDED_LEXER_ACTIONS() {
+  return Guid("AADB8D7E-AEEF-4415-AD2B-8204D6CF042E");
+}
+
+Guid ATNDeserializer::ADDED_UNICODE_SMP() {
+  return Guid("59627784-3BE5-417A-B9EB-8131A7286089");
+}
+
+Guid ATNDeserializer::SERIALIZED_UUID() {
+  return ADDED_UNICODE_SMP();
+}
+
+Guid ATNDeserializer::BASE_SERIALIZED_UUID() {
+  return Guid("33761B2D-78BB-4A43-8B0B-4F5BEE8AACF3");
+}
+
+std::vector<Guid>& ATNDeserializer::SUPPORTED_UUIDS() {
+  static std::vector<Guid> singleton = { BASE_SERIALIZED_UUID(), ADDED_PRECEDENCE_TRANSITIONS(), ADDED_LEXER_ACTIONS(), ADDED_UNICODE_SMP() };
+  return singleton;
+}
+
+bool ATNDeserializer::isFeatureSupported(const Guid &feature, const Guid &actualUuid) {
+  auto featureIterator = std::find(SUPPORTED_UUIDS().begin(), SUPPORTED_UUIDS().end(), feature);
+  if (featureIterator == SUPPORTED_UUIDS().end()) {
+    return false;
+  }
+  auto actualIterator = std::find(SUPPORTED_UUIDS().begin(), SUPPORTED_UUIDS().end(), actualUuid);
+  if (actualIterator == SUPPORTED_UUIDS().end()) {
+    return false;
+  }
+
+  return std::distance(featureIterator, actualIterator) >= 0;
+}
+
+ATN ATNDeserializer::deserialize(const std::vector<uint16_t>& input) {
+  // Don't adjust the first value since that's the version number.
+  std::vector<uint16_t> data(input.size());
+  data[0] = input[0];
+  for (size_t i = 1; i < input.size(); ++i) {
+    data[i] = input[i] - 2;
+  }
+
+  int p = 0;
+  int version = data[p++];
+  if (version != SERIALIZED_VERSION) {
+    std::string reason = "Could not deserialize ATN with version" + std::to_string(version) + "(expected " + std::to_string(SERIALIZED_VERSION) + ").";
+
+    throw UnsupportedOperationException(reason);
+  }
+
+  Guid uuid = toUUID(data.data(), p);
+  p += 8;
+  auto uuidIterator = std::find(SUPPORTED_UUIDS().begin(), SUPPORTED_UUIDS().end(), uuid);
+  if (uuidIterator == SUPPORTED_UUIDS().end()) {
+    std::string reason = "Could not deserialize ATN with UUID " + uuid.toString() + " (expected " +
+      SERIALIZED_UUID().toString() + " or a legacy UUID).";
+
+    throw UnsupportedOperationException(reason);
+  }
+
+  bool supportsPrecedencePredicates = isFeatureSupported(ADDED_PRECEDENCE_TRANSITIONS(), uuid);
+  bool supportsLexerActions = isFeatureSupported(ADDED_LEXER_ACTIONS(), uuid);
+
+  ATNType grammarType = (ATNType)data[p++];
+  size_t maxTokenType = data[p++];
+  ATN atn(grammarType, maxTokenType);
+
+  //
+  // STATES
+  //
+  std::vector<std::pair<LoopEndState*, size_t>> loopBackStateNumbers;
+  std::vector<std::pair<BlockStartState*, size_t>> endStateNumbers;
+  size_t nstates = data[p++];
+  for (size_t i = 0; i < nstates; i++) {
+    size_t stype = data[p++];
+    // ignore bad type of states
+    if (stype == ATNState::ATN_INVALID_TYPE) {
+      atn.addState(nullptr);
+      continue;
+    }
+
+    size_t ruleIndex = data[p++];
+    if (ruleIndex == 0xFFFF) {
+      ruleIndex = INVALID_INDEX;
+    }
+
+    ATNState *s = stateFactory(stype, ruleIndex);
+    if (stype == ATNState::LOOP_END) { // special case
+      int loopBackStateNumber = data[p++];
+      loopBackStateNumbers.push_back({ (LoopEndState*)s,  loopBackStateNumber });
+    } else if (is<BlockStartState*>(s)) {
+      int endStateNumber = data[p++];
+      endStateNumbers.push_back({ (BlockStartState*)s, endStateNumber });
+    }
+    atn.addState(s);
+  }
+
+  // delay the assignment of loop back and end states until we know all the state instances have been initialized
+  for (auto &pair : loopBackStateNumbers) {
+    pair.first->loopBackState = atn.states[pair.second];
+  }
+
+  for (auto &pair : endStateNumbers) {
+    pair.first->endState = (BlockEndState*)atn.states[pair.second];
+  }
+
+  size_t numNonGreedyStates = data[p++];
+  for (size_t i = 0; i < numNonGreedyStates; i++) {
+    size_t stateNumber = data[p++];
+    // The serialized ATN must be specifying the right states, so that the
+    // cast below is correct.
+    ((DecisionState *)atn.states[stateNumber])->nonGreedy = true;
+  }
+
+  if (supportsPrecedencePredicates) {
+    size_t numPrecedenceStates = data[p++];
+    for (size_t i = 0; i < numPrecedenceStates; i++) {
+      size_t stateNumber = data[p++];
+      ((RuleStartState *)atn.states[stateNumber])->isLeftRecursiveRule = true;
+    }
+  }
+
+  //
+  // RULES
+  //
+  size_t nrules = data[p++];
+  for (size_t i = 0; i < nrules; i++) {
+    size_t s = data[p++];
+    // Also here, the serialized atn must ensure to point to the correct class type.
+    RuleStartState *startState = (RuleStartState*)atn.states[s];
+    atn.ruleToStartState.push_back(startState);
+    if (atn.grammarType == ATNType::LEXER) {
+      size_t tokenType = data[p++];
+      if (tokenType == 0xFFFF) {
+        tokenType = Token::EOF;
+      }
+
+      atn.ruleToTokenType.push_back(tokenType);
+
+      if (!isFeatureSupported(ADDED_LEXER_ACTIONS(), uuid)) {
+        // this piece of unused metadata was serialized prior to the
+        // addition of LexerAction
+        //int actionIndexIgnored = data[p++];
+        p++;
+      }
+    }
+  }
+
+  atn.ruleToStopState.resize(nrules);
+  for (ATNState *state : atn.states) {
+    if (!is<RuleStopState*>(state)) {
+      continue;
+    }
+
+    RuleStopState *stopState = static_cast<RuleStopState*>(state);
+    atn.ruleToStopState[state->ruleIndex] = stopState;
+    atn.ruleToStartState[state->ruleIndex]->stopState = stopState;
+  }
+
+  //
+  // MODES
+  //
+  size_t nmodes = data[p++];
+  for (size_t i = 0; i < nmodes; i++) {
+    size_t s = data[p++];
+    atn.modeToStartState.push_back(static_cast<TokensStartState*>(atn.states[s]));
+  }
+
+  //
+  // SETS
+  //
+  std::vector<misc::IntervalSet> sets;
+
+  // First, deserialize sets with 16-bit arguments <= U+FFFF.
+  deserializeSets(data, p, sets, readUnicodeInt);
+
+  // Next, if the ATN was serialized with the Unicode SMP feature,
+  // deserialize sets with 32-bit arguments <= U+10FFFF.
+  if (isFeatureSupported(ADDED_UNICODE_SMP(), uuid)) {
+    deserializeSets(data, p, sets, readUnicodeInt32);
+  }
+
+  //
+  // EDGES
+  //
+  int nedges = data[p++];
+  for (int i = 0; i < nedges; i++) {
+    size_t src = data[p];
+    size_t trg = data[p + 1];
+    size_t ttype = data[p + 2];
+    size_t arg1 = data[p + 3];
+    size_t arg2 = data[p + 4];
+    size_t arg3 = data[p + 5];
+    Transition *trans = edgeFactory(atn, ttype, src, trg, arg1, arg2, arg3, sets);
+    ATNState *srcState = atn.states[src];
+    srcState->addTransition(trans);
+    p += 6;
+  }
+
+  // edges for rule stop states can be derived, so they aren't serialized
+  for (ATNState *state : atn.states) {
+    for (size_t i = 0; i < state->transitions.size(); i++) {
+      Transition *t = state->transitions[i];
+      if (!is<RuleTransition*>(t)) {
+        continue;
+      }
+
+      RuleTransition *ruleTransition = static_cast<RuleTransition*>(t);
+      size_t outermostPrecedenceReturn = INVALID_INDEX;
+      if (atn.ruleToStartState[ruleTransition->target->ruleIndex]->isLeftRecursiveRule) {
+        if (ruleTransition->precedence == 0) {
+          outermostPrecedenceReturn = ruleTransition->target->ruleIndex;
+        }
+      }
+
+      EpsilonTransition *returnTransition = new EpsilonTransition(ruleTransition->followState, outermostPrecedenceReturn); /* mem check: freed in ANTState d-tor */
+      atn.ruleToStopState[ruleTransition->target->ruleIndex]->addTransition(returnTransition);
+    }
+  }
+
+  for (ATNState *state : atn.states) {
+    if (is<BlockStartState *>(state)) {
+      BlockStartState *startState = static_cast<BlockStartState *>(state);
+
+      // we need to know the end state to set its start state
+      if (startState->endState == nullptr) {
+        throw IllegalStateException();
+      }
+
+      // block end states can only be associated to a single block start state
+      if (startState->endState->startState != nullptr) {
+        throw IllegalStateException();
+      }
+
+      startState->endState->startState = static_cast<BlockStartState*>(state);
+    }
+
+    if (is<PlusLoopbackState*>(state)) {
+      PlusLoopbackState *loopbackState = static_cast<PlusLoopbackState *>(state);
+      for (size_t i = 0; i < loopbackState->transitions.size(); i++) {
+        ATNState *target = loopbackState->transitions[i]->target;
+        if (is<PlusBlockStartState *>(target)) {
+          (static_cast<PlusBlockStartState *>(target))->loopBackState = loopbackState;
+        }
+      }
+    } else if (is<StarLoopbackState *>(state)) {
+      StarLoopbackState *loopbackState = static_cast<StarLoopbackState *>(state);
+      for (size_t i = 0; i < loopbackState->transitions.size(); i++) {
+        ATNState *target = loopbackState->transitions[i]->target;
+        if (is<StarLoopEntryState *>(target)) {
+          (static_cast<StarLoopEntryState*>(target))->loopBackState = loopbackState;
+        }
+      }
+    }
+  }
+
+  //
+  // DECISIONS
+  //
+  size_t ndecisions = data[p++];
+  for (size_t i = 1; i <= ndecisions; i++) {
+    size_t s = data[p++];
+    DecisionState *decState = dynamic_cast<DecisionState*>(atn.states[s]);
+    if (decState == nullptr)
+      throw IllegalStateException();
+
+    atn.decisionToState.push_back(decState);
+    decState->decision = (int)i - 1;
+  }
+
+  //
+  // LEXER ACTIONS
+  //
+  if (atn.grammarType == ATNType::LEXER) {
+    if (supportsLexerActions) {
+      atn.lexerActions.resize(data[p++]);
+      for (size_t i = 0; i < atn.lexerActions.size(); i++) {
+        LexerActionType actionType = (LexerActionType)data[p++];
+        int data1 = data[p++];
+        if (data1 == 0xFFFF) {
+          data1 = -1;
+        }
+
+        int data2 = data[p++];
+        if (data2 == 0xFFFF) {
+          data2 = -1;
+        }
+
+        atn.lexerActions[i] = lexerActionFactory(actionType, data1, data2);
+      }
+    } else {
+      // for compatibility with older serialized ATNs, convert the old
+      // serialized action index for action transitions to the new
+      // form, which is the index of a LexerCustomAction
+      for (ATNState *state : atn.states) {
+        for (size_t i = 0; i < state->transitions.size(); i++) {
+          Transition *transition = state->transitions[i];
+          if (!is<ActionTransition *>(transition)) {
+            continue;
+          }
+
+          size_t ruleIndex = static_cast<ActionTransition *>(transition)->ruleIndex;
+          size_t actionIndex = static_cast<ActionTransition *>(transition)->actionIndex;
+          Ref<LexerCustomAction> lexerAction = std::make_shared<LexerCustomAction>(ruleIndex, actionIndex);
+          state->transitions[i] = new ActionTransition(transition->target, ruleIndex, atn.lexerActions.size(), false); /* mem-check freed in ATNState d-tor */
+          delete transition; // ml: no longer needed since we just replaced it.
+          atn.lexerActions.push_back(lexerAction);
+        }
+      }
+    }
+  }
+
+  markPrecedenceDecisions(atn);
+
+  if (deserializationOptions.isVerifyATN()) {
+    verifyATN(atn);
+  }
+
+  if (deserializationOptions.isGenerateRuleBypassTransitions() && atn.grammarType == ATNType::PARSER) {
+    atn.ruleToTokenType.resize(atn.ruleToStartState.size());
+    for (size_t i = 0; i < atn.ruleToStartState.size(); i++) {
+      atn.ruleToTokenType[i] = int(atn.maxTokenType + i + 1);
+    }
+
+    for (std::vector<RuleStartState*>::size_type i = 0; i < atn.ruleToStartState.size(); i++) {
+      BasicBlockStartState *bypassStart = new BasicBlockStartState(); /* mem check: freed in ATN d-tor */
+      bypassStart->ruleIndex = (int)i;
+      atn.addState(bypassStart);
+
+      BlockEndState *bypassStop = new BlockEndState(); /* mem check: freed in ATN d-tor */
+      bypassStop->ruleIndex = (int)i;
+      atn.addState(bypassStop);
+
+      bypassStart->endState = bypassStop;
+      atn.defineDecisionState(bypassStart);
+
+      bypassStop->startState = bypassStart;
+
+      ATNState *endState;
+      Transition *excludeTransition = nullptr;
+      if (atn.ruleToStartState[i]->isLeftRecursiveRule) {
+        // wrap from the beginning of the rule to the StarLoopEntryState
+        endState = nullptr;
+        for (ATNState *state : atn.states) {
+          if (state->ruleIndex != i) {
+            continue;
+          }
+
+          if (!is<StarLoopEntryState*>(state)) {
+            continue;
+          }
+
+          ATNState *maybeLoopEndState = state->transitions[state->transitions.size() - 1]->target;
+          if (!is<LoopEndState*>(maybeLoopEndState)) {
+            continue;
+          }
+
+          if (maybeLoopEndState->epsilonOnlyTransitions && is<RuleStopState*>(maybeLoopEndState->transitions[0]->target)) {
+            endState = state;
+            break;
+          }
+        }
+
+        if (endState == nullptr) {
+          throw UnsupportedOperationException("Couldn't identify final state of the precedence rule prefix section.");
+
+        }
+
+        excludeTransition = (static_cast<StarLoopEntryState*>(endState))->loopBackState->transitions[0];
+      } else {
+        endState = atn.ruleToStopState[i];
+      }
+
+      // all non-excluded transitions that currently target end state need to target blockEnd instead
+      for (ATNState *state : atn.states) {
+        for (Transition *transition : state->transitions) {
+          if (transition == excludeTransition) {
+            continue;
+          }
+
+          if (transition->target == endState) {
+            transition->target = bypassStop;
+          }
+        }
+      }
+
+      // all transitions leaving the rule start state need to leave blockStart instead
+      while (atn.ruleToStartState[i]->transitions.size() > 0) {
+        Transition *transition = atn.ruleToStartState[i]->removeTransition(atn.ruleToStartState[i]->transitions.size() - 1);
+        bypassStart->addTransition(transition);
+      }
+
+      // link the new states
+      atn.ruleToStartState[i]->addTransition(new EpsilonTransition(bypassStart));  /* mem check: freed in ATNState d-tor */
+      bypassStop->addTransition(new EpsilonTransition(endState)); /* mem check: freed in ATNState d-tor */
+
+      ATNState *matchState = new BasicState(); /* mem check: freed in ATN d-tor */
+      atn.addState(matchState);
+      matchState->addTransition(new AtomTransition(bypassStop, atn.ruleToTokenType[i])); /* mem check: freed in ATNState d-tor */
+      bypassStart->addTransition(new EpsilonTransition(matchState)); /* mem check: freed in ATNState d-tor */
+    }
+
+    if (deserializationOptions.isVerifyATN()) {
+      // reverify after modification
+      verifyATN(atn);
+    }
+  }
+
+  return atn;
+}
+
+/**
+ * Analyze the {@link StarLoopEntryState} states in the specified ATN to set
+ * the {@link StarLoopEntryState#isPrecedenceDecision} field to the
+ * correct value.
+ *
+ * @param atn The ATN.
+ */
+void ATNDeserializer::markPrecedenceDecisions(const ATN &atn) {
+  for (ATNState *state : atn.states) {
+    if (!is<StarLoopEntryState *>(state)) {
+      continue;
+    }
+
+    /* We analyze the ATN to determine if this ATN decision state is the
+     * decision for the closure block that determines whether a
+     * precedence rule should continue or complete.
+     */
+    if (atn.ruleToStartState[state->ruleIndex]->isLeftRecursiveRule) {
+      ATNState *maybeLoopEndState = state->transitions[state->transitions.size() - 1]->target;
+      if (is<LoopEndState *>(maybeLoopEndState)) {
+        if (maybeLoopEndState->epsilonOnlyTransitions && is<RuleStopState *>(maybeLoopEndState->transitions[0]->target)) {
+          static_cast<StarLoopEntryState *>(state)->isPrecedenceDecision = true;
+        }
+      }
+    }
+  }
+}
+
+void ATNDeserializer::verifyATN(const ATN &atn) {
+  // verify assumptions
+  for (ATNState *state : atn.states) {
+    if (state == nullptr) {
+      continue;
+    }
+
+    checkCondition(state->epsilonOnlyTransitions || state->transitions.size() <= 1);
+
+    if (is<PlusBlockStartState *>(state)) {
+      checkCondition((static_cast<PlusBlockStartState *>(state))->loopBackState != nullptr);
+    }
+
+    if (is<StarLoopEntryState *>(state)) {
+      StarLoopEntryState *starLoopEntryState = static_cast<StarLoopEntryState*>(state);
+      checkCondition(starLoopEntryState->loopBackState != nullptr);
+      checkCondition(starLoopEntryState->transitions.size() == 2);
+
+      if (is<StarBlockStartState *>(starLoopEntryState->transitions[0]->target)) {
+        checkCondition(static_cast<LoopEndState *>(starLoopEntryState->transitions[1]->target) != nullptr);
+        checkCondition(!starLoopEntryState->nonGreedy);
+      } else if (is<LoopEndState *>(starLoopEntryState->transitions[0]->target)) {
+        checkCondition(is<StarBlockStartState *>(starLoopEntryState->transitions[1]->target));
+        checkCondition(starLoopEntryState->nonGreedy);
+      } else {
+        throw IllegalStateException();
+
+      }
+    }
+
+    if (is<StarLoopbackState *>(state)) {
+      checkCondition(state->transitions.size() == 1);
+      checkCondition(is<StarLoopEntryState *>(state->transitions[0]->target));
+    }
+
+    if (is<LoopEndState *>(state)) {
+      checkCondition((static_cast<LoopEndState *>(state))->loopBackState != nullptr);
+    }
+
+    if (is<RuleStartState *>(state)) {
+      checkCondition((static_cast<RuleStartState *>(state))->stopState != nullptr);
+    }
+
+    if (is<BlockStartState *>(state)) {
+      checkCondition((static_cast<BlockStartState *>(state))->endState != nullptr);
+    }
+
+    if (is<BlockEndState *>(state)) {
+      checkCondition((static_cast<BlockEndState *>(state))->startState != nullptr);
+    }
+
+    if (is<DecisionState *>(state)) {
+      DecisionState *decisionState = static_cast<DecisionState *>(state);
+      checkCondition(decisionState->transitions.size() <= 1 || decisionState->decision >= 0);
+    } else {
+      checkCondition(state->transitions.size() <= 1 || is<RuleStopState *>(state));
+    }
+  }
+}
+
+void ATNDeserializer::checkCondition(bool condition) {
+  checkCondition(condition, "");
+}
+
+void ATNDeserializer::checkCondition(bool condition, const std::string &message) {
+  if (!condition) {
+    throw IllegalStateException(message);
+  }
+}
+
+Guid ATNDeserializer::toUUID(const unsigned short *data, size_t offset) {
+  return Guid((uint16_t *)data + offset, true);
+}
+
+/* mem check: all created instances are freed in the d-tor of the ATNState they are added to. */
+Transition *ATNDeserializer::edgeFactory(const ATN &atn, size_t type, size_t /*src*/, size_t trg, size_t arg1,
+                                         size_t arg2, size_t arg3,
+  const std::vector<misc::IntervalSet> &sets) {
+
+  ATNState *target = atn.states[trg];
+  switch (type) {
+    case Transition::EPSILON:
+      return new EpsilonTransition(target);
+    case Transition::RANGE:
+      if (arg3 != 0) {
+        return new RangeTransition(target, Token::EOF, arg2);
+      } else {
+        return new RangeTransition(target, arg1, arg2);
+      }
+    case Transition::RULE:
+      return new RuleTransition(static_cast<RuleStartState*>(atn.states[arg1]), arg2, (int)arg3, target);
+    case Transition::PREDICATE:
+      return new PredicateTransition(target, arg1, arg2, arg3 != 0);
+    case Transition::PRECEDENCE:
+      return new PrecedencePredicateTransition(target, (int)arg1);
+    case Transition::ATOM:
+      if (arg3 != 0) {
+        return new AtomTransition(target, Token::EOF);
+      } else {
+        return new AtomTransition(target, arg1);
+      }
+    case Transition::ACTION:
+      return new ActionTransition(target, arg1, arg2, arg3 != 0);
+    case Transition::SET:
+      return new SetTransition(target, sets[arg1]);
+    case Transition::NOT_SET:
+      return new NotSetTransition(target, sets[arg1]);
+    case Transition::WILDCARD:
+      return new WildcardTransition(target);
+  }
+
+  throw IllegalArgumentException("The specified transition type is not valid.");
+}
+
+/* mem check: all created instances are freed in the d-tor of the ATN. */
+ATNState* ATNDeserializer::stateFactory(size_t type, size_t ruleIndex) {
+  ATNState *s;
+  switch (type) {
+    case ATNState::ATN_INVALID_TYPE:
+      return nullptr;
+    case ATNState::BASIC :
+      s = new BasicState();
+      break;
+    case ATNState::RULE_START :
+      s = new RuleStartState();
+      break;
+    case ATNState::BLOCK_START :
+      s = new BasicBlockStartState();
+      break;
+    case ATNState::PLUS_BLOCK_START :
+      s = new PlusBlockStartState();
+      break;
+    case ATNState::STAR_BLOCK_START :
+      s = new StarBlockStartState();
+      break;
+    case ATNState::TOKEN_START :
+      s = new TokensStartState();
+      break;
+    case ATNState::RULE_STOP :
+      s = new RuleStopState();
+      break;
+    case ATNState::BLOCK_END :
+      s = new BlockEndState();
+      break;
+    case ATNState::STAR_LOOP_BACK :
+      s = new StarLoopbackState();
+      break;
+    case ATNState::STAR_LOOP_ENTRY :
+      s = new StarLoopEntryState();
+      break;
+    case ATNState::PLUS_LOOP_BACK :
+      s = new PlusLoopbackState();
+      break;
+    case ATNState::LOOP_END :
+      s = new LoopEndState();
+      break;
+    default :
+      std::string message = "The specified state type " + std::to_string(type) + " is not valid.";
+      throw IllegalArgumentException(message);
+  }
+
+  s->ruleIndex = ruleIndex;
+  return s;
+}
+
+Ref<LexerAction> ATNDeserializer::lexerActionFactory(LexerActionType type, int data1, int data2) {
+  switch (type) {
+    case LexerActionType::CHANNEL:
+      return std::make_shared<LexerChannelAction>(data1);
+
+    case LexerActionType::CUSTOM:
+      return std::make_shared<LexerCustomAction>(data1, data2);
+
+    case LexerActionType::MODE:
+      return std::make_shared< LexerModeAction>(data1);
+
+    case LexerActionType::MORE:
+      return LexerMoreAction::getInstance();
+
+    case LexerActionType::POP_MODE:
+      return LexerPopModeAction::getInstance();
+
+    case LexerActionType::PUSH_MODE:
+      return std::make_shared<LexerPushModeAction>(data1);
+
+    case LexerActionType::SKIP:
+      return LexerSkipAction::getInstance();
+
+    case LexerActionType::TYPE:
+      return std::make_shared<LexerTypeAction>(data1);
+
+    default:
+      throw IllegalArgumentException("The specified lexer action type " + std::to_string(static_cast<size_t>(type)) +
+                                     " is not valid.");
+  }
+}

runtime-linux/antlr4-runtime/atn/ATNDeserializer.h

@@ -0,0 +1,85 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/ATNDeserializationOptions.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC ATNDeserializer {
+  public:
+    static const size_t SERIALIZED_VERSION;
+
+    /// This is the current serialized UUID.
+    // ml: defined as function to avoid the “static initialization order fiasco”.
+    static Guid SERIALIZED_UUID();
+
+    ATNDeserializer();
+    ATNDeserializer(const ATNDeserializationOptions& dso);
+    virtual ~ATNDeserializer();
+
+    static Guid toUUID(const unsigned short *data, size_t offset);
+
+    virtual ATN deserialize(const std::vector<uint16_t> &input);
+    virtual void verifyATN(const ATN &atn);
+
+    static void checkCondition(bool condition);
+    static void checkCondition(bool condition, const std::string &message);
+
+    static Transition *edgeFactory(const ATN &atn, size_t type, size_t src, size_t trg, size_t arg1, size_t arg2,
+                                   size_t arg3, const std::vector<misc::IntervalSet> &sets);
+
+    static ATNState *stateFactory(size_t type, size_t ruleIndex);
+
+  protected:
+    /// Determines if a particular serialized representation of an ATN supports
+    /// a particular feature, identified by the <seealso cref="UUID"/> used for serializing
+    /// the ATN at the time the feature was first introduced.
+    ///
+    /// <param name="feature"> The <seealso cref="UUID"/> marking the first time the feature was
+    /// supported in the serialized ATN. </param>
+    /// <param name="actualUuid"> The <seealso cref="UUID"/> of the actual serialized ATN which is
+    /// currently being deserialized. </param>
+    /// <returns> {@code true} if the {@code actualUuid} value represents a
+    /// serialized ATN at or after the feature identified by {@code feature} was
+    /// introduced; otherwise, {@code false}. </returns>
+    virtual bool isFeatureSupported(const Guid &feature, const Guid &actualUuid);
+    void markPrecedenceDecisions(const ATN &atn);
+    Ref<LexerAction> lexerActionFactory(LexerActionType type, int data1, int data2);
+
+  private:
+    /// This is the earliest supported serialized UUID.
+    static Guid BASE_SERIALIZED_UUID();
+
+    /// This UUID indicates an extension of <seealso cref="BASE_SERIALIZED_UUID"/> for the
+    /// addition of precedence predicates.
+    static Guid ADDED_PRECEDENCE_TRANSITIONS();
+
+    /**
+     * This UUID indicates an extension of ADDED_PRECEDENCE_TRANSITIONS
+     * for the addition of lexer actions encoded as a sequence of
+     * LexerAction instances.
+     */
+    static Guid ADDED_LEXER_ACTIONS();
+
+    /**
+     * This UUID indicates the serialized ATN contains two sets of
+     * IntervalSets, where the second set's values are encoded as
+     * 32-bit integers to support the full Unicode SMP range up to U+10FFFF.
+     */
+    static Guid ADDED_UNICODE_SMP();
+
+    /// This list contains all of the currently supported UUIDs, ordered by when
+    /// the feature first appeared in this branch.
+    static std::vector<Guid>& SUPPORTED_UUIDS();
+
+    ATNDeserializationOptions deserializationOptions;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ATNSerializer.cpp

@@ -0,0 +1,621 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/IntervalSet.h"
+#include "atn/ATNType.h"
+#include "atn/ATNState.h"
+#include "atn/BlockEndState.h"
+
+#include "atn/DecisionState.h"
+#include "atn/RuleStartState.h"
+#include "atn/LoopEndState.h"
+#include "atn/BlockStartState.h"
+#include "atn/Transition.h"
+#include "atn/SetTransition.h"
+#include "Token.h"
+#include "misc/Interval.h"
+#include "atn/ATN.h"
+
+#include "atn/RuleTransition.h"
+#include "atn/PrecedencePredicateTransition.h"
+#include "atn/PredicateTransition.h"
+#include "atn/RangeTransition.h"
+#include "atn/AtomTransition.h"
+#include "atn/ActionTransition.h"
+#include "atn/ATNDeserializer.h"
+
+#include "atn/TokensStartState.h"
+#include "Exceptions.h"
+#include "support/CPPUtils.h"
+
+#include "atn/LexerChannelAction.h"
+#include "atn/LexerCustomAction.h"
+#include "atn/LexerModeAction.h"
+#include "atn/LexerPushModeAction.h"
+#include "atn/LexerTypeAction.h"
+
+#include "Exceptions.h"
+
+#include "atn/ATNSerializer.h"
+
+using namespace antlrcpp;
+using namespace antlr4::atn;
+
+ATNSerializer::ATNSerializer(ATN *atn) { this->atn = atn; }
+
+ATNSerializer::ATNSerializer(ATN *atn, const std::vector<std::string> &tokenNames) {
+  this->atn = atn;
+  _tokenNames = tokenNames;
+}
+
+ATNSerializer::~ATNSerializer() { }
+
+std::vector<size_t> ATNSerializer::serialize() {
+  std::vector<size_t> data;
+  data.push_back(ATNDeserializer::SERIALIZED_VERSION);
+  serializeUUID(data, ATNDeserializer::SERIALIZED_UUID());
+
+  // convert grammar type to ATN const to avoid dependence on ANTLRParser
+  data.push_back(static_cast<size_t>(atn->grammarType));
+  data.push_back(atn->maxTokenType);
+  size_t nedges = 0;
+
+  std::unordered_map<misc::IntervalSet, int> setIndices;
+  std::vector<misc::IntervalSet> sets;
+
+  // dump states, count edges and collect sets while doing so
+  std::vector<size_t> nonGreedyStates;
+  std::vector<size_t> precedenceStates;
+  data.push_back(atn->states.size());
+  for (ATNState *s : atn->states) {
+    if (s == nullptr) {  // might be optimized away
+      data.push_back(ATNState::ATN_INVALID_TYPE);
+      continue;
+    }
+
+    size_t stateType = s->getStateType();
+    if (is<DecisionState *>(s) && (static_cast<DecisionState *>(s))->nonGreedy) {
+      nonGreedyStates.push_back(s->stateNumber);
+    }
+
+    if (is<RuleStartState *>(s) && (static_cast<RuleStartState *>(s))->isLeftRecursiveRule) {
+      precedenceStates.push_back(s->stateNumber);
+    }
+
+    data.push_back(stateType);
+
+    if (s->ruleIndex == INVALID_INDEX) {
+      data.push_back(0xFFFF);
+    }
+    else {
+      data.push_back(s->ruleIndex);
+    }
+
+    if (s->getStateType() == ATNState::LOOP_END) {
+      data.push_back((static_cast<LoopEndState *>(s))->loopBackState->stateNumber);
+    }
+    else if (is<BlockStartState *>(s)) {
+      data.push_back((static_cast<BlockStartState *>(s))->endState->stateNumber);
+    }
+
+    if (s->getStateType() != ATNState::RULE_STOP) {
+      // the deserializer can trivially derive these edges, so there's no need
+      // to serialize them
+      nedges += s->transitions.size();
+    }
+
+    for (size_t i = 0; i < s->transitions.size(); i++) {
+      Transition *t = s->transitions[i];
+      Transition::SerializationType edgeType = t->getSerializationType();
+      if (edgeType == Transition::SET || edgeType == Transition::NOT_SET) {
+        SetTransition *st = static_cast<SetTransition *>(t);
+        if (setIndices.find(st->set) == setIndices.end()) {
+          sets.push_back(st->set);
+          setIndices.insert({ st->set, (int)sets.size() - 1 });
+        }
+      }
+    }
+  }
+
+  // non-greedy states
+  data.push_back(nonGreedyStates.size());
+  for (size_t i = 0; i < nonGreedyStates.size(); i++) {
+    data.push_back(nonGreedyStates.at(i));
+  }
+
+  // precedence states
+  data.push_back(precedenceStates.size());
+  for (size_t i = 0; i < precedenceStates.size(); i++) {
+    data.push_back(precedenceStates.at(i));
+  }
+
+  size_t nrules = atn->ruleToStartState.size();
+  data.push_back(nrules);
+  for (size_t r = 0; r < nrules; r++) {
+    ATNState *ruleStartState = atn->ruleToStartState[r];
+    data.push_back(ruleStartState->stateNumber);
+    if (atn->grammarType == ATNType::LEXER) {
+      if (atn->ruleToTokenType[r] == Token::EOF) {
+        data.push_back(0xFFFF);
+      }
+      else {
+        data.push_back(atn->ruleToTokenType[r]);
+      }
+    }
+  }
+
+  size_t nmodes = atn->modeToStartState.size();
+  data.push_back(nmodes);
+  if (nmodes > 0) {
+    for (const auto &modeStartState : atn->modeToStartState) {
+      data.push_back(modeStartState->stateNumber);
+    }
+  }
+
+  size_t nsets = sets.size();
+  data.push_back(nsets);
+  for (auto set : sets) {
+    bool containsEof = set.contains(Token::EOF);
+    if (containsEof && set.getIntervals().at(0).b == -1) {
+      data.push_back(set.getIntervals().size() - 1);
+    }
+    else {
+      data.push_back(set.getIntervals().size());
+    }
+
+    data.push_back(containsEof ? 1 : 0);
+    for (auto &interval : set.getIntervals()) {
+      if (interval.a == -1) {
+        if (interval.b == -1) {
+          continue;
+        } else {
+          data.push_back(0);
+        }
+      }
+      else {
+        data.push_back(interval.a);
+      }
+
+      data.push_back(interval.b);
+    }
+  }
+
+  data.push_back(nedges);
+  for (ATNState *s : atn->states) {
+    if (s == nullptr) {
+      // might be optimized away
+      continue;
+    }
+
+    if (s->getStateType() == ATNState::RULE_STOP) {
+      continue;
+    }
+
+    for (size_t i = 0; i < s->transitions.size(); i++) {
+      Transition *t = s->transitions[i];
+
+      if (atn->states[t->target->stateNumber] == nullptr) {
+        throw IllegalStateException("Cannot serialize a transition to a removed state.");
+      }
+
+      size_t src = s->stateNumber;
+      size_t trg = t->target->stateNumber;
+      Transition::SerializationType edgeType = t->getSerializationType();
+      size_t arg1 = 0;
+      size_t arg2 = 0;
+      size_t arg3 = 0;
+      switch (edgeType) {
+        case Transition::RULE:
+          trg = (static_cast<RuleTransition *>(t))->followState->stateNumber;
+          arg1 = (static_cast<RuleTransition *>(t))->target->stateNumber;
+          arg2 = (static_cast<RuleTransition *>(t))->ruleIndex;
+          arg3 = (static_cast<RuleTransition *>(t))->precedence;
+          break;
+        case Transition::PRECEDENCE:
+        {
+          PrecedencePredicateTransition *ppt =
+          static_cast<PrecedencePredicateTransition *>(t);
+          arg1 = ppt->precedence;
+        }
+          break;
+        case Transition::PREDICATE:
+        {
+          PredicateTransition *pt = static_cast<PredicateTransition *>(t);
+          arg1 = pt->ruleIndex;
+          arg2 = pt->predIndex;
+          arg3 = pt->isCtxDependent ? 1 : 0;
+        }
+          break;
+        case Transition::RANGE:
+          arg1 = (static_cast<RangeTransition *>(t))->from;
+          arg2 = (static_cast<RangeTransition *>(t))->to;
+          if (arg1 == Token::EOF) {
+            arg1 = 0;
+            arg3 = 1;
+          }
+
+          break;
+        case Transition::ATOM:
+          arg1 = (static_cast<AtomTransition *>(t))->_label;
+          if (arg1 == Token::EOF) {
+            arg1 = 0;
+            arg3 = 1;
+          }
+
+          break;
+        case Transition::ACTION:
+        {
+          ActionTransition *at = static_cast<ActionTransition *>(t);
+          arg1 = at->ruleIndex;
+          arg2 = at->actionIndex;
+          if (arg2 == INVALID_INDEX) {
+            arg2 = 0xFFFF;
+          }
+
+          arg3 = at->isCtxDependent ? 1 : 0;
+        }
+          break;
+        case Transition::SET:
+          arg1 = setIndices[(static_cast<SetTransition *>(t))->set];
+          break;
+
+        case Transition::NOT_SET:
+          arg1 = setIndices[(static_cast<SetTransition *>(t))->set];
+          break;
+
+        default:
+          break;
+      }
+
+      data.push_back(src);
+      data.push_back(trg);
+      data.push_back(edgeType);
+      data.push_back(arg1);
+      data.push_back(arg2);
+      data.push_back(arg3);
+    }
+  }
+
+  size_t ndecisions = atn->decisionToState.size();
+  data.push_back(ndecisions);
+  for (DecisionState *decStartState : atn->decisionToState) {
+    data.push_back(decStartState->stateNumber);
+  }
+
+  // LEXER ACTIONS
+  if (atn->grammarType == ATNType::LEXER) {
+    data.push_back(atn->lexerActions.size());
+    for (Ref<LexerAction> &action : atn->lexerActions) {
+      data.push_back(static_cast<size_t>(action->getActionType()));
+      switch (action->getActionType()) {
+        case LexerActionType::CHANNEL:
+        {
+          int channel = std::dynamic_pointer_cast<LexerChannelAction>(action)->getChannel();
+          data.push_back(channel != -1 ? channel : 0xFFFF);
+          data.push_back(0);
+          break;
+        }
+
+        case LexerActionType::CUSTOM:
+        {
+          size_t ruleIndex = std::dynamic_pointer_cast<LexerCustomAction>(action)->getRuleIndex();
+          size_t actionIndex = std::dynamic_pointer_cast<LexerCustomAction>(action)->getActionIndex();
+          data.push_back(ruleIndex != INVALID_INDEX ? ruleIndex : 0xFFFF);
+          data.push_back(actionIndex != INVALID_INDEX ? actionIndex : 0xFFFF);
+          break;
+        }
+
+        case LexerActionType::MODE:
+        {
+          int mode = std::dynamic_pointer_cast<LexerModeAction>(action)->getMode();
+          data.push_back(mode != -1 ? mode : 0xFFFF);
+          data.push_back(0);
+          break;
+        }
+
+        case LexerActionType::MORE:
+          data.push_back(0);
+          data.push_back(0);
+          break;
+
+        case LexerActionType::POP_MODE:
+          data.push_back(0);
+          data.push_back(0);
+          break;
+
+        case LexerActionType::PUSH_MODE:
+        {
+          int mode = std::dynamic_pointer_cast<LexerPushModeAction>(action)->getMode();
+          data.push_back(mode != -1 ? mode : 0xFFFF);
+          data.push_back(0);
+          break;
+        }
+
+        case LexerActionType::SKIP:
+          data.push_back(0);
+          data.push_back(0);
+          break;
+
+        case LexerActionType::TYPE:
+        {
+          int type = std::dynamic_pointer_cast<LexerTypeAction>(action)->getType();
+          data.push_back(type != -1 ? type : 0xFFFF);
+          data.push_back(0);
+          break;
+        }
+
+        default:
+          throw IllegalArgumentException("The specified lexer action type " +
+                                         std::to_string(static_cast<size_t>(action->getActionType())) +
+                                         " is not valid.");
+      }
+    }
+  }
+
+  // don't adjust the first value since that's the version number
+  for (size_t i = 1; i < data.size(); i++) {
+    if (data.at(i) > 0xFFFF) {
+      throw UnsupportedOperationException("Serialized ATN data element out of range.");
+    }
+
+    size_t value = (data.at(i) + 2) & 0xFFFF;
+    data.at(i) = value;
+  }
+
+  return data;
+}
+
+//------------------------------------------------------------------------------------------------------------
+
+std::string ATNSerializer::decode(const std::wstring &inpdata) {
+  if (inpdata.size() < 10)
+    throw IllegalArgumentException("Not enough data to decode");
+
+  std::vector<uint16_t> data(inpdata.size());
+  data[0] = (uint16_t)inpdata[0];
+
+  // Don't adjust the first value since that's the version number.
+  for (size_t i = 1; i < inpdata.size(); ++i) {
+    data[i] = (uint16_t)inpdata[i] - 2;
+  }
+
+  std::string buf;
+  size_t p = 0;
+  size_t version = data[p++];
+  if (version != ATNDeserializer::SERIALIZED_VERSION) {
+    std::string reason = "Could not deserialize ATN with version " + std::to_string(version) + "(expected " +
+    std::to_string(ATNDeserializer::SERIALIZED_VERSION) + ").";
+    throw UnsupportedOperationException("ATN Serializer" + reason);
+  }
+
+  Guid uuid = ATNDeserializer::toUUID(data.data(), p);
+  p += 8;
+  if (uuid != ATNDeserializer::SERIALIZED_UUID()) {
+    std::string reason = "Could not deserialize ATN with UUID " + uuid.toString() + " (expected " +
+    ATNDeserializer::SERIALIZED_UUID().toString() + ").";
+    throw UnsupportedOperationException("ATN Serializer" + reason);
+  }
+
+  p++;  // skip grammarType
+  size_t maxType = data[p++];
+  buf.append("max type ").append(std::to_string(maxType)).append("\n");
+  size_t nstates = data[p++];
+  for (size_t i = 0; i < nstates; i++) {
+    size_t stype = data[p++];
+    if (stype == ATNState::ATN_INVALID_TYPE) {  // ignore bad type of states
+      continue;
+    }
+    size_t ruleIndex = data[p++];
+    if (ruleIndex == 0xFFFF) {
+      ruleIndex = INVALID_INDEX;
+    }
+
+    std::string arg = "";
+    if (stype == ATNState::LOOP_END) {
+      int loopBackStateNumber = data[p++];
+      arg = std::string(" ") + std::to_string(loopBackStateNumber);
+    }
+    else if (stype == ATNState::PLUS_BLOCK_START ||
+             stype == ATNState::STAR_BLOCK_START ||
+             stype == ATNState::BLOCK_START) {
+      int endStateNumber = data[p++];
+      arg = std::string(" ") + std::to_string(endStateNumber);
+    }
+    buf.append(std::to_string(i))
+    .append(":")
+    .append(ATNState::serializationNames[stype])
+    .append(" ")
+    .append(std::to_string(ruleIndex))
+    .append(arg)
+    .append("\n");
+  }
+  size_t numNonGreedyStates = data[p++];
+  p += numNonGreedyStates; // Instead of that useless loop below.
+  /*
+   for (int i = 0; i < numNonGreedyStates; i++) {
+   int stateNumber = data[p++];
+   }
+   */
+
+  size_t numPrecedenceStates = data[p++];
+  p += numPrecedenceStates;
+  /*
+   for (int i = 0; i < numPrecedenceStates; i++) {
+   int stateNumber = data[p++];
+   }
+   */
+
+  size_t nrules = data[p++];
+  for (size_t i = 0; i < nrules; i++) {
+    size_t s = data[p++];
+    if (atn->grammarType == ATNType::LEXER) {
+      size_t arg1 = data[p++];
+      buf.append("rule ")
+      .append(std::to_string(i))
+      .append(":")
+      .append(std::to_string(s))
+      .append(" ")
+      .append(std::to_string(arg1))
+      .append("\n");
+    }
+    else {
+      buf.append("rule ")
+      .append(std::to_string(i))
+      .append(":")
+      .append(std::to_string(s))
+      .append("\n");
+    }
+  }
+  size_t nmodes = data[p++];
+  for (size_t i = 0; i < nmodes; i++) {
+    size_t s = data[p++];
+    buf.append("mode ")
+    .append(std::to_string(i))
+    .append(":")
+    .append(std::to_string(s))
+    .append("\n");
+  }
+  size_t nsets = data[p++];
+  for (size_t i = 0; i < nsets; i++) {
+    size_t nintervals = data[p++];
+    buf.append(std::to_string(i)).append(":");
+    bool containsEof = data[p++] != 0;
+    if (containsEof) {
+      buf.append(getTokenName(Token::EOF));
+    }
+
+    for (size_t j = 0; j < nintervals; j++) {
+      if (containsEof || j > 0) {
+        buf.append(", ");
+      }
+
+      buf.append(getTokenName(data[p]))
+      .append("..")
+      .append(getTokenName(data[p + 1]));
+      p += 2;
+    }
+    buf.append("\n");
+  }
+  size_t nedges = data[p++];
+  for (size_t i = 0; i < nedges; i++) {
+    size_t src = data[p];
+    size_t trg = data[p + 1];
+    size_t ttype = data[p + 2];
+    size_t arg1 = data[p + 3];
+    size_t arg2 = data[p + 4];
+    size_t arg3 = data[p + 5];
+    buf.append(std::to_string(src))
+    .append("->")
+    .append(std::to_string(trg))
+    .append(" ")
+    .append(Transition::serializationNames[ttype])
+    .append(" ")
+    .append(std::to_string(arg1))
+    .append(",")
+    .append(std::to_string(arg2))
+    .append(",")
+    .append(std::to_string(arg3))
+    .append("\n");
+    p += 6;
+  }
+  size_t ndecisions = data[p++];
+  for (size_t i = 0; i < ndecisions; i++) {
+    size_t s = data[p++];
+    buf += std::to_string(i) + ":" + std::to_string(s) + "\n";
+  }
+
+  if (atn->grammarType == ATNType::LEXER) {
+    //int lexerActionCount = data[p++];
+
+    //p += lexerActionCount * 3; // Instead of useless loop below.
+    /*
+    for (int i = 0; i < lexerActionCount; i++) {
+      LexerActionType actionType = (LexerActionType)data[p++];
+      int data1 = data[p++];
+      int data2 = data[p++];
+    }
+     */
+  }
+
+  return buf;
+}
+
+std::string ATNSerializer::getTokenName(size_t t) {
+  if (t == Token::EOF) {
+    return "EOF";
+  }
+
+  if (atn->grammarType == ATNType::LEXER && t <= 0x10FFFF) {
+    switch (t) {
+      case '\n':
+        return "'\\n'";
+      case '\r':
+        return "'\\r'";
+      case '\t':
+        return "'\\t'";
+      case '\b':
+        return "'\\b'";
+      case '\f':
+        return "'\\f'";
+      case '\\':
+        return "'\\\\'";
+      case '\'':
+        return "'\\''";
+      default:
+        std::string s_hex = antlrcpp::toHexString((int)t);
+        if (s_hex >= "0" && s_hex <= "7F" && !iscntrl((int)t)) {
+          return "'" + std::to_string(t) + "'";
+        }
+
+        // turn on the bit above max "\u10FFFF" value so that we pad with zeros
+        // then only take last 6 digits
+        std::string hex = antlrcpp::toHexString((int)t | 0x1000000).substr(1, 6);
+        std::string unicodeStr = std::string("'\\u") + hex + std::string("'");
+        return unicodeStr;
+    }
+  }
+
+  if (_tokenNames.size() > 0 && t < _tokenNames.size()) {
+    return _tokenNames[t];
+  }
+
+  return std::to_string(t);
+}
+
+std::wstring ATNSerializer::getSerializedAsString(ATN *atn) {
+  std::vector<size_t> data = getSerialized(atn);
+  std::wstring result;
+  for (size_t entry : data)
+    result.push_back((wchar_t)entry);
+
+  return result;
+}
+
+std::vector<size_t> ATNSerializer::getSerialized(ATN *atn) {
+  return ATNSerializer(atn).serialize();
+}
+
+std::string ATNSerializer::getDecoded(ATN *atn, std::vector<std::string> &tokenNames) {
+  std::wstring serialized = getSerializedAsString(atn);
+  return ATNSerializer(atn, tokenNames).decode(serialized);
+}
+
+void ATNSerializer::serializeUUID(std::vector<size_t> &data, Guid uuid) {
+  unsigned int twoBytes = 0;
+  bool firstByte = true;
+  for( std::vector<unsigned char>::const_reverse_iterator rit = uuid.rbegin(); rit != uuid.rend(); ++rit )
+  {
+     if (firstByte) {
+       twoBytes = *rit;
+       firstByte = false;
+     } else {
+       twoBytes |= (*rit << 8);
+       data.push_back(twoBytes);
+       firstByte = true;
+     }
+  }
+  if (!firstByte)
+     throw IllegalArgumentException( "The UUID provided is not valid (odd number of bytes)." );
+}

runtime-linux/antlr4-runtime/atn/ATNSerializer.h

@@ -0,0 +1,61 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC ATNSerializer {
+  public:
+    ATN *atn;
+
+    ATNSerializer(ATN *atn);
+    ATNSerializer(ATN *atn, const std::vector<std::string> &tokenNames);
+    virtual ~ATNSerializer();
+
+    /// <summary>
+    /// Serialize state descriptors, edge descriptors, and decision->state map
+    ///  into list of ints:
+    ///
+    /// 		grammar-type, (ANTLRParser.LEXER, ...)
+    ///  	max token type,
+    ///  	num states,
+    ///  	state-0-type ruleIndex, state-1-type ruleIndex, ... state-i-type
+    ///  ruleIndex optional-arg ...
+    ///  	num rules,
+    ///  	rule-1-start-state rule-1-args, rule-2-start-state  rule-2-args, ...
+    ///  	(args are token type,actionIndex in lexer else 0,0)
+    ///      num modes,
+    ///      mode-0-start-state, mode-1-start-state, ... (parser has 0 modes)
+    ///      num sets
+    ///      set-0-interval-count intervals, set-1-interval-count intervals, ...
+    ///  	num total edges,
+    ///      src, trg, edge-type, edge arg1, optional edge arg2 (present always),
+    ///      ...
+    ///      num decisions,
+    ///      decision-0-start-state, decision-1-start-state, ...
+    ///
+    ///  Convenient to pack into unsigned shorts to make as Java string.
+    /// </summary>
+    virtual std::vector<size_t> serialize();
+
+    virtual std::string decode(const std::wstring& data);
+    virtual std::string getTokenName(size_t t);
+
+    /// Used by Java target to encode short/int array as chars in string.
+    static std::wstring getSerializedAsString(ATN *atn);
+    static std::vector<size_t> getSerialized(ATN *atn);
+
+    static std::string getDecoded(ATN *atn, std::vector<std::string> &tokenNames);
+
+  private:
+    std::vector<std::string> _tokenNames;
+
+    void serializeUUID(std::vector<size_t> &data, Guid uuid);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ATNSimulator.cpp

@@ -0,0 +1,63 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ATNType.h"
+#include "atn/ATNConfigSet.h"
+#include "dfa/DFAState.h"
+#include "atn/ATNDeserializer.h"
+#include "atn/EmptyPredictionContext.h"
+
+#include "atn/ATNSimulator.h"
+
+using namespace antlr4;
+using namespace antlr4::dfa;
+using namespace antlr4::atn;
+
+const Ref<DFAState> ATNSimulator::ERROR = std::make_shared<DFAState>(INT32_MAX);
+antlrcpp::SingleWriteMultipleReadLock ATNSimulator::_stateLock;
+antlrcpp::SingleWriteMultipleReadLock ATNSimulator::_edgeLock;
+
+ATNSimulator::ATNSimulator(const ATN &atn, PredictionContextCache &sharedContextCache)
+: atn(atn), _sharedContextCache(sharedContextCache) {
+}
+
+ATNSimulator::~ATNSimulator() {
+}
+
+void ATNSimulator::clearDFA() {
+  throw UnsupportedOperationException("This ATN simulator does not support clearing the DFA.");
+}
+
+PredictionContextCache& ATNSimulator::getSharedContextCache() {
+  return _sharedContextCache;
+}
+
+Ref<PredictionContext> ATNSimulator::getCachedContext(Ref<PredictionContext> const& context) {
+  // This function must only be called with an active state lock, as we are going to change a shared structure.
+  std::map<Ref<PredictionContext>, Ref<PredictionContext>> visited;
+  return PredictionContext::getCachedContext(context, _sharedContextCache, visited);
+}
+
+ATN ATNSimulator::deserialize(const std::vector<uint16_t> &data) {
+  ATNDeserializer deserializer;
+  return deserializer.deserialize(data);
+}
+
+void ATNSimulator::checkCondition(bool condition) {
+  ATNDeserializer::checkCondition(condition);
+}
+
+void ATNSimulator::checkCondition(bool condition, const std::string &message) {
+  ATNDeserializer::checkCondition(condition, message);
+}
+
+Transition *ATNSimulator::edgeFactory(const ATN &atn, int type, int src, int trg, int arg1, int arg2, int arg3,
+                                      const std::vector<misc::IntervalSet> &sets) {
+  return ATNDeserializer::edgeFactory(atn, type, src, trg, arg1, arg2, arg3, sets);
+}
+
+ATNState *ATNSimulator::stateFactory(int type, int ruleIndex) {
+  return ATNDeserializer::stateFactory(type, ruleIndex);
+}

runtime-linux/antlr4-runtime/atn/ATNSimulator.h

@@ -0,0 +1,87 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATN.h"
+#include "misc/IntervalSet.h"
+#include "support/CPPUtils.h"
+#include "atn/PredictionContext.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC ATNSimulator {
+  public:
+    /// Must distinguish between missing edge and edge we know leads nowhere.
+    static const Ref<dfa::DFAState> ERROR;
+    const ATN &atn;
+
+    ATNSimulator(const ATN &atn, PredictionContextCache &sharedContextCache);
+    virtual ~ATNSimulator();
+
+    virtual void reset() = 0;
+
+    /**
+     * 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
+     */
+    virtual void clearDFA();
+    virtual PredictionContextCache& getSharedContextCache();
+    virtual Ref<PredictionContext> getCachedContext(Ref<PredictionContext> const& context);
+
+    /// @deprecated Use <seealso cref="ATNDeserializer#deserialize"/> instead.
+    static ATN deserialize(const std::vector<uint16_t> &data);
+
+    /// @deprecated Use <seealso cref="ATNDeserializer#checkCondition(boolean)"/> instead.
+    static void checkCondition(bool condition);
+
+    /// @deprecated Use <seealso cref="ATNDeserializer#checkCondition(boolean, String)"/> instead.
+    static void checkCondition(bool condition, const std::string &message);
+
+    /// @deprecated Use <seealso cref="ATNDeserializer#edgeFactory"/> instead.
+    static Transition *edgeFactory(const ATN &atn, int type, int src, int trg, int arg1, int arg2, int arg3,
+                                   const std::vector<misc::IntervalSet> &sets);
+
+    /// @deprecated Use <seealso cref="ATNDeserializer#stateFactory"/> instead.
+    static ATNState *stateFactory(int type, int ruleIndex);
+
+  protected:
+    static antlrcpp::SingleWriteMultipleReadLock _stateLock; // Lock for DFA states.
+    static antlrcpp::SingleWriteMultipleReadLock _edgeLock; // Lock for the sparse edge map in DFA states.
+
+    /// <summary>
+    /// 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().
+    ///  <p/>
+    ///  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.
+    ///  <p/>
+    ///  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.
+    /// </summary>
+    PredictionContextCache &_sharedContextCache;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ATNState.cpp

@@ -0,0 +1,71 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ATN.h"
+#include "atn/Transition.h"
+#include "misc/IntervalSet.h"
+#include "support/CPPUtils.h"
+
+#include "atn/ATNState.h"
+
+using namespace antlr4::atn;
+using namespace antlrcpp;
+
+ATNState::ATNState() {
+}
+
+ATNState::~ATNState() {
+  for (auto transition : transitions) {
+    delete transition;
+  }
+}
+
+const std::vector<std::string> ATNState::serializationNames = {
+  "INVALID", "BASIC", "RULE_START", "BLOCK_START",
+  "PLUS_BLOCK_START", "STAR_BLOCK_START", "TOKEN_START", "RULE_STOP",
+  "BLOCK_END", "STAR_LOOP_BACK", "STAR_LOOP_ENTRY", "PLUS_LOOP_BACK", "LOOP_END"
+};
+
+size_t ATNState::hashCode() {
+  return stateNumber;
+}
+
+bool ATNState::operator == (const ATNState &other) {
+  return stateNumber == other.stateNumber;
+}
+
+bool ATNState::isNonGreedyExitState() {
+  return false;
+}
+
+std::string ATNState::toString() const {
+  return std::to_string(stateNumber);
+}
+
+void ATNState::addTransition(Transition *e) {
+  addTransition(transitions.size(), e);
+}
+
+void ATNState::addTransition(size_t index, Transition *e) {
+  for (Transition *transition : transitions)
+    if (transition->target->stateNumber == e->target->stateNumber) {
+      delete e;
+      return;
+    }
+
+  if (transitions.empty()) {
+    epsilonOnlyTransitions = e->isEpsilon();
+  } else if (epsilonOnlyTransitions != e->isEpsilon()) {
+    std::cerr << "ATN state %d has both epsilon and non-epsilon transitions.\n" << stateNumber;
+    epsilonOnlyTransitions = false;
+  }
+
+  transitions.insert(transitions.begin() + index, e);
+}
+
+Transition *ATNState::removeTransition(size_t index) {
+  transitions.erase(transitions.begin() + index);
+  return nullptr;
+}

runtime-linux/antlr4-runtime/atn/ATNState.h

@@ -0,0 +1,133 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "misc/IntervalSet.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// The following images show the relation of states and
+  /// <seealso cref="ATNState#transitions"/> for various grammar constructs.
+  ///
+  /// <ul>
+  ///
+  /// <li>Solid edges marked with an &#0949; indicate a required
+  /// <seealso cref="EpsilonTransition"/>.</li>
+  ///
+  /// <li>Dashed edges indicate locations where any transition derived from
+  /// <seealso cref="Transition"/> might appear.</li>
+  ///
+  /// <li>Dashed nodes are place holders for either a sequence of linked
+  /// <seealso cref="BasicState"/> states or the inclusion of a block representing a nested
+  /// construct in one of the forms below.</li>
+  ///
+  /// <li>Nodes showing multiple outgoing alternatives with a {@code ...} support
+  /// any number of alternatives (one or more). Nodes without the {@code ...} only
+  /// support the exact number of alternatives shown in the diagram.</li>
+  ///
+  /// </ul>
+  ///
+  /// <h2>Basic Blocks</h2>
+  ///
+  /// <h3>Rule</h3>
+  ///
+  /// <embed src="images/Rule.svg" type="image/svg+xml"/>
+  ///
+  /// <h3>Block of 1 or more alternatives</h3>
+  ///
+  /// <embed src="images/Block.svg" type="image/svg+xml"/>
+  ///
+  /// <h2>Greedy Loops</h2>
+  ///
+  /// <h3>Greedy Closure: {@code (...)*}</h3>
+  ///
+  /// <embed src="images/ClosureGreedy.svg" type="image/svg+xml"/>
+  ///
+  /// <h3>Greedy Positive Closure: {@code (...)+}</h3>
+  ///
+  /// <embed src="images/PositiveClosureGreedy.svg" type="image/svg+xml"/>
+  ///
+  /// <h3>Greedy Optional: {@code (...)?}</h3>
+  ///
+  /// <embed src="images/OptionalGreedy.svg" type="image/svg+xml"/>
+  ///
+  /// <h2>Non-Greedy Loops</h2>
+  ///
+  /// <h3>Non-Greedy Closure: {@code (...)*?}</h3>
+  ///
+  /// <embed src="images/ClosureNonGreedy.svg" type="image/svg+xml"/>
+  ///
+  /// <h3>Non-Greedy Positive Closure: {@code (...)+?}</h3>
+  ///
+  /// <embed src="images/PositiveClosureNonGreedy.svg" type="image/svg+xml"/>
+  ///
+  /// <h3>Non-Greedy Optional: {@code (...)??}</h3>
+  ///
+  /// <embed src="images/OptionalNonGreedy.svg" type="image/svg+xml"/>
+  /// </summary>
+  class ANTLR4CPP_PUBLIC ATN;
+
+  class ANTLR4CPP_PUBLIC ATNState {
+  public:
+    ATNState();
+    ATNState(ATNState const&) = delete;
+
+    virtual ~ATNState();
+
+    ATNState& operator=(ATNState const&) = delete;
+
+    static const size_t INITIAL_NUM_TRANSITIONS = 4;
+    static const size_t INVALID_STATE_NUMBER = static_cast<size_t>(-1); // std::numeric_limits<size_t>::max();
+
+    enum {
+      ATN_INVALID_TYPE = 0,
+      BASIC = 1,
+      RULE_START = 2,
+      BLOCK_START = 3,
+      PLUS_BLOCK_START = 4,
+      STAR_BLOCK_START = 5,
+      TOKEN_START = 6,
+      RULE_STOP = 7,
+      BLOCK_END = 8,
+      STAR_LOOP_BACK = 9,
+      STAR_LOOP_ENTRY = 10,
+      PLUS_LOOP_BACK = 11,
+      LOOP_END = 12
+    };
+
+    static const std::vector<std::string> serializationNames;
+
+    size_t stateNumber = INVALID_STATE_NUMBER;
+    size_t ruleIndex = 0; // at runtime, we don't have Rule objects
+    bool epsilonOnlyTransitions = false;
+
+  public:
+    virtual size_t hashCode();
+    bool operator == (const ATNState &other);
+
+    /// Track the transitions emanating from this ATN state.
+    std::vector<Transition*> transitions;
+
+    virtual bool isNonGreedyExitState();
+    virtual std::string toString() const;
+    virtual void addTransition(Transition *e);
+    virtual void addTransition(size_t index, Transition *e);
+    virtual Transition* removeTransition(size_t index);
+    virtual size_t getStateType() = 0;
+
+  private:
+    /// Used to cache lookahead during parsing, not used during construction.
+
+    misc::IntervalSet _nextTokenWithinRule;
+    std::atomic<bool> _nextTokenUpdated { false };
+
+    friend class ATN;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ATNType.h

@@ -0,0 +1,20 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "antlr4-common.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Represents the type of recognizer an ATN applies to.
+  enum class ATNType {
+    LEXER = 0,
+    PARSER = 1,
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/AbstractPredicateTransition.cpp

@@ -0,0 +1,14 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/AbstractPredicateTransition.h"
+
+using namespace antlr4::atn;
+
+AbstractPredicateTransition::AbstractPredicateTransition(ATNState *target) : Transition(target) {
+}
+
+AbstractPredicateTransition::~AbstractPredicateTransition() {
+}

runtime-linux/antlr4-runtime/atn/AbstractPredicateTransition.h

@@ -0,0 +1,24 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/Transition.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTState;
+
+  class ANTLR4CPP_PUBLIC AbstractPredicateTransition : public Transition {
+
+  public:
+    AbstractPredicateTransition(ATNState *target);
+    ~AbstractPredicateTransition();
+
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ActionTransition.cpp

@@ -0,0 +1,33 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ActionTransition.h"
+
+using namespace antlr4::atn;
+
+ActionTransition::ActionTransition(ATNState *target, size_t ruleIndex)
+  : Transition(target), ruleIndex(ruleIndex), actionIndex(INVALID_INDEX), isCtxDependent(false) {
+}
+
+ActionTransition::ActionTransition(ATNState *target, size_t ruleIndex, size_t actionIndex, bool isCtxDependent)
+  : Transition(target), ruleIndex(ruleIndex), actionIndex(actionIndex), isCtxDependent(isCtxDependent) {
+}
+
+Transition::SerializationType ActionTransition::getSerializationType() const {
+  return ACTION;
+}
+
+bool ActionTransition::isEpsilon() const {
+  return true; // we are to be ignored by analysis 'cept for predicates
+}
+
+bool ActionTransition::matches(size_t /*symbol*/, size_t /*minVocabSymbol*/, size_t /*maxVocabSymbol*/) const {
+  return false;
+}
+
+std::string ActionTransition::toString() const {
+  return " ACTION " + Transition::toString() + " { ruleIndex: " + std::to_string(ruleIndex) + ", actionIndex: " +
+  std::to_string(actionIndex) + ", isCtxDependent: " + std::to_string(isCtxDependent) + " }";
+}

runtime-linux/antlr4-runtime/atn/ActionTransition.h

@@ -0,0 +1,33 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/Transition.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC ActionTransition final : public Transition {
+  public:
+    const size_t ruleIndex;
+    const size_t actionIndex;
+    const bool isCtxDependent; // e.g., $i ref in action
+
+    ActionTransition(ATNState *target, size_t ruleIndex);
+
+    ActionTransition(ATNState *target, size_t ruleIndex, size_t actionIndex, bool isCtxDependent);
+
+    virtual SerializationType getSerializationType() const override;
+
+    virtual bool isEpsilon() const override;
+
+    virtual bool matches(size_t symbol, size_t minVocabSymbol, size_t maxVocabSymbol) const override;
+
+    virtual std::string toString() const override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/AmbiguityInfo.cpp

@@ -0,0 +1,16 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/AmbiguityInfo.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+
+AmbiguityInfo::AmbiguityInfo(size_t decision, ATNConfigSet *configs, const antlrcpp::BitSet &ambigAlts,
+                             TokenStream *input, size_t startIndex, size_t stopIndex, bool fullCtx)
+  : DecisionEventInfo(decision, configs, input, startIndex, stopIndex, fullCtx) {
+
+  this->ambigAlts = ambigAlts;
+}

runtime-linux/antlr4-runtime/atn/AmbiguityInfo.h

@@ -0,0 +1,68 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionEventInfo.h"
+#include "support/BitSet.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// This class represents profiling event information for an ambiguity.
+  /// Ambiguities are decisions where a particular input resulted in an SLL
+  /// conflict, followed by LL prediction also reaching a conflict state
+  /// (indicating a true ambiguity in the grammar).
+  ///
+  /// <para>
+  /// This event may be reported during SLL prediction in cases where the
+  /// conflicting SLL configuration set provides sufficient information to
+  /// determine that the SLL conflict is truly an ambiguity. For example, if none
+  /// of the ATN configurations in the conflicting SLL configuration set have
+  /// traversed a global follow transition (i.e.
+  /// <seealso cref="ATNConfig#reachesIntoOuterContext"/> is 0 for all configurations), then
+  /// the result of SLL prediction for that input is known to be equivalent to the
+  /// result of LL prediction for that input.</para>
+  ///
+  /// <para>
+  /// In some cases, the minimum represented alternative in the conflicting LL
+  /// configuration set is not equal to the minimum represented alternative in the
+  /// conflicting SLL configuration set. Grammars and inputs which result in this
+  /// scenario are unable to use <seealso cref="PredictionMode#SLL"/>, which in turn means
+  /// they cannot use the two-stage parsing strategy to improve parsing performance
+  /// for that input.</para>
+  /// </summary>
+  /// <seealso cref= ParserATNSimulator#reportAmbiguity </seealso>
+  /// <seealso cref= ANTLRErrorListener#reportAmbiguity
+  ///
+  /// @since 4.3 </seealso>
+  class ANTLR4CPP_PUBLIC AmbiguityInfo : public DecisionEventInfo {
+  public:
+    /// The set of alternative numbers for this decision event that lead to a valid parse.
+    antlrcpp::BitSet ambigAlts;
+
+    /// <summary>
+    /// Constructs a new instance of the <seealso cref="AmbiguityInfo"/> class with the
+    /// specified detailed ambiguity information.
+    /// </summary>
+    /// <param name="decision"> The decision number </param>
+    /// <param name="configs"> The final configuration set identifying the ambiguous
+    /// alternatives for the current input </param>
+    /// <param name="ambigAlts"> The set of alternatives in the decision that lead to a valid parse.
+    ///                  The predicted alt is the min(ambigAlts) </param>
+    /// <param name="input"> The input token stream </param>
+    /// <param name="startIndex"> The start index for the current prediction </param>
+    /// <param name="stopIndex"> The index at which the ambiguity was identified during
+    /// prediction </param>
+    /// <param name="fullCtx"> {@code true} if the ambiguity was identified during LL
+    /// prediction; otherwise, {@code false} if the ambiguity was identified
+    /// during SLL prediction </param>
+    AmbiguityInfo(size_t decision, ATNConfigSet *configs, const antlrcpp::BitSet &ambigAlts, TokenStream *input,
+                  size_t startIndex, size_t stopIndex, bool fullCtx);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ArrayPredictionContext.cpp

@@ -0,0 +1,82 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "support/Arrays.h"
+#include "atn/SingletonPredictionContext.h"
+
+#include "atn/ArrayPredictionContext.h"
+
+using namespace antlr4::atn;
+
+ArrayPredictionContext::ArrayPredictionContext(Ref<SingletonPredictionContext> const& a)
+  : ArrayPredictionContext({ a->parent }, { a->returnState }) {
+}
+
+ArrayPredictionContext::ArrayPredictionContext(std::vector<Ref<PredictionContext>> const& parents_,
+                                               std::vector<size_t> const& returnStates)
+  : PredictionContext(calculateHashCode(parents_, returnStates)), parents(parents_), returnStates(returnStates) {
+    assert(parents.size() > 0);
+    assert(returnStates.size() > 0);
+}
+
+ArrayPredictionContext::~ArrayPredictionContext() {
+}
+
+bool ArrayPredictionContext::isEmpty() const {
+  // Since EMPTY_RETURN_STATE can only appear in the last position, we don't need to verify that size == 1.
+  return returnStates[0] == EMPTY_RETURN_STATE;
+}
+
+size_t ArrayPredictionContext::size() const {
+  return returnStates.size();
+}
+
+Ref<PredictionContext> ArrayPredictionContext::getParent(size_t index) const {
+  return parents[index];
+}
+
+size_t ArrayPredictionContext::getReturnState(size_t index) const {
+  return returnStates[index];
+}
+
+bool ArrayPredictionContext::operator == (PredictionContext const& o) const {
+  if (this == &o) {
+    return true;
+  }
+
+  const ArrayPredictionContext *other = dynamic_cast<const ArrayPredictionContext*>(&o);
+  if (other == nullptr || hashCode() != other->hashCode()) {
+    return false; // can't be same if hash is different
+  }
+
+  return antlrcpp::Arrays::equals(returnStates, other->returnStates) &&
+    antlrcpp::Arrays::equals(parents, other->parents);
+}
+
+std::string ArrayPredictionContext::toString() const {
+  if (isEmpty()) {
+    return "[]";
+  }
+
+  std::stringstream ss;
+  ss << "[";
+  for (size_t i = 0; i < returnStates.size(); i++) {
+    if (i > 0) {
+      ss << ", ";
+    }
+    if (returnStates[i] == EMPTY_RETURN_STATE) {
+      ss << "$";
+      continue;
+    }
+    ss << returnStates[i];
+    if (parents[i] != nullptr) {
+      ss << " " << parents[i]->toString();
+    } else {
+      ss << "nul";
+    }
+  }
+  ss << "]";
+  return ss.str();
+}

runtime-linux/antlr4-runtime/atn/ArrayPredictionContext.h

@@ -0,0 +1,43 @@
+
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/PredictionContext.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class SingletonPredictionContext;
+
+  class ANTLR4CPP_PUBLIC ArrayPredictionContext : public PredictionContext {
+  public:
+    /// Parent can be empty only if full ctx mode and we make an array
+    /// from EMPTY and non-empty. We merge EMPTY by using null parent and
+    /// returnState == EMPTY_RETURN_STATE.
+    // Also here: we use a strong reference to our parents to avoid having them freed prematurely.
+    //            See also SinglePredictionContext.
+    const std::vector<Ref<PredictionContext>> parents;
+
+    /// Sorted for merge, no duplicates; if present, EMPTY_RETURN_STATE is always last.
+    const std::vector<size_t> returnStates;
+
+    ArrayPredictionContext(Ref<SingletonPredictionContext> const& a);
+    ArrayPredictionContext(std::vector<Ref<PredictionContext>> const& parents_, std::vector<size_t> const& returnStates);
+    virtual ~ArrayPredictionContext();
+
+    virtual bool isEmpty() const override;
+    virtual size_t size() const override;
+    virtual Ref<PredictionContext> getParent(size_t index) const override;
+    virtual size_t getReturnState(size_t index) const override;
+    bool operator == (const PredictionContext &o) const override;
+
+    virtual std::string toString() const override;
+  };
+
+} // namespace atn
+} // namespace antlr4
+

runtime-linux/antlr4-runtime/atn/AtomTransition.cpp

@@ -0,0 +1,31 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/IntervalSet.h"
+#include "atn/Transition.h"
+
+#include "atn/AtomTransition.h"
+
+using namespace antlr4::misc;
+using namespace antlr4::atn;
+
+AtomTransition::AtomTransition(ATNState *target, size_t label) : Transition(target), _label(label) {
+}
+
+Transition::SerializationType AtomTransition::getSerializationType() const {
+  return ATOM;
+}
+
+IntervalSet AtomTransition::label() const {
+  return IntervalSet::of((int)_label);
+}
+
+bool AtomTransition::matches(size_t symbol, size_t /*minVocabSymbol*/, size_t /*maxVocabSymbol*/) const {
+  return _label == symbol;
+}
+
+std::string AtomTransition::toString() const {
+  return "ATOM " + Transition::toString() + " { label: " + std::to_string(_label) + " }";
+}

runtime-linux/antlr4-runtime/atn/AtomTransition.h

@@ -0,0 +1,30 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/Transition.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// TO_DO: make all transitions sets? no, should remove set edges.
+  class ANTLR4CPP_PUBLIC AtomTransition final : public Transition {
+  public:
+    /// The token type or character value; or, signifies special label.
+    const size_t _label;
+
+    AtomTransition(ATNState *target, size_t label);
+
+    virtual SerializationType getSerializationType() const override;
+
+    virtual misc::IntervalSet label() const override;
+    virtual bool matches(size_t symbol, size_t minVocabSymbol, size_t maxVocabSymbol) const override;
+
+    virtual std::string toString() const override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/BasicBlockStartState.cpp

@@ -0,0 +1,12 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/BasicBlockStartState.h"
+
+using namespace antlr4::atn;
+
+size_t BasicBlockStartState::getStateType() {
+  return BLOCK_START;
+}

runtime-linux/antlr4-runtime/atn/BasicBlockStartState.h

@@ -0,0 +1,22 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "antlr4-common.h"
+#include "atn/BlockStartState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC BasicBlockStartState final : public BlockStartState {
+
+  public:
+    virtual size_t getStateType() override;
+
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/BasicState.cpp

@@ -0,0 +1,12 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/BasicState.h"
+
+using namespace antlr4::atn;
+
+size_t BasicState::getStateType() {
+  return BASIC;
+}

runtime-linux/antlr4-runtime/atn/BasicState.h

@@ -0,0 +1,21 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATNState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC BasicState final : public ATNState {
+
+  public:
+    virtual size_t getStateType() override;
+
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/BlockEndState.cpp

@@ -0,0 +1,15 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/BlockEndState.h"
+
+using namespace antlr4::atn;
+
+BlockEndState::BlockEndState() : startState(nullptr) {
+}
+
+size_t BlockEndState::getStateType() {
+  return BLOCK_END;
+}

runtime-linux/antlr4-runtime/atn/BlockEndState.h

@@ -0,0 +1,24 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATNState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Terminal node of a simple {@code (a|b|c)} block.
+  class ANTLR4CPP_PUBLIC BlockEndState final : public ATNState {
+  public:
+    BlockStartState *startState = nullptr;
+
+    BlockEndState();
+
+    virtual size_t getStateType() override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/BlockStartState.cpp

@@ -0,0 +1,9 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "BlockStartState.h"
+
+antlr4::atn::BlockStartState::~BlockStartState() {
+}

runtime-linux/antlr4-runtime/atn/BlockStartState.h

@@ -0,0 +1,21 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  ///  The start of a regular {@code (...)} block.
+  class ANTLR4CPP_PUBLIC BlockStartState : public DecisionState {
+  public:
+    ~BlockStartState();
+    BlockEndState *endState = nullptr;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ContextSensitivityInfo.cpp

@@ -0,0 +1,14 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ContextSensitivityInfo.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+
+ContextSensitivityInfo::ContextSensitivityInfo(size_t decision, ATNConfigSet *configs, TokenStream *input,
+  size_t startIndex, size_t stopIndex)
+  : DecisionEventInfo(decision, configs, input, startIndex, stopIndex, true) {
+}

runtime-linux/antlr4-runtime/atn/ContextSensitivityInfo.h

@@ -0,0 +1,47 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionEventInfo.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// This class represents profiling event information for a context sensitivity.
+  /// Context sensitivities are decisions where a particular input resulted in an
+  /// SLL conflict, but LL prediction produced a single unique alternative.
+  ///
+  /// <para>
+  /// In some cases, the unique alternative identified by LL prediction is not
+  /// equal to the minimum represented alternative in the conflicting SLL
+  /// configuration set. Grammars and inputs which result in this scenario are
+  /// unable to use <seealso cref="PredictionMode#SLL"/>, which in turn means they cannot use
+  /// the two-stage parsing strategy to improve parsing performance for that
+  /// input.</para>
+  /// </summary>
+  /// <seealso cref= ParserATNSimulator#reportContextSensitivity </seealso>
+  /// <seealso cref= ANTLRErrorListener#reportContextSensitivity
+  ///
+  /// @since 4.3 </seealso>
+  class ANTLR4CPP_PUBLIC ContextSensitivityInfo : public DecisionEventInfo {
+  public:
+    /// <summary>
+    /// Constructs a new instance of the <seealso cref="ContextSensitivityInfo"/> class
+    /// with the specified detailed context sensitivity information.
+    /// </summary>
+    /// <param name="decision"> The decision number </param>
+    /// <param name="configs"> The final configuration set containing the unique
+    /// alternative identified by full-context prediction </param>
+    /// <param name="input"> The input token stream </param>
+    /// <param name="startIndex"> The start index for the current prediction </param>
+    /// <param name="stopIndex"> The index at which the context sensitivity was
+    /// identified during full-context prediction </param>
+    ContextSensitivityInfo(size_t decision, ATNConfigSet *configs, TokenStream *input, size_t startIndex, size_t stopIndex);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/DecisionEventInfo.cpp

@@ -0,0 +1,14 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/DecisionEventInfo.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+
+DecisionEventInfo::DecisionEventInfo(size_t decision, ATNConfigSet *configs, TokenStream *input, size_t startIndex,
+  size_t stopIndex, bool fullCtx)
+  : decision(decision), configs(configs), input(input), startIndex(startIndex), stopIndex(stopIndex), fullCtx(fullCtx) {
+}

runtime-linux/antlr4-runtime/atn/DecisionEventInfo.h

@@ -0,0 +1,70 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "antlr4-common.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// This is the base class for gathering detailed information about prediction
+  /// events which occur during parsing.
+  ///
+  /// Note that we could record the parser call stack at the time this event
+  /// occurred but in the presence of left recursive rules, the stack is kind of
+  /// meaningless. It's better to look at the individual configurations for their
+  /// individual stacks. Of course that is a <seealso cref="PredictionContext"/> object
+  /// not a parse tree node and so it does not have information about the extent
+  /// (start...stop) of the various subtrees. Examining the stack tops of all
+  /// configurations provide the return states for the rule invocations.
+  /// From there you can get the enclosing rule.
+  ///
+  /// @since 4.3
+  /// </summary>
+  class ANTLR4CPP_PUBLIC DecisionEventInfo {
+  public:
+    /// <summary>
+    /// The invoked decision number which this event is related to.
+    /// </summary>
+    /// <seealso cref= ATN#decisionToState </seealso>
+    const size_t decision;
+
+    /// <summary>
+    /// The configuration set containing additional information relevant to the
+    /// prediction state when the current event occurred, or {@code null} if no
+    /// additional information is relevant or available.
+    /// </summary>
+    const ATNConfigSet *configs;
+
+    /// <summary>
+    /// The input token stream which is being parsed.
+    /// </summary>
+    const TokenStream *input;
+
+    /// <summary>
+    /// The token index in the input stream at which the current prediction was
+    /// originally invoked.
+    /// </summary>
+    const size_t startIndex;
+
+    /// <summary>
+    /// The token index in the input stream at which the current event occurred.
+    /// </summary>
+    const size_t stopIndex;
+
+    /// <summary>
+    /// {@code true} if the current event occurred during LL prediction;
+    /// otherwise, {@code false} if the input occurred during SLL prediction.
+    /// </summary>
+    const bool fullCtx;
+
+    DecisionEventInfo(size_t decision, ATNConfigSet *configs, TokenStream *input, size_t startIndex,
+                      size_t stopIndex, bool fullCtx);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/DecisionInfo.cpp

@@ -0,0 +1,25 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ErrorInfo.h"
+#include "atn/LookaheadEventInfo.h"
+
+#include "atn/DecisionInfo.h"
+
+using namespace antlr4::atn;
+
+DecisionInfo::DecisionInfo(size_t decision) : decision(decision) {
+}
+
+std::string DecisionInfo::toString() const {
+  std::stringstream ss;
+
+  ss << "{decision=" << decision << ", contextSensitivities=" << contextSensitivities.size() << ", errors=";
+  ss << errors.size() << ", ambiguities=" << ambiguities.size() << ", SLL_lookahead=" << SLL_TotalLook;
+  ss << ", SLL_ATNTransitions=" << SLL_ATNTransitions << ", SLL_DFATransitions=" << SLL_DFATransitions;
+  ss << ", LL_Fallback=" << LL_Fallback << ", LL_lookahead=" << LL_TotalLook << ", LL_ATNTransitions=" << LL_ATNTransitions << '}';
+
+  return ss.str();
+}

runtime-linux/antlr4-runtime/atn/DecisionInfo.h

@@ -0,0 +1,227 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ContextSensitivityInfo.h"
+#include "atn/AmbiguityInfo.h"
+#include "atn/PredicateEvalInfo.h"
+#include "atn/ErrorInfo.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class LookaheadEventInfo;
+
+  /// <summary>
+  /// This class contains profiling gathered for a particular decision.
+  ///
+  /// <para>
+  /// Parsing performance in ANTLR 4 is heavily influenced by both static factors
+  /// (e.g. the form of the rules in the grammar) and dynamic factors (e.g. the
+  /// choice of input and the state of the DFA cache at the time profiling
+  /// operations are started). For best results, gather and use aggregate
+  /// statistics from a large sample of inputs representing the inputs expected in
+  /// production before using the results to make changes in the grammar.</para>
+  ///
+  /// @since 4.3
+  /// </summary>
+  class ANTLR4CPP_PUBLIC DecisionInfo {
+  public:
+    /// <summary>
+    /// The decision number, which is an index into <seealso cref="ATN#decisionToState"/>.
+    /// </summary>
+    const size_t decision;
+
+    /// <summary>
+    /// The total number of times <seealso cref="ParserATNSimulator#adaptivePredict"/> was
+    /// invoked for this decision.
+    /// </summary>
+    long long invocations = 0;
+
+    /// <summary>
+    /// The total time spent in <seealso cref="ParserATNSimulator#adaptivePredict"/> for
+    /// this decision, in nanoseconds.
+    ///
+    /// <para>
+    /// The value of this field contains the sum of differential results obtained
+    /// by <seealso cref="System#nanoTime()"/>, and is not adjusted to compensate for JIT
+    /// and/or garbage collection overhead. For best accuracy, use a modern JVM
+    /// implementation that provides precise results from
+    /// <seealso cref="System#nanoTime()"/>, and perform profiling in a separate process
+    /// which is warmed up by parsing the input prior to profiling. If desired,
+    /// call <seealso cref="ATNSimulator#clearDFA"/> to reset the DFA cache to its initial
+    /// state before starting the profiling measurement pass.</para>
+    /// </summary>
+    long long timeInPrediction = 0;
+
+    /// <summary>
+    /// The sum of the lookahead required for SLL prediction for this decision.
+    /// Note that SLL prediction is used before LL prediction for performance
+    /// reasons even when <seealso cref="PredictionMode#LL"/> or
+    /// <seealso cref="PredictionMode#LL_EXACT_AMBIG_DETECTION"/> is used.
+    /// </summary>
+    long long SLL_TotalLook = 0;
+
+    /// <summary>
+    /// Gets the minimum lookahead required for any single SLL prediction to
+    /// complete for this decision, by reaching a unique prediction, reaching an
+    /// SLL conflict state, or encountering a syntax error.
+    /// </summary>
+    long long SLL_MinLook = 0;
+
+    /// <summary>
+    /// Gets the maximum lookahead required for any single SLL prediction to
+    /// complete for this decision, by reaching a unique prediction, reaching an
+    /// SLL conflict state, or encountering a syntax error.
+    /// </summary>
+    long long SLL_MaxLook = 0;
+
+    /// Gets the <seealso cref="LookaheadEventInfo"/> associated with the event where the
+    /// <seealso cref="#SLL_MaxLook"/> value was set.
+    Ref<LookaheadEventInfo> SLL_MaxLookEvent;
+
+    /// <summary>
+    /// The sum of the lookahead required for LL prediction for this decision.
+    /// Note that LL prediction is only used when SLL prediction reaches a
+    /// conflict state.
+    /// </summary>
+    long long LL_TotalLook = 0;
+
+    /// <summary>
+    /// Gets the minimum lookahead required for any single LL prediction to
+    /// complete for this decision. An LL prediction completes when the algorithm
+    /// reaches a unique prediction, a conflict state (for
+    /// <seealso cref="PredictionMode#LL"/>, an ambiguity state (for
+    /// <seealso cref="PredictionMode#LL_EXACT_AMBIG_DETECTION"/>, or a syntax error.
+    /// </summary>
+    long long LL_MinLook = 0;
+
+    /// <summary>
+    /// Gets the maximum lookahead required for any single LL prediction to
+    /// complete for this decision. An LL prediction completes when the algorithm
+    /// reaches a unique prediction, a conflict state (for
+    /// <seealso cref="PredictionMode#LL"/>, an ambiguity state (for
+    /// <seealso cref="PredictionMode#LL_EXACT_AMBIG_DETECTION"/>, or a syntax error.
+    /// </summary>
+    long long LL_MaxLook = 0;
+
+    /// <summary>
+    /// Gets the <seealso cref="LookaheadEventInfo"/> associated with the event where the
+    /// <seealso cref="#LL_MaxLook"/> value was set.
+    /// </summary>
+    Ref<LookaheadEventInfo> LL_MaxLookEvent;
+
+    /// <summary>
+    /// A collection of <seealso cref="ContextSensitivityInfo"/> instances describing the
+    /// context sensitivities encountered during LL prediction for this decision.
+    /// </summary>
+    /// <seealso cref= ContextSensitivityInfo </seealso>
+    std::vector<ContextSensitivityInfo> contextSensitivities;
+
+    /// <summary>
+    /// A collection of <seealso cref="ErrorInfo"/> instances describing the parse errors
+    /// identified during calls to <seealso cref="ParserATNSimulator#adaptivePredict"/> for
+    /// this decision.
+    /// </summary>
+    /// <seealso cref= ErrorInfo </seealso>
+    std::vector<ErrorInfo> errors;
+
+    /// <summary>
+    /// A collection of <seealso cref="AmbiguityInfo"/> instances describing the
+    /// ambiguities encountered during LL prediction for this decision.
+    /// </summary>
+    /// <seealso cref= AmbiguityInfo </seealso>
+    std::vector<AmbiguityInfo> ambiguities;
+
+    /// <summary>
+    /// A collection of <seealso cref="PredicateEvalInfo"/> instances describing the
+    /// results of evaluating individual predicates during prediction for this
+    /// decision.
+    /// </summary>
+    /// <seealso cref= PredicateEvalInfo </seealso>
+    std::vector<PredicateEvalInfo> predicateEvals;
+
+    /// <summary>
+    /// The total number of ATN transitions required during SLL prediction for
+    /// this decision. An ATN transition is determined by the number of times the
+    /// DFA does not contain an edge that is required for prediction, resulting
+    /// in on-the-fly computation of that edge.
+    ///
+    /// <para>
+    /// If DFA caching of SLL transitions is employed by the implementation, ATN
+    /// computation may cache the computed edge for efficient lookup during
+    /// future parsing of this decision. Otherwise, the SLL parsing algorithm
+    /// will use ATN transitions exclusively.</para>
+    /// </summary>
+    /// <seealso cref= #SLL_ATNTransitions </seealso>
+    /// <seealso cref= ParserATNSimulator#computeTargetState </seealso>
+    /// <seealso cref= LexerATNSimulator#computeTargetState </seealso>
+    long long SLL_ATNTransitions = 0;
+
+    /// <summary>
+    /// The total number of DFA transitions required during SLL prediction for
+    /// this decision.
+    ///
+    /// <para>If the ATN simulator implementation does not use DFA caching for SLL
+    /// transitions, this value will be 0.</para>
+    /// </summary>
+    /// <seealso cref= ParserATNSimulator#getExistingTargetState </seealso>
+    /// <seealso cref= LexerATNSimulator#getExistingTargetState </seealso>
+    long long SLL_DFATransitions = 0;
+
+    /// <summary>
+    /// Gets the total number of times SLL prediction completed in a conflict
+    /// state, resulting in fallback to LL prediction.
+    ///
+    /// <para>Note that this value is not related to whether or not
+    /// <seealso cref="PredictionMode#SLL"/> may be used successfully with a particular
+    /// grammar. If the ambiguity resolution algorithm applied to the SLL
+    /// conflicts for this decision produce the same result as LL prediction for
+    /// this decision, <seealso cref="PredictionMode#SLL"/> would produce the same overall
+    /// parsing result as <seealso cref="PredictionMode#LL"/>.</para>
+    /// </summary>
+    long long LL_Fallback = 0;
+
+    /// <summary>
+    /// The total number of ATN transitions required during LL prediction for
+    /// this decision. An ATN transition is determined by the number of times the
+    /// DFA does not contain an edge that is required for prediction, resulting
+    /// in on-the-fly computation of that edge.
+    ///
+    /// <para>
+    /// If DFA caching of LL transitions is employed by the implementation, ATN
+    /// computation may cache the computed edge for efficient lookup during
+    /// future parsing of this decision. Otherwise, the LL parsing algorithm will
+    /// use ATN transitions exclusively.</para>
+    /// </summary>
+    /// <seealso cref= #LL_DFATransitions </seealso>
+    /// <seealso cref= ParserATNSimulator#computeTargetState </seealso>
+    /// <seealso cref= LexerATNSimulator#computeTargetState </seealso>
+    long long LL_ATNTransitions = 0;
+
+    /// <summary>
+    /// The total number of DFA transitions required during LL prediction for
+    /// this decision.
+    ///
+    /// <para>If the ATN simulator implementation does not use DFA caching for LL
+    /// transitions, this value will be 0.</para>
+    /// </summary>
+    /// <seealso cref= ParserATNSimulator#getExistingTargetState </seealso>
+    /// <seealso cref= LexerATNSimulator#getExistingTargetState </seealso>
+    long long LL_DFATransitions = 0;
+
+    /// <summary>
+    /// Constructs a new instance of the <seealso cref="DecisionInfo"/> class to contain
+    /// statistics for a particular decision.
+    /// </summary>
+    /// <param name="decision"> The decision number </param>
+    DecisionInfo(size_t decision);
+
+    std::string toString() const;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/DecisionState.cpp

@@ -0,0 +1,17 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/DecisionState.h"
+
+using namespace antlr4::atn;
+
+void DecisionState::InitializeInstanceFields() {
+  decision = -1;
+  nonGreedy = false;
+}
+
+std::string DecisionState::toString() const {
+  return "DECISION " + ATNState::toString();
+}

runtime-linux/antlr4-runtime/atn/DecisionState.h

@@ -0,0 +1,30 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATNState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC DecisionState : public ATNState {
+  public:
+    int decision;
+    bool nonGreedy;
+
+  private:
+    void InitializeInstanceFields();
+
+  public:
+    DecisionState() {
+      InitializeInstanceFields();
+    }
+
+    virtual std::string toString() const override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/EmptyPredictionContext.cpp

@@ -0,0 +1,35 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/EmptyPredictionContext.h"
+
+using namespace antlr4::atn;
+
+EmptyPredictionContext::EmptyPredictionContext() : SingletonPredictionContext(nullptr, EMPTY_RETURN_STATE) {
+}
+
+bool EmptyPredictionContext::isEmpty() const {
+  return true;
+}
+
+size_t EmptyPredictionContext::size() const {
+  return 1;
+}
+
+Ref<PredictionContext> EmptyPredictionContext::getParent(size_t /*index*/) const {
+  return nullptr;
+}
+
+size_t EmptyPredictionContext::getReturnState(size_t /*index*/) const {
+  return returnState;
+}
+
+bool EmptyPredictionContext::operator == (const PredictionContext &o) const {
+  return this == &o;
+}
+
+std::string EmptyPredictionContext::toString() const {
+  return "$";
+}

runtime-linux/antlr4-runtime/atn/EmptyPredictionContext.h

@@ -0,0 +1,27 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/SingletonPredictionContext.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC EmptyPredictionContext : public SingletonPredictionContext {
+  public:
+    EmptyPredictionContext();
+
+    virtual bool isEmpty() const override;
+    virtual size_t size() const override;
+    virtual Ref<PredictionContext> getParent(size_t index) const override;
+    virtual size_t getReturnState(size_t index) const override;
+    virtual std::string toString() const override;
+
+    virtual bool operator == (const PredictionContext &o) const override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/EpsilonTransition.cpp

@@ -0,0 +1,35 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/EpsilonTransition.h"
+
+using namespace antlr4::atn;
+
+EpsilonTransition::EpsilonTransition(ATNState *target) : EpsilonTransition(target, INVALID_INDEX) {
+}
+
+EpsilonTransition::EpsilonTransition(ATNState *target, size_t outermostPrecedenceReturn)
+  : Transition(target), _outermostPrecedenceReturn(outermostPrecedenceReturn) {
+}
+
+size_t EpsilonTransition::outermostPrecedenceReturn() {
+  return _outermostPrecedenceReturn;
+}
+
+Transition::SerializationType EpsilonTransition::getSerializationType() const {
+  return EPSILON;
+}
+
+bool EpsilonTransition::isEpsilon() const {
+  return true;
+}
+
+bool EpsilonTransition::matches(size_t /*symbol*/, size_t /*minVocabSymbol*/, size_t /*maxVocabSymbol*/) const {
+  return false;
+}
+
+std::string EpsilonTransition::toString() const {
+  return "EPSILON " + Transition::toString() + " {}";
+}

runtime-linux/antlr4-runtime/atn/EpsilonTransition.h

@@ -0,0 +1,39 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/Transition.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC EpsilonTransition final : public Transition {
+  public:
+    EpsilonTransition(ATNState *target);
+    EpsilonTransition(ATNState *target, size_t outermostPrecedenceReturn);
+
+    /**
+     * @return the rule index of a precedence rule for which this transition is
+     * returning from, where the precedence value is 0; otherwise, INVALID_INDEX.
+     *
+     * @see ATNConfig#isPrecedenceFilterSuppressed()
+     * @see ParserATNSimulator#applyPrecedenceFilter(ATNConfigSet)
+     * @since 4.4.1
+     */
+    size_t outermostPrecedenceReturn();
+    virtual SerializationType getSerializationType() const override;
+
+    virtual bool isEpsilon() const override;
+    virtual bool matches(size_t symbol, size_t minVocabSymbol, size_t maxVocabSymbol) const override;
+
+    virtual std::string toString() const override;
+
+  private:
+    const size_t _outermostPrecedenceReturn; // A rule index.
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ErrorInfo.cpp

@@ -0,0 +1,15 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ATNConfigSet.h"
+
+#include "atn/ErrorInfo.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+
+ErrorInfo::ErrorInfo(size_t decision, ATNConfigSet *configs, TokenStream *input, size_t startIndex, size_t stopIndex, bool fullCtx)
+  : DecisionEventInfo(decision, configs, input, startIndex, stopIndex, fullCtx) {
+}

runtime-linux/antlr4-runtime/atn/ErrorInfo.h

@@ -0,0 +1,43 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionEventInfo.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// This class represents profiling event information for a syntax error
+  /// identified during prediction. Syntax errors occur when the prediction
+  /// algorithm is unable to identify an alternative which would lead to a
+  /// successful parse.
+  /// </summary>
+  /// <seealso cref= Parser#notifyErrorListeners(Token, String, RecognitionException) </seealso>
+  /// <seealso cref= ANTLRErrorListener#syntaxError
+  ///
+  /// @since 4.3 </seealso>
+  class ANTLR4CPP_PUBLIC ErrorInfo : public DecisionEventInfo {
+  public:
+    /// <summary>
+    /// Constructs a new instance of the <seealso cref="ErrorInfo"/> class with the
+    /// specified detailed syntax error information.
+    /// </summary>
+    /// <param name="decision"> The decision number </param>
+    /// <param name="configs"> The final configuration set reached during prediction
+    /// prior to reaching the <seealso cref="ATNSimulator#ERROR"/> state </param>
+    /// <param name="input"> The input token stream </param>
+    /// <param name="startIndex"> The start index for the current prediction </param>
+    /// <param name="stopIndex"> The index at which the syntax error was identified </param>
+    /// <param name="fullCtx"> {@code true} if the syntax error was identified during LL
+    /// prediction; otherwise, {@code false} if the syntax error was identified
+    /// during SLL prediction </param>
+    ErrorInfo(size_t decision, ATNConfigSet *configs, TokenStream *input, size_t startIndex, size_t stopIndex,
+              bool fullCtx);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LL1Analyzer.cpp

@@ -0,0 +1,158 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/RuleStopState.h"
+#include "atn/Transition.h"
+#include "atn/RuleTransition.h"
+#include "atn/SingletonPredictionContext.h"
+#include "atn/AbstractPredicateTransition.h"
+#include "atn/WildcardTransition.h"
+#include "atn/NotSetTransition.h"
+#include "misc/IntervalSet.h"
+#include "atn/ATNConfig.h"
+#include "atn/EmptyPredictionContext.h"
+
+#include "support/CPPUtils.h"
+
+#include "atn/LL1Analyzer.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlrcpp;
+
+LL1Analyzer::LL1Analyzer(const ATN &atn) : _atn(atn) {
+}
+
+LL1Analyzer::~LL1Analyzer() {
+}
+
+std::vector<misc::IntervalSet> LL1Analyzer::getDecisionLookahead(ATNState *s) const {
+  std::vector<misc::IntervalSet> look;
+
+  if (s == nullptr) {
+    return look;
+  }
+
+  look.resize(s->transitions.size()); // Fills all interval sets with defaults.
+  for (size_t alt = 0; alt < s->transitions.size(); alt++) {
+    bool seeThruPreds = false; // fail to get lookahead upon pred
+
+    ATNConfig::Set lookBusy;
+    antlrcpp::BitSet callRuleStack;
+    _LOOK(s->transitions[alt]->target, nullptr, PredictionContext::EMPTY,
+          look[alt], lookBusy, callRuleStack, seeThruPreds, false);
+
+    // Wipe out lookahead for this alternative if we found nothing
+    // or we had a predicate when we !seeThruPreds
+    if (look[alt].size() == 0 || look[alt].contains(HIT_PRED)) {
+      look[alt].clear();
+    }
+  }
+  return look;
+}
+
+misc::IntervalSet LL1Analyzer::LOOK(ATNState *s, RuleContext *ctx) const {
+  return LOOK(s, nullptr, ctx);
+}
+
+misc::IntervalSet LL1Analyzer::LOOK(ATNState *s, ATNState *stopState, RuleContext *ctx) const {
+  misc::IntervalSet r;
+  bool seeThruPreds = true; // ignore preds; get all lookahead
+  Ref<PredictionContext> lookContext = ctx != nullptr ? PredictionContext::fromRuleContext(_atn, ctx) : nullptr;
+
+  ATNConfig::Set lookBusy;
+  antlrcpp::BitSet callRuleStack;
+  _LOOK(s, stopState, lookContext, r, lookBusy, callRuleStack, seeThruPreds, true);
+
+  return r;
+}
+
+void LL1Analyzer::_LOOK(ATNState *s, ATNState *stopState, Ref<PredictionContext> const& ctx, misc::IntervalSet &look,
+  ATNConfig::Set &lookBusy, antlrcpp::BitSet &calledRuleStack, bool seeThruPreds, bool addEOF) const {
+
+  Ref<ATNConfig> c = std::make_shared<ATNConfig>(s, 0, ctx);
+
+  if (lookBusy.count(c) > 0) // Keep in mind comparison is based on members of the class, not the actual instance.
+    return;
+
+  lookBusy.insert(c);
+
+  // ml: s can never be null, hence no need to check if stopState is != null.
+  if (s == stopState) {
+    if (ctx == nullptr) {
+      look.add(Token::EPSILON);
+      return;
+    } else if (ctx->isEmpty() && addEOF) {
+      look.add(Token::EOF);
+      return;
+    }
+  }
+
+  if (s->getStateType() == ATNState::RULE_STOP) {
+    if (ctx == nullptr) {
+      look.add(Token::EPSILON);
+      return;
+    } else if (ctx->isEmpty() && addEOF) {
+      look.add(Token::EOF);
+      return;
+    }
+
+    if (ctx != PredictionContext::EMPTY) {
+      // run thru all possible stack tops in ctx
+      for (size_t i = 0; i < ctx->size(); i++) {
+        ATNState *returnState = _atn.states[ctx->getReturnState(i)];
+
+        bool removed = calledRuleStack.test(returnState->ruleIndex);
+        auto onExit = finally([removed, &calledRuleStack, returnState] {
+          if (removed) {
+            calledRuleStack.set(returnState->ruleIndex);
+          }
+        });
+
+        calledRuleStack[returnState->ruleIndex] = false;
+        _LOOK(returnState, stopState, ctx->getParent(i), look, lookBusy, calledRuleStack, seeThruPreds, addEOF);
+      }
+      return;
+    }
+  }
+
+  size_t n = s->transitions.size();
+  for (size_t i = 0; i < n; i++) {
+    Transition *t = s->transitions[i];
+
+    if (t->getSerializationType() == Transition::RULE) {
+      if (calledRuleStack[(static_cast<RuleTransition*>(t))->target->ruleIndex]) {
+        continue;
+      }
+
+      Ref<PredictionContext> newContext = SingletonPredictionContext::create(ctx, (static_cast<RuleTransition*>(t))->followState->stateNumber);
+      auto onExit = finally([t, &calledRuleStack] {
+        calledRuleStack[(static_cast<RuleTransition*>(t))->target->ruleIndex] = false;
+      });
+
+      calledRuleStack.set((static_cast<RuleTransition*>(t))->target->ruleIndex);
+      _LOOK(t->target, stopState, newContext, look, lookBusy, calledRuleStack, seeThruPreds, addEOF);
+
+    } else if (is<AbstractPredicateTransition *>(t)) {
+      if (seeThruPreds) {
+        _LOOK(t->target, stopState, ctx, look, lookBusy, calledRuleStack, seeThruPreds, addEOF);
+      } else {
+        look.add(HIT_PRED);
+      }
+    } else if (t->isEpsilon()) {
+      _LOOK(t->target, stopState, ctx, look, lookBusy, calledRuleStack, seeThruPreds, addEOF);
+    } else if (t->getSerializationType() == Transition::WILDCARD) {
+      look.addAll(misc::IntervalSet::of(Token::MIN_USER_TOKEN_TYPE, static_cast<ssize_t>(_atn.maxTokenType)));
+    } else {
+      misc::IntervalSet set = t->label();
+      if (!set.isEmpty()) {
+        if (is<NotSetTransition*>(t)) {
+          set = set.complement(misc::IntervalSet::of(Token::MIN_USER_TOKEN_TYPE, static_cast<ssize_t>(_atn.maxTokenType)));
+        }
+        look.addAll(set);
+      }
+    }
+  }
+}

runtime-linux/antlr4-runtime/atn/LL1Analyzer.h

@@ -0,0 +1,109 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "Token.h"
+#include "support/BitSet.h"
+#include "atn/PredictionContext.h"
+#include "atn/ATNConfig.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC LL1Analyzer {
+  public:
+    /// Special value added to the lookahead sets to indicate that we hit
+    ///  a predicate during analysis if {@code seeThruPreds==false}.
+    static const size_t HIT_PRED = Token::INVALID_TYPE;
+
+    const atn::ATN &_atn;
+
+    LL1Analyzer(const atn::ATN &atn);
+    virtual ~LL1Analyzer();
+
+    /// <summary>
+    /// Calculates the SLL(1) expected lookahead set for each outgoing transition
+    /// of an <seealso cref="ATNState"/>. The returned array has one element for each
+    /// outgoing transition in {@code s}. If the closure from transition
+    /// <em>i</em> leads to a semantic predicate before matching a symbol, the
+    /// element at index <em>i</em> of the result will be {@code null}.
+    /// </summary>
+    /// <param name="s"> the ATN state </param>
+    /// <returns> the expected symbols for each outgoing transition of {@code s}. </returns>
+    virtual std::vector<misc::IntervalSet> getDecisionLookahead(ATNState *s) const;
+
+    /// <summary>
+    /// Compute set of tokens that can follow {@code s} in the ATN in the
+    /// specified {@code ctx}.
+    /// <p/>
+    /// If {@code ctx} is {@code null} and the end of the rule containing
+    /// {@code s} is reached, <seealso cref="Token#EPSILON"/> is added to the result set.
+    /// If {@code ctx} is not {@code null} and the end of the outermost rule is
+    /// reached, <seealso cref="Token#EOF"/> is added to the result set.
+    /// </summary>
+    /// <param name="s"> the ATN state </param>
+    /// <param name="ctx"> the complete parser context, or {@code null} if the context
+    /// should be ignored
+    /// </param>
+    /// <returns> The set of tokens that can follow {@code s} in the ATN in the
+    /// specified {@code ctx}. </returns>
+    virtual misc::IntervalSet LOOK(ATNState *s, RuleContext *ctx) const;
+
+    /// <summary>
+    /// Compute set of tokens that can follow {@code s} in the ATN in the
+    /// specified {@code ctx}.
+    /// <p/>
+    /// If {@code ctx} is {@code null} and the end of the rule containing
+    /// {@code s} is reached, <seealso cref="Token#EPSILON"/> is added to the result set.
+    /// If {@code ctx} is not {@code null} and the end of the outermost rule is
+    /// reached, <seealso cref="Token#EOF"/> is added to the result set.
+    /// </summary>
+    /// <param name="s"> the ATN state </param>
+    /// <param name="stopState"> the ATN state to stop at. This can be a
+    /// <seealso cref="BlockEndState"/> to detect epsilon paths through a closure. </param>
+    /// <param name="ctx"> the complete parser context, or {@code null} if the context
+    /// should be ignored
+    /// </param>
+    /// <returns> The set of tokens that can follow {@code s} in the ATN in the
+    /// specified {@code ctx}. </returns>
+    virtual misc::IntervalSet LOOK(ATNState *s, ATNState *stopState, RuleContext *ctx) const;
+
+    /// <summary>
+    /// Compute set of tokens that can follow {@code s} in the ATN in the
+    /// specified {@code ctx}.
+    /// <p/>
+    /// If {@code ctx} is {@code null} and {@code stopState} or the end of the
+    /// rule containing {@code s} is reached, <seealso cref="Token#EPSILON"/> is added to
+    /// the result set. If {@code ctx} is not {@code null} and {@code addEOF} is
+    /// {@code true} and {@code stopState} or the end of the outermost rule is
+    /// reached, <seealso cref="Token#EOF"/> is added to the result set.
+    /// </summary>
+    /// <param name="s"> the ATN state. </param>
+    /// <param name="stopState"> the ATN state to stop at. This can be a
+    /// <seealso cref="BlockEndState"/> to detect epsilon paths through a closure. </param>
+    /// <param name="ctx"> The outer context, or {@code null} if the outer context should
+    /// not be used. </param>
+    /// <param name="look"> The result lookahead set. </param>
+    /// <param name="lookBusy"> A set used for preventing epsilon closures in the ATN
+    /// from causing a stack overflow. Outside code should pass
+    /// {@code new HashSet<ATNConfig>} for this argument. </param>
+    /// <param name="calledRuleStack"> A set used for preventing left recursion in the
+    /// ATN from causing a stack overflow. Outside code should pass
+    /// {@code new BitSet()} for this argument. </param>
+    /// <param name="seeThruPreds"> {@code true} to true semantic predicates as
+    /// implicitly {@code true} and "see through them", otherwise {@code false}
+    /// to treat semantic predicates as opaque and add <seealso cref="#HIT_PRED"/> to the
+    /// result if one is encountered. </param>
+    /// <param name="addEOF"> Add <seealso cref="Token#EOF"/> to the result if the end of the
+    /// outermost context is reached. This parameter has no effect if {@code ctx}
+    /// is {@code null}. </param>
+  protected:
+    virtual void _LOOK(ATNState *s, ATNState *stopState, Ref<PredictionContext> const& ctx, misc::IntervalSet &look,
+      ATNConfig::Set &lookBusy, antlrcpp::BitSet &calledRuleStack, bool seeThruPreds, bool addEOF) const;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerATNConfig.cpp

@@ -0,0 +1,84 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "atn/DecisionState.h"
+#include "atn/PredictionContext.h"
+#include "SemanticContext.h"
+#include "atn/LexerActionExecutor.h"
+
+#include "support/CPPUtils.h"
+
+#include "atn/LexerATNConfig.h"
+
+using namespace antlr4::atn;
+using namespace antlrcpp;
+
+LexerATNConfig::LexerATNConfig(ATNState *state, int alt, Ref<PredictionContext> const& context)
+  : ATNConfig(state, alt, context, SemanticContext::NONE), _passedThroughNonGreedyDecision(false) {
+}
+
+LexerATNConfig::LexerATNConfig(ATNState *state, int alt, Ref<PredictionContext> const& context,
+                               Ref<LexerActionExecutor> const& lexerActionExecutor)
+  : ATNConfig(state, alt, context, SemanticContext::NONE), _lexerActionExecutor(lexerActionExecutor),
+    _passedThroughNonGreedyDecision(false) {
+}
+
+LexerATNConfig::LexerATNConfig(Ref<LexerATNConfig> const& c, ATNState *state)
+  : ATNConfig(c, state, c->context, c->semanticContext), _lexerActionExecutor(c->_lexerActionExecutor),
+   _passedThroughNonGreedyDecision(checkNonGreedyDecision(c, state)) {
+}
+
+LexerATNConfig::LexerATNConfig(Ref<LexerATNConfig> const& c, ATNState *state, Ref<LexerActionExecutor> const& lexerActionExecutor)
+  : ATNConfig(c, state, c->context, c->semanticContext), _lexerActionExecutor(lexerActionExecutor),
+    _passedThroughNonGreedyDecision(checkNonGreedyDecision(c, state)) {
+}
+
+LexerATNConfig::LexerATNConfig(Ref<LexerATNConfig> const& c, ATNState *state, Ref<PredictionContext> const& context)
+  : ATNConfig(c, state, context, c->semanticContext), _lexerActionExecutor(c->_lexerActionExecutor),
+    _passedThroughNonGreedyDecision(checkNonGreedyDecision(c, state)) {
+}
+
+Ref<LexerActionExecutor> LexerATNConfig::getLexerActionExecutor() const {
+  return _lexerActionExecutor;
+}
+
+bool LexerATNConfig::hasPassedThroughNonGreedyDecision() {
+  return _passedThroughNonGreedyDecision;
+}
+
+size_t LexerATNConfig::hashCode() const {
+  size_t hashCode = misc::MurmurHash::initialize(7);
+  hashCode = misc::MurmurHash::update(hashCode, state->stateNumber);
+  hashCode = misc::MurmurHash::update(hashCode, alt);
+  hashCode = misc::MurmurHash::update(hashCode, context);
+  hashCode = misc::MurmurHash::update(hashCode, semanticContext);
+  hashCode = misc::MurmurHash::update(hashCode, _passedThroughNonGreedyDecision ? 1 : 0);
+  hashCode = misc::MurmurHash::update(hashCode, _lexerActionExecutor);
+  hashCode = misc::MurmurHash::finish(hashCode, 6);
+  return hashCode;
+}
+
+bool LexerATNConfig::operator == (const LexerATNConfig& other) const
+{
+  if (this == &other)
+    return true;
+
+  if (_passedThroughNonGreedyDecision != other._passedThroughNonGreedyDecision)
+    return false;
+
+  if (_lexerActionExecutor == nullptr)
+    return other._lexerActionExecutor == nullptr;
+  if (*_lexerActionExecutor != *(other._lexerActionExecutor)) {
+    return false;
+  }
+
+  return ATNConfig::operator == (other);
+}
+
+bool LexerATNConfig::checkNonGreedyDecision(Ref<LexerATNConfig> const& source, ATNState *target) {
+  return source->_passedThroughNonGreedyDecision ||
+    (is<DecisionState*>(target) && (static_cast<DecisionState*>(target))->nonGreedy);
+}

runtime-linux/antlr4-runtime/atn/LexerATNConfig.h

@@ -0,0 +1,44 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATNConfig.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC LexerATNConfig : public ATNConfig {
+  public:
+    LexerATNConfig(ATNState *state, int alt, Ref<PredictionContext> const& context);
+    LexerATNConfig(ATNState *state, int alt, Ref<PredictionContext> const& context, Ref<LexerActionExecutor> const& lexerActionExecutor);
+
+    LexerATNConfig(Ref<LexerATNConfig> const& c, ATNState *state);
+    LexerATNConfig(Ref<LexerATNConfig> const& c, ATNState *state, Ref<LexerActionExecutor> const& lexerActionExecutor);
+    LexerATNConfig(Ref<LexerATNConfig> const& c, ATNState *state, Ref<PredictionContext> const& context);
+
+    /**
+     * Gets the {@link LexerActionExecutor} capable of executing the embedded
+     * action(s) for the current configuration.
+     */
+    Ref<LexerActionExecutor> getLexerActionExecutor() const;
+    bool hasPassedThroughNonGreedyDecision();
+
+    virtual size_t hashCode() const override;
+
+    bool operator == (const LexerATNConfig& other) const;
+
+  private:
+    /**
+     * This is the backing field for {@link #getLexerActionExecutor}.
+     */
+    const Ref<LexerActionExecutor> _lexerActionExecutor;
+    const bool _passedThroughNonGreedyDecision;
+
+    static bool checkNonGreedyDecision(Ref<LexerATNConfig> const& source, ATNState *target);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerATNSimulator.cpp

@@ -0,0 +1,628 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "IntStream.h"
+#include "atn/OrderedATNConfigSet.h"
+#include "Token.h"
+#include "LexerNoViableAltException.h"
+#include "atn/RuleStopState.h"
+#include "atn/RuleTransition.h"
+#include "atn/SingletonPredictionContext.h"
+#include "atn/PredicateTransition.h"
+#include "atn/ActionTransition.h"
+#include "atn/TokensStartState.h"
+#include "misc/Interval.h"
+#include "dfa/DFA.h"
+#include "Lexer.h"
+
+#include "dfa/DFAState.h"
+#include "atn/LexerATNConfig.h"
+#include "atn/LexerActionExecutor.h"
+#include "atn/EmptyPredictionContext.h"
+
+#include "atn/LexerATNSimulator.h"
+
+#define DEBUG_ATN 0
+#define DEBUG_DFA 0
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlrcpp;
+
+LexerATNSimulator::SimState::~SimState() {
+}
+
+void LexerATNSimulator::SimState::reset() {
+  index = INVALID_INDEX;
+  line = 0;
+  charPos = INVALID_INDEX;
+  dfaState = nullptr; // Don't delete. It's just a reference.
+}
+
+void LexerATNSimulator::SimState::InitializeInstanceFields() {
+  index = INVALID_INDEX;
+  line = 0;
+  charPos = INVALID_INDEX;
+}
+
+int LexerATNSimulator::match_calls = 0;
+
+
+LexerATNSimulator::LexerATNSimulator(const ATN &atn, std::vector<dfa::DFA> &decisionToDFA,
+                                     PredictionContextCache &sharedContextCache)
+  : LexerATNSimulator(nullptr, atn, decisionToDFA, sharedContextCache) {
+}
+
+LexerATNSimulator::LexerATNSimulator(Lexer *recog, const ATN &atn, std::vector<dfa::DFA> &decisionToDFA,
+                                     PredictionContextCache &sharedContextCache)
+  : ATNSimulator(atn, sharedContextCache), _recog(recog), _decisionToDFA(decisionToDFA) {
+  InitializeInstanceFields();
+}
+
+void LexerATNSimulator::copyState(LexerATNSimulator *simulator) {
+  _charPositionInLine = simulator->_charPositionInLine;
+  _line = simulator->_line;
+  _mode = simulator->_mode;
+  _startIndex = simulator->_startIndex;
+}
+
+size_t LexerATNSimulator::match(CharStream *input, size_t mode) {
+  match_calls++;
+  _mode = mode;
+  ssize_t mark = input->mark();
+
+  auto onExit = finally([input, mark] {
+    input->release(mark);
+  });
+
+  _startIndex = input->index();
+  _prevAccept.reset();
+  const dfa::DFA &dfa = _decisionToDFA[mode];
+  if (dfa.s0 == nullptr) {
+    return matchATN(input);
+  } else {
+    return execATN(input, dfa.s0);
+  }
+}
+
+void LexerATNSimulator::reset() {
+  _prevAccept.reset();
+  _startIndex = 0;
+  _line = 1;
+  _charPositionInLine = 0;
+  _mode = Lexer::DEFAULT_MODE;
+}
+
+void LexerATNSimulator::clearDFA() {
+  size_t size = _decisionToDFA.size();
+  _decisionToDFA.clear();
+  for (size_t d = 0; d < size; ++d) {
+    _decisionToDFA.push_back(dfa::DFA(atn.getDecisionState(d), d));
+  }
+}
+
+size_t LexerATNSimulator::matchATN(CharStream *input) {
+  ATNState *startState = atn.modeToStartState[_mode];
+
+  std::unique_ptr<ATNConfigSet> s0_closure = computeStartState(input, startState);
+
+  bool suppressEdge = s0_closure->hasSemanticContext;
+  s0_closure->hasSemanticContext = false;
+
+  dfa::DFAState *next = addDFAState(s0_closure.release());
+  if (!suppressEdge) {
+    _decisionToDFA[_mode].s0 = next;
+  }
+
+  size_t predict = execATN(input, next);
+
+  return predict;
+}
+
+size_t LexerATNSimulator::execATN(CharStream *input, dfa::DFAState *ds0) {
+  if (ds0->isAcceptState) {
+    // allow zero-length tokens
+    // ml: in Java code this method uses 3 params. The first is a member var of the class anyway (_prevAccept), so why pass it here?
+    captureSimState(input, ds0);
+  }
+
+  size_t t = input->LA(1);
+  dfa::DFAState *s = ds0; // s is current/from DFA state
+
+  while (true) { // while more work
+    // As we move src->trg, src->trg, we keep track of the previous trg to
+    // avoid looking up the DFA state again, which is expensive.
+    // If the previous target was already part of the DFA, we might
+    // be able to avoid doing a reach operation upon t. If s!=null,
+    // it means that semantic predicates didn't prevent us from
+    // creating a DFA state. Once we know s!=null, we check to see if
+    // the DFA state has an edge already for t. If so, we can just reuse
+    // it's configuration set; there's no point in re-computing it.
+    // This is kind of like doing DFA simulation within the ATN
+    // simulation because DFA simulation is really just a way to avoid
+    // computing reach/closure sets. Technically, once we know that
+    // we have a previously added DFA state, we could jump over to
+    // the DFA simulator. But, that would mean popping back and forth
+    // a lot and making things more complicated algorithmically.
+    // This optimization makes a lot of sense for loops within DFA.
+    // A character will take us back to an existing DFA state
+    // that already has lots of edges out of it. e.g., .* in comments.
+    dfa::DFAState *target = getExistingTargetState(s, t);
+    if (target == nullptr) {
+      target = computeTargetState(input, s, t);
+    }
+
+    if (target == ERROR.get()) {
+      break;
+    }
+
+    // If this is a consumable input element, make sure to consume before
+    // capturing the accept state so the input index, line, and char
+    // position accurately reflect the state of the interpreter at the
+    // end of the token.
+    if (t != Token::EOF) {
+      consume(input);
+    }
+
+    if (target->isAcceptState) {
+      captureSimState(input, target);
+      if (t == Token::EOF) {
+        break;
+      }
+    }
+
+    t = input->LA(1);
+    s = target; // flip; current DFA target becomes new src/from state
+  }
+
+  return failOrAccept(input, s->configs.get(), t);
+}
+
+dfa::DFAState *LexerATNSimulator::getExistingTargetState(dfa::DFAState *s, size_t t) {
+  dfa::DFAState* retval = nullptr;
+  _edgeLock.readLock();
+  if (t <= MAX_DFA_EDGE) {
+    auto iterator = s->edges.find(t - MIN_DFA_EDGE);
+#if DEBUG_ATN == 1
+    if (iterator != s->edges.end()) {
+      std::cout << std::string("reuse state ") << s->stateNumber << std::string(" edge to ") << iterator->second->stateNumber << std::endl;
+    }
+#endif
+
+    if (iterator != s->edges.end())
+	retval = iterator->second;
+  }
+  _edgeLock.readUnlock();
+  return retval;
+}
+
+dfa::DFAState *LexerATNSimulator::computeTargetState(CharStream *input, dfa::DFAState *s, size_t t) {
+  OrderedATNConfigSet *reach = new OrderedATNConfigSet(); /* mem-check: deleted on error or managed by new DFA state. */
+
+  // if we don't find an existing DFA state
+  // Fill reach starting from closure, following t transitions
+  getReachableConfigSet(input, s->configs.get(), reach, t);
+
+  if (reach->isEmpty()) { // we got nowhere on t from s
+    if (!reach->hasSemanticContext) {
+      // we got nowhere on t, don't throw out this knowledge; it'd
+      // cause a failover from DFA later.
+      delete reach;
+      addDFAEdge(s, t, ERROR.get());
+    }
+
+    // stop when we can't match any more char
+    return ERROR.get();
+  }
+
+  // Add an edge from s to target DFA found/created for reach
+  return addDFAEdge(s, t, reach);
+}
+
+size_t LexerATNSimulator::failOrAccept(CharStream *input, ATNConfigSet *reach, size_t t) {
+  if (_prevAccept.dfaState != nullptr) {
+    Ref<LexerActionExecutor> lexerActionExecutor = _prevAccept.dfaState->lexerActionExecutor;
+    accept(input, lexerActionExecutor, _startIndex, _prevAccept.index, _prevAccept.line, _prevAccept.charPos);
+    return _prevAccept.dfaState->prediction;
+  } else {
+    // if no accept and EOF is first char, return EOF
+    if (t == Token::EOF && input->index() == _startIndex) {
+      return Token::EOF;
+    }
+
+    throw LexerNoViableAltException(_recog, input, _startIndex, reach);
+  }
+}
+
+void LexerATNSimulator::getReachableConfigSet(CharStream *input, ATNConfigSet *closure_, ATNConfigSet *reach, size_t t) {
+  // this is used to skip processing for configs which have a lower priority
+  // than a config that already reached an accept state for the same rule
+  size_t skipAlt = ATN::INVALID_ALT_NUMBER;
+
+  for (auto c : closure_->configs) {
+    bool currentAltReachedAcceptState = c->alt == skipAlt;
+    if (currentAltReachedAcceptState && (std::static_pointer_cast<LexerATNConfig>(c))->hasPassedThroughNonGreedyDecision()) {
+      continue;
+    }
+
+#if DEBUG_ATN == 1
+      std::cout << "testing " << getTokenName((int)t) << " at " << c->toString(true) << std::endl;
+#endif
+
+    size_t n = c->state->transitions.size();
+    for (size_t ti = 0; ti < n; ti++) { // for each transition
+      Transition *trans = c->state->transitions[ti];
+      ATNState *target = getReachableTarget(trans, (int)t);
+      if (target != nullptr) {
+        Ref<LexerActionExecutor> lexerActionExecutor = std::static_pointer_cast<LexerATNConfig>(c)->getLexerActionExecutor();
+        if (lexerActionExecutor != nullptr) {
+          lexerActionExecutor = lexerActionExecutor->fixOffsetBeforeMatch((int)input->index() - (int)_startIndex);
+        }
+
+        bool treatEofAsEpsilon = t == Token::EOF;
+        Ref<LexerATNConfig> config = std::make_shared<LexerATNConfig>(std::static_pointer_cast<LexerATNConfig>(c),
+          target, lexerActionExecutor);
+
+        if (closure(input, config, reach, currentAltReachedAcceptState, true, treatEofAsEpsilon)) {
+          // any remaining configs for this alt have a lower priority than
+          // the one that just reached an accept state.
+          skipAlt = c->alt;
+          break;
+        }
+      }
+    }
+  }
+}
+
+void LexerATNSimulator::accept(CharStream *input, const Ref<LexerActionExecutor> &lexerActionExecutor, size_t /*startIndex*/,
+                               size_t index, size_t line, size_t charPos) {
+#if DEBUG_ATN == 1
+    std::cout << "ACTION ";
+    std::cout << toString(lexerActionExecutor) << std::endl;
+#endif
+
+  // seek to after last char in token
+  input->seek(index);
+  _line = line;
+  _charPositionInLine = (int)charPos;
+
+  if (lexerActionExecutor != nullptr && _recog != nullptr) {
+    lexerActionExecutor->execute(_recog, input, _startIndex);
+  }
+}
+
+atn::ATNState *LexerATNSimulator::getReachableTarget(Transition *trans, size_t t) {
+  if (trans->matches(t, Lexer::MIN_CHAR_VALUE, Lexer::MAX_CHAR_VALUE)) {
+    return trans->target;
+  }
+
+  return nullptr;
+}
+
+std::unique_ptr<ATNConfigSet> LexerATNSimulator::computeStartState(CharStream *input, ATNState *p) {
+  Ref<PredictionContext> initialContext = PredictionContext::EMPTY; // ml: the purpose of this assignment is unclear
+  std::unique_ptr<ATNConfigSet> configs(new OrderedATNConfigSet());
+  for (size_t i = 0; i < p->transitions.size(); i++) {
+    ATNState *target = p->transitions[i]->target;
+    Ref<LexerATNConfig> c = std::make_shared<LexerATNConfig>(target, (int)(i + 1), initialContext);
+    closure(input, c, configs.get(), false, false, false);
+  }
+
+  return configs;
+}
+
+bool LexerATNSimulator::closure(CharStream *input, const Ref<LexerATNConfig> &config, ATNConfigSet *configs,
+                                bool currentAltReachedAcceptState, bool speculative, bool treatEofAsEpsilon) {
+#if DEBUG_ATN == 1
+    std::cout << "closure(" << config->toString(true) << ")" << std::endl;
+#endif
+
+  if (is<RuleStopState *>(config->state)) {
+#if DEBUG_ATN == 1
+      if (_recog != nullptr) {
+        std::cout << "closure at " << _recog->getRuleNames()[config->state->ruleIndex] << " rule stop " << config << std::endl;
+      } else {
+        std::cout << "closure at rule stop " << config << std::endl;
+      }
+#endif
+
+    if (config->context == nullptr || config->context->hasEmptyPath()) {
+      if (config->context == nullptr || config->context->isEmpty()) {
+        configs->add(config);
+        return true;
+      } else {
+        configs->add(std::make_shared<LexerATNConfig>(config, config->state, PredictionContext::EMPTY));
+        currentAltReachedAcceptState = true;
+      }
+    }
+
+    if (config->context != nullptr && !config->context->isEmpty()) {
+      for (size_t i = 0; i < config->context->size(); i++) {
+        if (config->context->getReturnState(i) != PredictionContext::EMPTY_RETURN_STATE) {
+          std::weak_ptr<PredictionContext> newContext = config->context->getParent(i); // "pop" return state
+          ATNState *returnState = atn.states[config->context->getReturnState(i)];
+          Ref<LexerATNConfig> c = std::make_shared<LexerATNConfig>(config, returnState, newContext.lock());
+          currentAltReachedAcceptState = closure(input, c, configs, currentAltReachedAcceptState, speculative, treatEofAsEpsilon);
+        }
+      }
+    }
+
+    return currentAltReachedAcceptState;
+  }
+
+  // optimization
+  if (!config->state->epsilonOnlyTransitions) {
+    if (!currentAltReachedAcceptState || !config->hasPassedThroughNonGreedyDecision()) {
+      configs->add(config);
+    }
+  }
+
+  ATNState *p = config->state;
+  for (size_t i = 0; i < p->transitions.size(); i++) {
+    Transition *t = p->transitions[i];
+    Ref<LexerATNConfig> c = getEpsilonTarget(input, config, t, configs, speculative, treatEofAsEpsilon);
+    if (c != nullptr) {
+      currentAltReachedAcceptState = closure(input, c, configs, currentAltReachedAcceptState, speculative, treatEofAsEpsilon);
+    }
+  }
+
+  return currentAltReachedAcceptState;
+}
+
+Ref<LexerATNConfig> LexerATNSimulator::getEpsilonTarget(CharStream *input, const Ref<LexerATNConfig> &config, Transition *t,
+  ATNConfigSet *configs, bool speculative, bool treatEofAsEpsilon) {
+
+  Ref<LexerATNConfig> c = nullptr;
+  switch (t->getSerializationType()) {
+    case Transition::RULE: {
+      RuleTransition *ruleTransition = static_cast<RuleTransition*>(t);
+      Ref<PredictionContext> newContext = SingletonPredictionContext::create(config->context, ruleTransition->followState->stateNumber);
+      c = std::make_shared<LexerATNConfig>(config, t->target, newContext);
+      break;
+    }
+
+    case Transition::PRECEDENCE:
+      throw UnsupportedOperationException("Precedence predicates are not supported in lexers.");
+
+    case Transition::PREDICATE: {
+      /*  Track traversing semantic predicates. If we traverse,
+       we cannot add a DFA state for this "reach" computation
+       because the DFA would not test the predicate again in the
+       future. Rather than creating collections of semantic predicates
+       like v3 and testing them on prediction, v4 will test them on the
+       fly all the time using the ATN not the DFA. This is slower but
+       semantically it's not used that often. One of the key elements to
+       this predicate mechanism is not adding DFA states that see
+       predicates immediately afterwards in the ATN. For example,
+
+       a : ID {p1}? | ID {p2}? ;
+
+       should create the start state for rule 'a' (to save start state
+       competition), but should not create target of ID state. The
+       collection of ATN states the following ID references includes
+       states reached by traversing predicates. Since this is when we
+       test them, we cannot cash the DFA state target of ID.
+       */
+      PredicateTransition *pt = static_cast<PredicateTransition*>(t);
+
+#if DEBUG_ATN == 1
+        std::cout << "EVAL rule " << pt->ruleIndex << ":" << pt->predIndex << std::endl;
+#endif
+
+      configs->hasSemanticContext = true;
+      if (evaluatePredicate(input, pt->ruleIndex, pt->predIndex, speculative)) {
+        c = std::make_shared<LexerATNConfig>(config, t->target);
+      }
+      break;
+    }
+
+    case Transition::ACTION:
+      if (config->context == nullptr|| config->context->hasEmptyPath()) {
+        // execute actions anywhere in the start rule for a token.
+        //
+        // TO_DO: if the entry rule is invoked recursively, some
+        // actions may be executed during the recursive call. The
+        // problem can appear when hasEmptyPath() is true but
+        // isEmpty() is false. In this case, the config needs to be
+        // split into two contexts - one with just the empty path
+        // and another with everything but the empty path.
+        // Unfortunately, the current algorithm does not allow
+        // getEpsilonTarget to return two configurations, so
+        // additional modifications are needed before we can support
+        // the split operation.
+        Ref<LexerActionExecutor> lexerActionExecutor = LexerActionExecutor::append(config->getLexerActionExecutor(),
+          atn.lexerActions[static_cast<ActionTransition *>(t)->actionIndex]);
+        c = std::make_shared<LexerATNConfig>(config, t->target, lexerActionExecutor);
+        break;
+      }
+      else {
+        // ignore actions in referenced rules
+        c = std::make_shared<LexerATNConfig>(config, t->target);
+        break;
+      }
+
+    case Transition::EPSILON:
+      c = std::make_shared<LexerATNConfig>(config, t->target);
+      break;
+
+    case Transition::ATOM:
+    case Transition::RANGE:
+    case Transition::SET:
+      if (treatEofAsEpsilon) {
+        if (t->matches(Token::EOF, Lexer::MIN_CHAR_VALUE, Lexer::MAX_CHAR_VALUE)) {
+          c = std::make_shared<LexerATNConfig>(config, t->target);
+          break;
+        }
+      }
+
+      break;
+
+    default: // To silence the compiler. Other transition types are not used here.
+      break;
+  }
+
+  return c;
+}
+
+bool LexerATNSimulator::evaluatePredicate(CharStream *input, size_t ruleIndex, size_t predIndex, bool speculative) {
+  // assume true if no recognizer was provided
+  if (_recog == nullptr) {
+    return true;
+  }
+
+  if (!speculative) {
+    return _recog->sempred(nullptr, ruleIndex, predIndex);
+  }
+
+  size_t savedCharPositionInLine = _charPositionInLine;
+  size_t savedLine = _line;
+  size_t index = input->index();
+  ssize_t marker = input->mark();
+
+  auto onExit = finally([this, input, savedCharPositionInLine, savedLine, index, marker] {
+    _charPositionInLine = savedCharPositionInLine;
+    _line = savedLine;
+    input->seek(index);
+    input->release(marker);
+  });
+
+  consume(input);
+  return _recog->sempred(nullptr, ruleIndex, predIndex);
+}
+
+void LexerATNSimulator::captureSimState(CharStream *input, dfa::DFAState *dfaState) {
+  _prevAccept.index = input->index();
+  _prevAccept.line = _line;
+  _prevAccept.charPos = _charPositionInLine;
+  _prevAccept.dfaState = dfaState;
+}
+
+dfa::DFAState *LexerATNSimulator::addDFAEdge(dfa::DFAState *from, size_t t, ATNConfigSet *q) {
+  /* leading to this call, ATNConfigSet.hasSemanticContext is used as a
+   * marker indicating dynamic predicate evaluation makes this edge
+   * dependent on the specific input sequence, so the static edge in the
+   * DFA should be omitted. The target DFAState is still created since
+   * execATN has the ability to resynchronize with the DFA state cache
+   * following the predicate evaluation step.
+   *
+   * TJP notes: next time through the DFA, we see a pred again and eval.
+   * If that gets us to a previously created (but dangling) DFA
+   * state, we can continue in pure DFA mode from there.
+   */
+  bool suppressEdge = q->hasSemanticContext;
+  q->hasSemanticContext = false;
+
+  dfa::DFAState *to = addDFAState(q);
+
+  if (suppressEdge) {
+    return to;
+  }
+
+  addDFAEdge(from, t, to);
+  return to;
+}
+
+void LexerATNSimulator::addDFAEdge(dfa::DFAState *p, size_t t, dfa::DFAState *q) {
+  if (/*t < MIN_DFA_EDGE ||*/ t > MAX_DFA_EDGE) { // MIN_DFA_EDGE is 0
+    // Only track edges within the DFA bounds
+    return;
+  }
+
+  _edgeLock.writeLock();
+  p->edges[t - MIN_DFA_EDGE] = q; // connect
+  _edgeLock.writeUnlock();
+}
+
+dfa::DFAState *LexerATNSimulator::addDFAState(ATNConfigSet *configs) {
+  /* the lexer evaluates predicates on-the-fly; by this point configs
+   * should not contain any configurations with unevaluated predicates.
+   */
+  assert(!configs->hasSemanticContext);
+
+  dfa::DFAState *proposed = new dfa::DFAState(std::unique_ptr<ATNConfigSet>(configs)); /* mem-check: managed by the DFA or deleted below */
+  Ref<ATNConfig> firstConfigWithRuleStopState = nullptr;
+  for (auto &c : configs->configs) {
+    if (is<RuleStopState *>(c->state)) {
+      firstConfigWithRuleStopState = c;
+      break;
+    }
+  }
+
+  if (firstConfigWithRuleStopState != nullptr) {
+    proposed->isAcceptState = true;
+    proposed->lexerActionExecutor = std::dynamic_pointer_cast<LexerATNConfig>(firstConfigWithRuleStopState)->getLexerActionExecutor();
+    proposed->prediction = atn.ruleToTokenType[firstConfigWithRuleStopState->state->ruleIndex];
+  }
+
+  dfa::DFA &dfa = _decisionToDFA[_mode];
+
+  _stateLock.writeLock();
+  if (!dfa.states.empty()) {
+    auto iterator = dfa.states.find(proposed);
+    if (iterator != dfa.states.end()) {
+      delete proposed;
+      _stateLock.writeUnlock();
+      return *iterator;
+    }
+  }
+
+  proposed->stateNumber = (int)dfa.states.size();
+  proposed->configs->setReadonly(true);
+
+  dfa.states.insert(proposed);
+  _stateLock.writeUnlock();
+
+  return proposed;
+}
+
+dfa::DFA& LexerATNSimulator::getDFA(size_t mode) {
+  return _decisionToDFA[mode];
+}
+
+std::string LexerATNSimulator::getText(CharStream *input) {
+  // index is first lookahead char, don't include.
+  return input->getText(misc::Interval(_startIndex, input->index() - 1));
+}
+
+size_t LexerATNSimulator::getLine() const {
+  return _line;
+}
+
+void LexerATNSimulator::setLine(size_t line) {
+  _line = line;
+}
+
+size_t LexerATNSimulator::getCharPositionInLine() {
+  return _charPositionInLine;
+}
+
+void LexerATNSimulator::setCharPositionInLine(size_t charPositionInLine) {
+  _charPositionInLine = charPositionInLine;
+}
+
+void LexerATNSimulator::consume(CharStream *input) {
+  size_t curChar = input->LA(1);
+  if (curChar == '\n') {
+    _line++;
+    _charPositionInLine = 0;
+  } else {
+    _charPositionInLine++;
+  }
+  input->consume();
+}
+
+std::string LexerATNSimulator::getTokenName(size_t t) {
+  if (t == Token::EOF) {
+    return "EOF";
+  }
+  return std::string("'") + static_cast<char>(t) + std::string("'");
+}
+
+void LexerATNSimulator::InitializeInstanceFields() {
+  _startIndex = 0;
+  _line = 1;
+  _charPositionInLine = 0;
+  _mode = antlr4::Lexer::DEFAULT_MODE;
+}

runtime-linux/antlr4-runtime/atn/LexerATNSimulator.h

@@ -0,0 +1,210 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATNSimulator.h"
+#include "atn/LexerATNConfig.h"
+#include "atn/ATNConfigSet.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// "dup" of ParserInterpreter
+  class ANTLR4CPP_PUBLIC LexerATNSimulator : public ATNSimulator {
+  protected:
+    class SimState {
+    public:
+      virtual ~SimState();
+
+    protected:
+      size_t index;
+      size_t line;
+      size_t charPos;
+      dfa::DFAState *dfaState;
+      virtual void reset();
+      friend class LexerATNSimulator;
+
+    private:
+      void InitializeInstanceFields();
+
+    public:
+      SimState() {
+        InitializeInstanceFields();
+      }
+    };
+
+
+  public:
+    static const size_t MIN_DFA_EDGE = 0;
+    static const size_t MAX_DFA_EDGE = 127; // forces unicode to stay in ATN
+
+  protected:
+    /// <summary>
+    /// When we hit an accept state in either the DFA or the ATN, we
+    ///  have to notify the character stream to start buffering characters
+    ///  via <seealso cref="IntStream#mark"/> and record the current state. The current sim state
+    ///  includes the current index into the input, the current line,
+    ///  and current character position in that line. Note that the Lexer is
+    ///  tracking the starting line and characterization of the token. These
+    ///  variables track the "state" of the simulator when it hits an accept state.
+    /// <p/>
+    ///  We track these variables separately for the DFA and ATN simulation
+    ///  because the DFA simulation often has to fail over to the ATN
+    ///  simulation. If the ATN simulation fails, we need the DFA to fall
+    ///  back to its previously accepted state, if any. If the ATN succeeds,
+    ///  then the ATN does the accept and the DFA simulator that invoked it
+    ///  can simply return the predicted token type.
+    /// </summary>
+    Lexer *const _recog;
+
+    /// The current token's starting index into the character stream.
+    ///  Shared across DFA to ATN simulation in case the ATN fails and the
+    ///  DFA did not have a previous accept state. In this case, we use the
+    ///  ATN-generated exception object.
+    size_t _startIndex;
+
+    /// line number 1..n within the input.
+    size_t _line;
+
+    /// The index of the character relative to the beginning of the line 0..n-1.
+    size_t _charPositionInLine;
+
+  public:
+    std::vector<dfa::DFA> &_decisionToDFA;
+
+  protected:
+    size_t _mode;
+
+    /// Used during DFA/ATN exec to record the most recent accept configuration info.
+    SimState _prevAccept;
+
+  public:
+    static int match_calls;
+
+    LexerATNSimulator(const ATN &atn, std::vector<dfa::DFA> &decisionToDFA, PredictionContextCache &sharedContextCache);
+    LexerATNSimulator(Lexer *recog, const ATN &atn, std::vector<dfa::DFA> &decisionToDFA, PredictionContextCache &sharedContextCache);
+    virtual ~LexerATNSimulator () {}
+
+    virtual void copyState(LexerATNSimulator *simulator);
+    virtual size_t match(CharStream *input, size_t mode);
+    virtual void reset() override;
+
+    virtual void clearDFA() override;
+
+  protected:
+    virtual size_t matchATN(CharStream *input);
+    virtual size_t execATN(CharStream *input, dfa::DFAState *ds0);
+
+    /// <summary>
+    /// Get an existing target state for an edge in the DFA. If the target state
+    /// for the edge has not yet been computed or is otherwise not available,
+    /// this method returns {@code null}.
+    /// </summary>
+    /// <param name="s"> The current DFA state </param>
+    /// <param name="t"> The next input symbol </param>
+    /// <returns> The existing target DFA state for the given input symbol
+    /// {@code t}, or {@code null} if the target state for this edge is not
+    /// already cached </returns>
+    virtual dfa::DFAState *getExistingTargetState(dfa::DFAState *s, size_t t);
+
+    /// <summary>
+    /// Compute a target state for an edge in the DFA, and attempt to add the
+    /// computed state and corresponding edge to the DFA.
+    /// </summary>
+    /// <param name="input"> The input stream </param>
+    /// <param name="s"> The current DFA state </param>
+    /// <param name="t"> The next input symbol
+    /// </param>
+    /// <returns> The computed target DFA state for the given input symbol
+    /// {@code t}. If {@code t} does not lead to a valid DFA state, this method
+    /// returns <seealso cref="#ERROR"/>. </returns>
+    virtual dfa::DFAState *computeTargetState(CharStream *input, dfa::DFAState *s, size_t t);
+
+    virtual size_t failOrAccept(CharStream *input, ATNConfigSet *reach, size_t t);
+
+    /// <summary>
+    /// Given a starting configuration set, figure out all ATN configurations
+    ///  we can reach upon input {@code t}. Parameter {@code reach} is a return
+    ///  parameter.
+    /// </summary>
+    void getReachableConfigSet(CharStream *input, ATNConfigSet *closure_, // closure_ as we have a closure() already
+                               ATNConfigSet *reach, size_t t);
+
+    virtual void accept(CharStream *input, const Ref<LexerActionExecutor> &lexerActionExecutor, size_t startIndex, size_t index,
+                        size_t line, size_t charPos);
+
+    virtual ATNState *getReachableTarget(Transition *trans, size_t t);
+
+    virtual std::unique_ptr<ATNConfigSet> computeStartState(CharStream *input, ATNState *p);
+
+    /// <summary>
+    /// Since the alternatives within any lexer decision are ordered by
+    /// preference, this method stops pursuing the closure as soon as an accept
+    /// state is reached. After the first accept state is reached by depth-first
+    /// search from {@code config}, all other (potentially reachable) states for
+    /// this rule would have a lower priority.
+    /// </summary>
+    /// <returns> {@code true} if an accept state is reached, otherwise
+    /// {@code false}. </returns>
+    virtual bool closure(CharStream *input, const Ref<LexerATNConfig> &config, ATNConfigSet *configs,
+                         bool currentAltReachedAcceptState, bool speculative, bool treatEofAsEpsilon);
+
+    // side-effect: can alter configs.hasSemanticContext
+    virtual Ref<LexerATNConfig> getEpsilonTarget(CharStream *input, const Ref<LexerATNConfig> &config, Transition *t,
+      ATNConfigSet *configs, bool speculative, bool treatEofAsEpsilon);
+
+    /// <summary>
+    /// Evaluate a predicate specified in the lexer.
+    /// <p/>
+    /// If {@code speculative} is {@code true}, this method was called before
+    /// <seealso cref="#consume"/> for the matched character. This method should call
+    /// <seealso cref="#consume"/> before evaluating the predicate to ensure position
+    /// sensitive values, including <seealso cref="Lexer#getText"/>, <seealso cref="Lexer#getLine"/>,
+    /// and <seealso cref="Lexer#getCharPositionInLine"/>, properly reflect the current
+    /// lexer state. This method should restore {@code input} and the simulator
+    /// to the original state before returning (i.e. undo the actions made by the
+    /// call to <seealso cref="#consume"/>.
+    /// </summary>
+    /// <param name="input"> The input stream. </param>
+    /// <param name="ruleIndex"> The rule containing the predicate. </param>
+    /// <param name="predIndex"> The index of the predicate within the rule. </param>
+    /// <param name="speculative"> {@code true} if the current index in {@code input} is
+    /// one character before the predicate's location.
+    /// </param>
+    /// <returns> {@code true} if the specified predicate evaluates to
+    /// {@code true}. </returns>
+    virtual bool evaluatePredicate(CharStream *input, size_t ruleIndex, size_t predIndex, bool speculative);
+
+    virtual void captureSimState(CharStream *input, dfa::DFAState *dfaState);
+    virtual dfa::DFAState* addDFAEdge(dfa::DFAState *from, size_t t, ATNConfigSet *q);
+    virtual void addDFAEdge(dfa::DFAState *p, size_t t, dfa::DFAState *q);
+
+    /// <summary>
+    /// Add a new DFA state if there isn't one with this set of
+    /// configurations already. This method also detects the first
+    /// configuration containing an ATN rule stop state. Later, when
+    /// traversing the DFA, we will know which rule to accept.
+    /// </summary>
+    virtual dfa::DFAState *addDFAState(ATNConfigSet *configs);
+
+  public:
+    dfa::DFA& getDFA(size_t mode);
+
+    /// Get the text matched so far for the current token.
+    virtual std::string getText(CharStream *input);
+    virtual size_t getLine() const;
+    virtual void setLine(size_t line);
+    virtual size_t getCharPositionInLine();
+    virtual void setCharPositionInLine(size_t charPositionInLine);
+    virtual void consume(CharStream *input);
+    virtual std::string getTokenName(size_t t);
+
+  private:
+    void InitializeInstanceFields();
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerAction.cpp

@@ -0,0 +1,9 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "LexerAction.h"
+
+antlr4::atn::LexerAction::~LexerAction() {
+}

runtime-linux/antlr4-runtime/atn/LexerAction.h

@@ -0,0 +1,66 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerActionType.h"
+#include "antlr4-common.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Represents a single action which can be executed following the successful
+  /// match of a lexer rule. Lexer actions are used for both embedded action syntax
+  /// and ANTLR 4's new lexer command syntax.
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerAction {
+  public:
+    virtual ~LexerAction();
+
+    /// <summary>
+    /// Gets the serialization type of the lexer action.
+    /// </summary>
+    /// <returns> The serialization type of the lexer action. </returns>
+    virtual LexerActionType getActionType() const = 0;
+
+    /// <summary>
+    /// Gets whether the lexer action is position-dependent. Position-dependent
+    /// actions may have different semantics depending on the <seealso cref="CharStream"/>
+    /// index at the time the action is executed.
+    ///
+    /// <para>Many lexer commands, including {@code type}, {@code skip}, and
+    /// {@code more}, do not check the input index during their execution.
+    /// Actions like this are position-independent, and may be stored more
+    /// efficiently as part of the <seealso cref="LexerATNConfig#lexerActionExecutor"/>.</para>
+    /// </summary>
+    /// <returns> {@code true} if the lexer action semantics can be affected by the
+    /// position of the input <seealso cref="CharStream"/> at the time it is executed;
+    /// otherwise, {@code false}. </returns>
+    virtual bool isPositionDependent() const = 0;
+
+    /// <summary>
+    /// Execute the lexer action in the context of the specified <seealso cref="Lexer"/>.
+    ///
+    /// <para>For position-dependent actions, the input stream must already be
+    /// positioned correctly prior to calling this method.</para>
+    /// </summary>
+    /// <param name="lexer"> The lexer instance. </param>
+    virtual void execute(Lexer *lexer) = 0;
+
+    virtual size_t hashCode() const = 0;
+    virtual bool operator == (const LexerAction &obj) const = 0;
+    virtual bool operator != (const LexerAction &obj) const {
+      return !(*this == obj);
+    }
+
+    virtual std::string toString() const = 0;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerActionExecutor.cpp

@@ -0,0 +1,107 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "atn/LexerIndexedCustomAction.h"
+#include "support/CPPUtils.h"
+#include "support/Arrays.h"
+
+#include "atn/LexerActionExecutor.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+using namespace antlrcpp;
+
+LexerActionExecutor::LexerActionExecutor(const std::vector<Ref<LexerAction>> &lexerActions)
+  : _lexerActions(lexerActions), _hashCode(generateHashCode()) {
+}
+
+LexerActionExecutor::~LexerActionExecutor() {
+}
+
+Ref<LexerActionExecutor> LexerActionExecutor::append(Ref<LexerActionExecutor> const& lexerActionExecutor,
+                                                     Ref<LexerAction> const& lexerAction) {
+  if (lexerActionExecutor == nullptr) {
+    return std::make_shared<LexerActionExecutor>(std::vector<Ref<LexerAction>> { lexerAction });
+  }
+
+  std::vector<Ref<LexerAction>> lexerActions = lexerActionExecutor->_lexerActions; // Make a copy.
+  lexerActions.push_back(lexerAction);
+  return std::make_shared<LexerActionExecutor>(lexerActions);
+}
+
+Ref<LexerActionExecutor> LexerActionExecutor::fixOffsetBeforeMatch(int offset) {
+  std::vector<Ref<LexerAction>> updatedLexerActions;
+  for (size_t i = 0; i < _lexerActions.size(); i++) {
+    if (_lexerActions[i]->isPositionDependent() && !is<LexerIndexedCustomAction>(_lexerActions[i])) {
+      if (updatedLexerActions.empty()) {
+        updatedLexerActions = _lexerActions; // Make a copy.
+      }
+
+      updatedLexerActions[i] = std::make_shared<LexerIndexedCustomAction>(offset, _lexerActions[i]);
+    }
+  }
+
+  if (updatedLexerActions.empty()) {
+    return shared_from_this();
+  }
+
+  return std::make_shared<LexerActionExecutor>(updatedLexerActions);
+}
+
+std::vector<Ref<LexerAction>> LexerActionExecutor::getLexerActions() const {
+  return _lexerActions;
+}
+
+void LexerActionExecutor::execute(Lexer *lexer, CharStream *input, size_t startIndex) {
+  bool requiresSeek = false;
+  size_t stopIndex = input->index();
+
+  auto onExit = finally([requiresSeek, input, stopIndex]() {
+    if (requiresSeek) {
+      input->seek(stopIndex);
+    }
+  });
+  for (auto lexerAction : _lexerActions) {
+    if (is<LexerIndexedCustomAction>(lexerAction)) {
+      int offset = (std::static_pointer_cast<LexerIndexedCustomAction>(lexerAction))->getOffset();
+      input->seek(startIndex + offset);
+      lexerAction = std::static_pointer_cast<LexerIndexedCustomAction>(lexerAction)->getAction();
+      requiresSeek = (startIndex + offset) != stopIndex;
+    } else if (lexerAction->isPositionDependent()) {
+      input->seek(stopIndex);
+      requiresSeek = false;
+    }
+
+    lexerAction->execute(lexer);
+  }
+}
+
+size_t LexerActionExecutor::hashCode() const {
+  return _hashCode;
+}
+
+bool LexerActionExecutor::operator == (const LexerActionExecutor &obj) const {
+  if (&obj == this) {
+    return true;
+  }
+
+  return _hashCode == obj._hashCode && Arrays::equals(_lexerActions, obj._lexerActions);
+}
+
+bool LexerActionExecutor::operator != (const LexerActionExecutor &obj) const {
+  return !operator==(obj);
+}
+
+size_t LexerActionExecutor::generateHashCode() const {
+  size_t hash = MurmurHash::initialize();
+  for (auto lexerAction : _lexerActions) {
+    hash = MurmurHash::update(hash, lexerAction);
+  }
+  hash = MurmurHash::finish(hash, _lexerActions.size());
+
+  return hash;
+}

runtime-linux/antlr4-runtime/atn/LexerActionExecutor.h

@@ -0,0 +1,115 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "CharStream.h"
+#include "atn/LexerAction.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Represents an executor for a sequence of lexer actions which traversed during
+  /// the matching operation of a lexer rule (token).
+  ///
+  /// <para>The executor tracks position information for position-dependent lexer actions
+  /// efficiently, ensuring that actions appearing only at the end of the rule do
+  /// not cause bloating of the <seealso cref="DFA"/> created for the lexer.</para>
+  class ANTLR4CPP_PUBLIC LexerActionExecutor : public std::enable_shared_from_this<LexerActionExecutor> {
+  public:
+    /// <summary>
+    /// Constructs an executor for a sequence of <seealso cref="LexerAction"/> actions. </summary>
+    /// <param name="lexerActions"> The lexer actions to execute. </param>
+    LexerActionExecutor(const std::vector<Ref<LexerAction>> &lexerActions);
+    virtual ~LexerActionExecutor();
+
+    /// <summary>
+    /// Creates a <seealso cref="LexerActionExecutor"/> which executes the actions for
+    /// the input {@code lexerActionExecutor} followed by a specified
+    /// {@code lexerAction}.
+    /// </summary>
+    /// <param name="lexerActionExecutor"> The executor for actions already traversed by
+    /// the lexer while matching a token within a particular
+    /// <seealso cref="LexerATNConfig"/>. If this is {@code null}, the method behaves as
+    /// though it were an empty executor. </param>
+    /// <param name="lexerAction"> The lexer action to execute after the actions
+    /// specified in {@code lexerActionExecutor}.
+    /// </param>
+    /// <returns> A <seealso cref="LexerActionExecutor"/> for executing the combine actions
+    /// of {@code lexerActionExecutor} and {@code lexerAction}. </returns>
+    static Ref<LexerActionExecutor> append(Ref<LexerActionExecutor> const& lexerActionExecutor,
+                                           Ref<LexerAction> const& lexerAction);
+
+    /// <summary>
+    /// Creates a <seealso cref="LexerActionExecutor"/> which encodes the current offset
+    /// for position-dependent lexer actions.
+    ///
+    /// <para>Normally, when the executor encounters lexer actions where
+    /// <seealso cref="LexerAction#isPositionDependent"/> returns {@code true}, it calls
+    /// <seealso cref="IntStream#seek"/> on the input <seealso cref="CharStream"/> to set the input
+    /// position to the <em>end</em> of the current token. This behavior provides
+    /// for efficient DFA representation of lexer actions which appear at the end
+    /// of a lexer rule, even when the lexer rule matches a variable number of
+    /// characters.</para>
+    ///
+    /// <para>Prior to traversing a match transition in the ATN, the current offset
+    /// from the token start index is assigned to all position-dependent lexer
+    /// actions which have not already been assigned a fixed offset. By storing
+    /// the offsets relative to the token start index, the DFA representation of
+    /// lexer actions which appear in the middle of tokens remains efficient due
+    /// to sharing among tokens of the same length, regardless of their absolute
+    /// position in the input stream.</para>
+    ///
+    /// <para>If the current executor already has offsets assigned to all
+    /// position-dependent lexer actions, the method returns {@code this}.</para>
+    /// </summary>
+    /// <param name="offset"> The current offset to assign to all position-dependent
+    /// lexer actions which do not already have offsets assigned.
+    /// </param>
+    /// <returns> A <seealso cref="LexerActionExecutor"/> which stores input stream offsets
+    /// for all position-dependent lexer actions. </returns>
+    virtual Ref<LexerActionExecutor> fixOffsetBeforeMatch(int offset);
+
+    /// <summary>
+    /// Gets the lexer actions to be executed by this executor. </summary>
+    /// <returns> The lexer actions to be executed by this executor. </returns>
+    virtual std::vector<Ref<LexerAction>> getLexerActions() const;
+
+    /// <summary>
+    /// Execute the actions encapsulated by this executor within the context of a
+    /// particular <seealso cref="Lexer"/>.
+    ///
+    /// <para>This method calls <seealso cref="IntStream#seek"/> to set the position of the
+    /// {@code input} <seealso cref="CharStream"/> prior to calling
+    /// <seealso cref="LexerAction#execute"/> on a position-dependent action. Before the
+    /// method returns, the input position will be restored to the same position
+    /// it was in when the method was invoked.</para>
+    /// </summary>
+    /// <param name="lexer"> The lexer instance. </param>
+    /// <param name="input"> The input stream which is the source for the current token.
+    /// When this method is called, the current <seealso cref="IntStream#index"/> for
+    /// {@code input} should be the start of the following token, i.e. 1
+    /// character past the end of the current token. </param>
+    /// <param name="startIndex"> The token start index. This value may be passed to
+    /// <seealso cref="IntStream#seek"/> to set the {@code input} position to the beginning
+    /// of the token. </param>
+    virtual void execute(Lexer *lexer, CharStream *input, size_t startIndex);
+
+    virtual size_t hashCode() const;
+    virtual bool operator == (const LexerActionExecutor &obj) const;
+    virtual bool operator != (const LexerActionExecutor &obj) const;
+
+  private:
+    const std::vector<Ref<LexerAction>> _lexerActions;
+
+    /// Caches the result of <seealso cref="#hashCode"/> since the hash code is an element
+    /// of the performance-critical <seealso cref="LexerATNConfig#hashCode"/> operation.
+    const size_t _hashCode;
+
+    size_t generateHashCode() const;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerActionType.h

@@ -0,0 +1,55 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "antlr4-common.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Represents the serialization type of a <seealso cref="LexerAction"/>.
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  enum class LexerActionType : size_t {
+    /// <summary>
+    /// The type of a <seealso cref="LexerChannelAction"/> action.
+    /// </summary>
+    CHANNEL,
+    /// <summary>
+    /// The type of a <seealso cref="LexerCustomAction"/> action.
+    /// </summary>
+    CUSTOM,
+    /// <summary>
+    /// The type of a <seealso cref="LexerModeAction"/> action.
+    /// </summary>
+    MODE,
+    /// <summary>
+    /// The type of a <seealso cref="LexerMoreAction"/> action.
+    /// </summary>
+    MORE,
+    /// <summary>
+    /// The type of a <seealso cref="LexerPopModeAction"/> action.
+    /// </summary>
+    POP_MODE,
+    /// <summary>
+    /// The type of a <seealso cref="LexerPushModeAction"/> action.
+    /// </summary>
+    PUSH_MODE,
+    /// <summary>
+    /// The type of a <seealso cref="LexerSkipAction"/> action.
+    /// </summary>
+    SKIP,
+    /// <summary>
+    /// The type of a <seealso cref="LexerTypeAction"/> action.
+    /// </summary>
+    TYPE,
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerChannelAction.cpp

@@ -0,0 +1,55 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+
+#include "atn/LexerChannelAction.h"
+
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+LexerChannelAction::LexerChannelAction(int channel) : _channel(channel) {
+}
+
+int LexerChannelAction::getChannel() const {
+  return _channel;
+}
+
+LexerActionType LexerChannelAction::getActionType() const {
+  return LexerActionType::CHANNEL;
+}
+
+bool LexerChannelAction::isPositionDependent() const {
+  return false;
+}
+
+void LexerChannelAction::execute(Lexer *lexer) {
+  lexer->setChannel(_channel);
+}
+
+size_t LexerChannelAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  hash = MurmurHash::update(hash, _channel);
+  return MurmurHash::finish(hash, 2);
+}
+
+bool LexerChannelAction::operator == (const LexerAction &obj) const {
+  if (&obj == this) {
+    return true;
+  }
+
+  const LexerChannelAction *action = dynamic_cast<const LexerChannelAction *>(&obj);
+  if (action == nullptr) {
+    return false;
+  }
+
+  return _channel == action->_channel;
+}
+
+std::string LexerChannelAction::toString() const {
+  return "channel(" + std::to_string(_channel) + ")";
+}

runtime-linux/antlr4-runtime/atn/LexerChannelAction.h

@@ -0,0 +1,63 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/LexerActionType.h"
+
+namespace antlr4 {
+namespace atn {
+
+  using antlr4::Lexer;
+
+  /// <summary>
+  /// Implements the {@code channel} lexer action by calling
+  /// <seealso cref="Lexer#setChannel"/> with the assigned channel.
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerChannelAction final : public LexerAction {
+  public:
+    /// <summary>
+    /// Constructs a new {@code channel} action with the specified channel value. </summary>
+    /// <param name="channel"> The channel value to pass to <seealso cref="Lexer#setChannel"/>. </param>
+    LexerChannelAction(int channel);
+
+    /// <summary>
+    /// Gets the channel to use for the <seealso cref="Token"/> created by the lexer.
+    /// </summary>
+    /// <returns> The channel to use for the <seealso cref="Token"/> created by the lexer. </returns>
+    int getChannel() const;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#CHANNEL"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code false}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>This action is implemented by calling <seealso cref="Lexer#setChannel"/> with the
+    /// value provided by <seealso cref="#getChannel"/>.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    const int _channel;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerCustomAction.cpp

@@ -0,0 +1,62 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "support/CPPUtils.h"
+#include "Lexer.h"
+
+#include "atn/LexerCustomAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+LexerCustomAction::LexerCustomAction(size_t ruleIndex, size_t actionIndex) : _ruleIndex(ruleIndex), _actionIndex(actionIndex) {
+}
+
+size_t LexerCustomAction::getRuleIndex() const {
+  return _ruleIndex;
+}
+
+size_t LexerCustomAction::getActionIndex() const {
+  return _actionIndex;
+}
+
+LexerActionType LexerCustomAction::getActionType() const {
+  return LexerActionType::CUSTOM;
+}
+
+bool LexerCustomAction::isPositionDependent() const {
+  return true;
+}
+
+void LexerCustomAction::execute(Lexer *lexer) {
+  lexer->action(nullptr, _ruleIndex, _actionIndex);
+}
+
+size_t LexerCustomAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  hash = MurmurHash::update(hash, _ruleIndex);
+  hash = MurmurHash::update(hash, _actionIndex);
+  return MurmurHash::finish(hash, 3);
+}
+
+bool LexerCustomAction::operator == (const LexerAction &obj) const {
+  if (&obj == this) {
+    return true;
+  }
+
+  const LexerCustomAction *action = dynamic_cast<const LexerCustomAction *>(&obj);
+  if (action == nullptr) {
+    return false;
+  }
+
+  return _ruleIndex == action->_ruleIndex && _actionIndex == action->_actionIndex;
+}
+
+std::string LexerCustomAction::toString() const {
+  return antlrcpp::toString(this);
+}

runtime-linux/antlr4-runtime/atn/LexerCustomAction.h

@@ -0,0 +1,87 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/LexerActionType.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Executes a custom lexer action by calling <seealso cref="Recognizer#action"/> with the
+  /// rule and action indexes assigned to the custom action. The implementation of
+  /// a custom action is added to the generated code for the lexer in an override
+  /// of <seealso cref="Recognizer#action"/> when the grammar is compiled.
+  ///
+  /// <para>This class may represent embedded actions created with the <code>{...}</code>
+  /// syntax in ANTLR 4, as well as actions created for lexer commands where the
+  /// command argument could not be evaluated when the grammar was compiled.</para>
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerCustomAction final : public LexerAction {
+  public:
+    /// <summary>
+    /// Constructs a custom lexer action with the specified rule and action
+    /// indexes.
+    /// </summary>
+    /// <param name="ruleIndex"> The rule index to use for calls to
+    /// <seealso cref="Recognizer#action"/>. </param>
+    /// <param name="actionIndex"> The action index to use for calls to
+    /// <seealso cref="Recognizer#action"/>. </param>
+    LexerCustomAction(size_t ruleIndex, size_t actionIndex);
+
+    /// <summary>
+    /// Gets the rule index to use for calls to <seealso cref="Recognizer#action"/>.
+    /// </summary>
+    /// <returns> The rule index for the custom action. </returns>
+    size_t getRuleIndex() const;
+
+    /// <summary>
+    /// Gets the action index to use for calls to <seealso cref="Recognizer#action"/>.
+    /// </summary>
+    /// <returns> The action index for the custom action. </returns>
+    size_t getActionIndex() const;
+
+    /// <summary>
+    /// {@inheritDoc}
+    /// </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#CUSTOM"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// Gets whether the lexer action is position-dependent. Position-dependent
+    /// actions may have different semantics depending on the <seealso cref="CharStream"/>
+    /// index at the time the action is executed.
+    ///
+    /// <para>Custom actions are position-dependent since they may represent a
+    /// user-defined embedded action which makes calls to methods like
+    /// <seealso cref="Lexer#getText"/>.</para>
+    /// </summary>
+    /// <returns> This method returns {@code true}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>Custom actions are implemented by calling <seealso cref="Lexer#action"/> with the
+    /// appropriate rule and action indexes.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    const size_t _ruleIndex;
+    const size_t _actionIndex;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerIndexedCustomAction.cpp

@@ -0,0 +1,63 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+#include "support/CPPUtils.h"
+
+#include "atn/LexerIndexedCustomAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+LexerIndexedCustomAction::LexerIndexedCustomAction(int offset, Ref<LexerAction> const& action)
+  : _offset(offset), _action(action) {
+}
+
+int LexerIndexedCustomAction::getOffset() const {
+  return _offset;
+}
+
+Ref<LexerAction> LexerIndexedCustomAction::getAction() const {
+  return _action;
+}
+
+LexerActionType LexerIndexedCustomAction::getActionType() const {
+  return _action->getActionType();
+}
+
+bool LexerIndexedCustomAction::isPositionDependent() const {
+  return true;
+}
+
+void LexerIndexedCustomAction::execute(Lexer *lexer) {
+  // assume the input stream position was properly set by the calling code
+  _action->execute(lexer);
+}
+
+size_t LexerIndexedCustomAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, _offset);
+  hash = MurmurHash::update(hash, _action);
+  return MurmurHash::finish(hash, 2);
+}
+
+bool LexerIndexedCustomAction::operator == (const LexerAction &obj) const {
+  if (&obj == this) {
+    return true;
+  }
+
+  const LexerIndexedCustomAction *action = dynamic_cast<const LexerIndexedCustomAction *>(&obj);
+  if (action == nullptr) {
+    return false;
+  }
+
+  return _offset == action->_offset && *_action == *action->_action;
+}
+
+std::string LexerIndexedCustomAction::toString() const {
+  return antlrcpp::toString(this);
+}

runtime-linux/antlr4-runtime/atn/LexerIndexedCustomAction.h

@@ -0,0 +1,82 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "RuleContext.h"
+#include "atn/LexerAction.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// This implementation of <seealso cref="LexerAction"/> is used for tracking input offsets
+  /// for position-dependent actions within a <seealso cref="LexerActionExecutor"/>.
+  ///
+  /// <para>This action is not serialized as part of the ATN, and is only required for
+  /// position-dependent lexer actions which appear at a location other than the
+  /// end of a rule. For more information about DFA optimizations employed for
+  /// lexer actions, see <seealso cref="LexerActionExecutor#append"/> and
+  /// <seealso cref="LexerActionExecutor#fixOffsetBeforeMatch"/>.</para>
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerIndexedCustomAction final : public LexerAction {
+  public:
+    /// <summary>
+    /// Constructs a new indexed custom action by associating a character offset
+    /// with a <seealso cref="LexerAction"/>.
+    ///
+    /// <para>Note: This class is only required for lexer actions for which
+    /// <seealso cref="LexerAction#isPositionDependent"/> returns {@code true}.</para>
+    /// </summary>
+    /// <param name="offset"> The offset into the input <seealso cref="CharStream"/>, relative to
+    /// the token start index, at which the specified lexer action should be
+    /// executed. </param>
+    /// <param name="action"> The lexer action to execute at a particular offset in the
+    /// input <seealso cref="CharStream"/>. </param>
+    LexerIndexedCustomAction(int offset, Ref<LexerAction> const& action);
+
+    /// <summary>
+    /// Gets the location in the input <seealso cref="CharStream"/> at which the lexer
+    /// action should be executed. The value is interpreted as an offset relative
+    /// to the token start index.
+    /// </summary>
+    /// <returns> The location in the input <seealso cref="CharStream"/> at which the lexer
+    /// action should be executed. </returns>
+    int getOffset() const;
+
+    /// <summary>
+    /// Gets the lexer action to execute.
+    /// </summary>
+    /// <returns> A <seealso cref="LexerAction"/> object which executes the lexer action. </returns>
+    Ref<LexerAction> getAction() const;
+
+    /// <summary>
+    /// {@inheritDoc}
+    /// </summary>
+    /// <returns> This method returns the result of calling <seealso cref="#getActionType"/>
+    /// on the <seealso cref="LexerAction"/> returned by <seealso cref="#getAction"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code true}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    virtual void execute(Lexer *lexer) override;
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    const int _offset;
+    const Ref<LexerAction> _action;
+  };
+
+} // namespace atn
+} // namespace antlr4
+

runtime-linux/antlr4-runtime/atn/LexerModeAction.cpp

@@ -0,0 +1,56 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+
+#include "atn/LexerModeAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+LexerModeAction::LexerModeAction(int mode) : _mode(mode) {
+}
+
+int LexerModeAction::getMode() {
+  return _mode;
+}
+
+LexerActionType LexerModeAction::getActionType() const {
+  return LexerActionType::MODE;
+}
+
+bool LexerModeAction::isPositionDependent() const {
+  return false;
+}
+
+void LexerModeAction::execute(Lexer *lexer) {
+  lexer->setMode(_mode);
+}
+
+size_t LexerModeAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  hash = MurmurHash::update(hash, _mode);
+  return MurmurHash::finish(hash, 2);
+}
+
+bool LexerModeAction::operator == (const LexerAction &obj) const {
+  if (&obj == this) {
+    return true;
+  }
+
+  const LexerModeAction *action = dynamic_cast<const LexerModeAction *>(&obj);
+  if (action == nullptr) {
+    return false;
+  }
+
+  return _mode == action->_mode;
+}
+
+std::string LexerModeAction::toString() const {
+  return "mode(" + std::to_string(_mode) + ")";
+}

runtime-linux/antlr4-runtime/atn/LexerModeAction.h

@@ -0,0 +1,61 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/LexerActionType.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Implements the {@code mode} lexer action by calling <seealso cref="Lexer#mode"/> with
+  /// the assigned mode.
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerModeAction final : public LexerAction {
+  public:
+    /// <summary>
+    /// Constructs a new {@code mode} action with the specified mode value. </summary>
+    /// <param name="mode"> The mode value to pass to <seealso cref="Lexer#mode"/>. </param>
+    LexerModeAction(int mode);
+
+    /// <summary>
+    /// Get the lexer mode this action should transition the lexer to.
+    /// </summary>
+    /// <returns> The lexer mode for this {@code mode} command. </returns>
+    int getMode();
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#MODE"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code false}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>This action is implemented by calling <seealso cref="Lexer#mode"/> with the
+    /// value provided by <seealso cref="#getMode"/>.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    const int _mode;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerMoreAction.cpp

@@ -0,0 +1,47 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+
+#include "atn/LexerMoreAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+const Ref<LexerMoreAction> LexerMoreAction::getInstance() {
+  static Ref<LexerMoreAction> instance(new LexerMoreAction());
+  return instance;
+}
+
+LexerMoreAction::LexerMoreAction() {
+}
+
+LexerActionType LexerMoreAction::getActionType() const {
+  return LexerActionType::MORE;
+}
+
+bool LexerMoreAction::isPositionDependent() const {
+  return false;
+}
+
+void LexerMoreAction::execute(Lexer *lexer) {
+  lexer->more();
+}
+
+size_t LexerMoreAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  return MurmurHash::finish(hash, 1);
+}
+
+bool LexerMoreAction::operator == (const LexerAction &obj) const {
+  return &obj == this;
+}
+
+std::string LexerMoreAction::toString() const {
+  return "more";
+}

runtime-linux/antlr4-runtime/atn/LexerMoreAction.h

@@ -0,0 +1,57 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/LexerActionType.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Implements the {@code more} lexer action by calling <seealso cref="Lexer#more"/>.
+  ///
+  /// <para>The {@code more} command does not have any parameters, so this action is
+  /// implemented as a singleton instance exposed by <seealso cref="#INSTANCE"/>.</para>
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerMoreAction final : public LexerAction {
+  public:
+    /// <summary>
+    /// Provides a singleton instance of this parameterless lexer action.
+    /// </summary>
+    static const Ref<LexerMoreAction> getInstance();
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#MORE"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code false}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>This action is implemented by calling <seealso cref="Lexer#more"/>.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    /// Constructs the singleton instance of the lexer {@code more} command.
+    LexerMoreAction();
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerPopModeAction.cpp

@@ -0,0 +1,47 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+
+#include "atn/LexerPopModeAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+const Ref<LexerPopModeAction> LexerPopModeAction::getInstance() {
+  static Ref<LexerPopModeAction> instance(new LexerPopModeAction());
+  return instance;
+}
+
+LexerPopModeAction::LexerPopModeAction() {
+}
+
+LexerActionType LexerPopModeAction::getActionType() const {
+  return LexerActionType::POP_MODE;
+}
+
+bool LexerPopModeAction::isPositionDependent() const {
+  return false;
+}
+
+void LexerPopModeAction::execute(Lexer *lexer) {
+  lexer->popMode();
+}
+
+size_t LexerPopModeAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  return MurmurHash::finish(hash, 1);
+}
+
+bool LexerPopModeAction::operator == (const LexerAction &obj) const {
+  return &obj == this;
+}
+
+std::string LexerPopModeAction::toString() const {
+  return "popMode";
+}

runtime-linux/antlr4-runtime/atn/LexerPopModeAction.h

@@ -0,0 +1,57 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/LexerActionType.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Implements the {@code popMode} lexer action by calling <seealso cref="Lexer#popMode"/>.
+  ///
+  /// <para>The {@code popMode} command does not have any parameters, so this action is
+  /// implemented as a singleton instance exposed by <seealso cref="#INSTANCE"/>.</para>
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerPopModeAction final : public LexerAction {
+  public:
+    /// <summary>
+    /// Provides a singleton instance of this parameterless lexer action.
+    /// </summary>
+    static const Ref<LexerPopModeAction> getInstance();
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#POP_MODE"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code false}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>This action is implemented by calling <seealso cref="Lexer#popMode"/>.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    /// Constructs the singleton instance of the lexer {@code popMode} command.
+    LexerPopModeAction();
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerPushModeAction.cpp

@@ -0,0 +1,56 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+
+#include "atn/LexerPushModeAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+LexerPushModeAction::LexerPushModeAction(int mode) : _mode(mode) {
+}
+
+int LexerPushModeAction::getMode() const {
+  return _mode;
+}
+
+LexerActionType LexerPushModeAction::getActionType() const {
+  return LexerActionType::PUSH_MODE;
+}
+
+bool LexerPushModeAction::isPositionDependent() const {
+  return false;
+}
+
+void LexerPushModeAction::execute(Lexer *lexer) {
+  lexer->pushMode(_mode);
+}
+
+size_t LexerPushModeAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  hash = MurmurHash::update(hash, _mode);
+  return MurmurHash::finish(hash, 2);
+}
+
+bool LexerPushModeAction::operator == (const LexerAction &obj) const {
+  if (&obj == this) {
+    return true;
+  }
+
+  const LexerPushModeAction *action = dynamic_cast<const LexerPushModeAction *>(&obj);
+  if (action == nullptr) {
+    return false;
+  }
+
+  return _mode == action->_mode;
+}
+
+std::string LexerPushModeAction::toString() const {
+  return "pushMode(" + std::to_string(_mode) + ")";
+}

runtime-linux/antlr4-runtime/atn/LexerPushModeAction.h

@@ -0,0 +1,61 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/LexerActionType.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Implements the {@code pushMode} lexer action by calling
+  /// <seealso cref="Lexer#pushMode"/> with the assigned mode.
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerPushModeAction final : public LexerAction {
+  public:
+    /// <summary>
+    /// Constructs a new {@code pushMode} action with the specified mode value. </summary>
+    /// <param name="mode"> The mode value to pass to <seealso cref="Lexer#pushMode"/>. </param>
+    LexerPushModeAction(int mode);
+
+    /// <summary>
+    /// Get the lexer mode this action should transition the lexer to.
+    /// </summary>
+    /// <returns> The lexer mode for this {@code pushMode} command. </returns>
+    int getMode() const;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#PUSH_MODE"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code false}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>This action is implemented by calling <seealso cref="Lexer#pushMode"/> with the
+    /// value provided by <seealso cref="#getMode"/>.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    const int _mode;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerSkipAction.cpp

@@ -0,0 +1,47 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+
+#include "atn/LexerSkipAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+const Ref<LexerSkipAction> LexerSkipAction::getInstance() {
+  static Ref<LexerSkipAction> instance(new LexerSkipAction());
+  return instance;
+}
+
+LexerSkipAction::LexerSkipAction() {
+}
+
+LexerActionType LexerSkipAction::getActionType() const {
+  return LexerActionType::SKIP;
+}
+
+bool LexerSkipAction::isPositionDependent() const {
+  return false;
+}
+
+void LexerSkipAction::execute(Lexer *lexer) {
+  lexer->skip();
+}
+
+size_t LexerSkipAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  return MurmurHash::finish(hash, 1);
+}
+
+bool LexerSkipAction::operator == (const LexerAction &obj) const {
+  return &obj == this;
+}
+
+std::string LexerSkipAction::toString() const {
+  return "skip";
+}

runtime-linux/antlr4-runtime/atn/LexerSkipAction.h

@@ -0,0 +1,55 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerAction.h"
+#include "atn/LexerActionType.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// Implements the {@code skip} lexer action by calling <seealso cref="Lexer#skip"/>.
+  ///
+  /// <para>The {@code skip} command does not have any parameters, so this action is
+  /// implemented as a singleton instance exposed by <seealso cref="#INSTANCE"/>.</para>
+  ///
+  /// @author Sam Harwell
+  /// @since 4.2
+  /// </summary>
+  class ANTLR4CPP_PUBLIC LexerSkipAction final : public LexerAction {
+  public:
+    /// Provides a singleton instance of this parameterless lexer action.
+    static const Ref<LexerSkipAction> getInstance();
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#SKIP"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code false}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>This action is implemented by calling <seealso cref="Lexer#skip"/>.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    /// Constructs the singleton instance of the lexer {@code skip} command.
+    LexerSkipAction();
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LexerTypeAction.cpp

@@ -0,0 +1,56 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "misc/MurmurHash.h"
+#include "Lexer.h"
+
+#include "atn/LexerTypeAction.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+using namespace antlr4::misc;
+
+LexerTypeAction::LexerTypeAction(int type) : _type(type) {
+}
+
+int LexerTypeAction::getType() const {
+  return _type;
+}
+
+LexerActionType LexerTypeAction::getActionType() const {
+  return LexerActionType::TYPE;
+}
+
+bool LexerTypeAction::isPositionDependent() const {
+  return false;
+}
+
+void LexerTypeAction::execute(Lexer *lexer) {
+  lexer->setType(_type);
+}
+
+size_t LexerTypeAction::hashCode() const {
+  size_t hash = MurmurHash::initialize();
+  hash = MurmurHash::update(hash, static_cast<size_t>(getActionType()));
+  hash = MurmurHash::update(hash, _type);
+  return MurmurHash::finish(hash, 2);
+}
+
+bool LexerTypeAction::operator == (const LexerAction &obj) const {
+  if (&obj == this) {
+    return true;
+  }
+
+  const LexerTypeAction *action = dynamic_cast<const LexerTypeAction *>(&obj);
+  if (action == nullptr) {
+    return false;
+  }
+
+  return _type == action->_type;
+}
+
+std::string LexerTypeAction::toString() const {
+  return "type(" + std::to_string(_type) + ")";
+}

runtime-linux/antlr4-runtime/atn/LexerTypeAction.h

@@ -0,0 +1,55 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/LexerActionType.h"
+#include "atn/LexerAction.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Implements the {@code type} lexer action by calling <seealso cref="Lexer#setType"/>
+  /// with the assigned type.
+  class ANTLR4CPP_PUBLIC LexerTypeAction : public LexerAction {
+  public:
+    /// <summary>
+    /// Constructs a new {@code type} action with the specified token type value. </summary>
+    /// <param name="type"> The type to assign to the token using <seealso cref="Lexer#setType"/>. </param>
+    LexerTypeAction(int type);
+
+    /// <summary>
+    /// Gets the type to assign to a token created by the lexer. </summary>
+    /// <returns> The type to assign to a token created by the lexer. </returns>
+    virtual int getType() const;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns <seealso cref="LexerActionType#TYPE"/>. </returns>
+    virtual LexerActionType getActionType() const override;
+
+    /// <summary>
+    /// {@inheritDoc} </summary>
+    /// <returns> This method returns {@code false}. </returns>
+    virtual bool isPositionDependent() const override;
+
+    /// <summary>
+    /// {@inheritDoc}
+    ///
+    /// <para>This action is implemented by calling <seealso cref="Lexer#setType"/> with the
+    /// value provided by <seealso cref="#getType"/>.</para>
+    /// </summary>
+    virtual void execute(Lexer *lexer) override;
+
+    virtual size_t hashCode() const override;
+    virtual bool operator == (const LexerAction &obj) const override;
+    virtual std::string toString() const override;
+
+  private:
+    const int _type;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LookaheadEventInfo.cpp

@@ -0,0 +1,16 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/LookaheadEventInfo.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+
+LookaheadEventInfo::LookaheadEventInfo(size_t decision, ATNConfigSet *configs, size_t predictedAlt,
+  TokenStream *input, size_t startIndex, size_t stopIndex, bool fullCtx)
+  : DecisionEventInfo(decision, configs, input, startIndex, stopIndex, fullCtx) {
+
+  this->predictedAlt = predictedAlt;
+}

runtime-linux/antlr4-runtime/atn/LookaheadEventInfo.h

@@ -0,0 +1,42 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionEventInfo.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// This class represents profiling event information for tracking the lookahead
+  /// depth required in order to make a prediction.
+  class ANTLR4CPP_PUBLIC LookaheadEventInfo : public DecisionEventInfo {
+  public:
+    /// The alternative chosen by adaptivePredict(), not necessarily
+    ///  the outermost alt shown for a rule; left-recursive rules have
+    ///  user-level alts that differ from the rewritten rule with a (...) block
+    ///  and a (..)* loop.
+    size_t predictedAlt = 0;
+
+    /// <summary>
+    /// Constructs a new instance of the <seealso cref="LookaheadEventInfo"/> class with
+    /// the specified detailed lookahead information.
+    /// </summary>
+    /// <param name="decision"> The decision number </param>
+    /// <param name="configs"> The final configuration set containing the necessary
+    /// information to determine the result of a prediction, or {@code null} if
+    /// the final configuration set is not available </param>
+    /// <param name="input"> The input token stream </param>
+    /// <param name="startIndex"> The start index for the current prediction </param>
+    /// <param name="stopIndex"> The index at which the prediction was finally made </param>
+    /// <param name="fullCtx"> {@code true} if the current lookahead is part of an LL
+    /// prediction; otherwise, {@code false} if the current lookahead is part of
+    /// an SLL prediction </param>
+    LookaheadEventInfo(size_t decision, ATNConfigSet *configs, size_t predictedAlt, TokenStream *input, size_t startIndex,
+                       size_t stopIndex, bool fullCtx);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/LoopEndState.cpp

@@ -0,0 +1,12 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/LoopEndState.h"
+
+using namespace antlr4::atn;
+
+size_t LoopEndState::getStateType() {
+  return LOOP_END;
+}

runtime-linux/antlr4-runtime/atn/LoopEndState.h

@@ -0,0 +1,22 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATNState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Mark the end of a * or + loop.
+  class ANTLR4CPP_PUBLIC LoopEndState final : public ATNState {
+  public:
+    ATNState *loopBackState = nullptr;
+
+    virtual size_t getStateType() override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/Makefile

@@ -0,0 +1,67 @@
+
+CXXFLAGS += -g -std=c++0x -Wall #-Wextra
+CXXFLAGS += -I. -I../ -I../misc/ -I../tree/ -I../dfa/ \
+	-I../../../../../antlrcpp/
+
+#TODO LDFLAGS += ?
+
+ALL_CXXFLAGS = $(CPPFLAGS) $(CXXFLAGS)
+ALL_LDFLAGS = $(LDFLAGS)
+
+# Escote's files
+SRCS = \
+    AbstractPredicateTransition.cpp \
+	ActionTransition.cpp \
+	ArrayPredictionContext.cpp \
+	ATNDeserializationOptions.cpp \
+	ATNDeserializer.cpp \
+	ATNState.cpp \
+	ATNType.cpp \
+	AtomTransition.cpp \
+	BasicBlockStartState.cpp \
+	BasicState.cpp \
+	BlockEndState.cpp \
+	BlockStartState.cpp \
+	DecisionState.cpp \
+	EmptyPredictionContext.cpp \
+	EpsilonTransition.cpp \
+	LexerATNConfig.cpp \
+	LoopEndState.cpp
+# Escote's TODO: LL1Analyzer.cpp LexerATNSimulator.cpp ATNSimulator.cpp \
+	ATNSerializer.cpp ATNConfigSet.cpp ATNConfig.cpp \
+	ATN.cpp
+
+# Alejandro's files
+SRCS += \
+	NotSetTransition.cpp \
+	OrderedATNConfigSet.cpp \
+	PlusBlockStartState.cpp \
+	PlusLoopbackState.cpp \
+	PredicateTransition.cpp \
+	PredictionMode.cpp \
+	RangeTransition.cpp \
+	RuleStartState.cpp \
+	RuleStopState.cpp \
+	RuleTransition.cpp \
+	SemanticContext.cpp \
+	SetTransition.cpp \
+	SingletonPredictionContext.cpp \
+	StarBlockStartState.cpp \
+	StarLoopbackState.cpp \
+	StarLoopEntryState.cpp \
+	TokensStartState.cpp \
+	Transition.cpp \
+	WildcardTransition.cpp
+# Alejandro's TODO: PredictionContext.cpp PredictionContextCache.cpp \
+	PrecedencePredicateTransition.cpp ParserATNSimulator.cpp
+
+OBJS = $(SRCS:.cpp=.o)
+
+all: $(OBJS)
+
+%.o: %.cpp
+	$(CXX) -c $(ALL_CXXFLAGS) $< -o $@ 
+
+clean:
+	$(RM) $(OBJS)
+

runtime-linux/antlr4-runtime/atn/NotSetTransition.cpp

@@ -0,0 +1,27 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/NotSetTransition.h"
+#include "atn/ATNState.h"
+#include "misc/IntervalSet.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+
+NotSetTransition::NotSetTransition(ATNState *target, const misc::IntervalSet &set) : SetTransition(target, set) {
+}
+
+Transition::SerializationType NotSetTransition::getSerializationType() const {
+  return NOT_SET;
+}
+
+bool NotSetTransition::matches(size_t symbol, size_t minVocabSymbol, size_t maxVocabSymbol) const {
+  return symbol >= minVocabSymbol && symbol <= maxVocabSymbol
+    && !SetTransition::matches(symbol, minVocabSymbol, maxVocabSymbol);
+}
+
+std::string NotSetTransition::toString() const {
+  return "NOT_SET " + Transition::toString() + " { " + SetTransition::toString() + " }";
+}

runtime-linux/antlr4-runtime/atn/NotSetTransition.h

@@ -0,0 +1,25 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/SetTransition.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC NotSetTransition final : public SetTransition {
+  public:
+    NotSetTransition(ATNState *target, const misc::IntervalSet &set);
+
+    virtual SerializationType getSerializationType() const override;
+
+    virtual bool matches(size_t symbol, size_t minVocabSymbol, size_t maxVocabSymbol) const override;
+
+    virtual std::string toString() const override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/OrderedATNConfigSet.cpp

@@ -0,0 +1,12 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/OrderedATNConfigSet.h"
+
+using namespace antlr4::atn;
+
+size_t OrderedATNConfigSet::getHash(ATNConfig *c) {
+  return c->hashCode();
+}

runtime-linux/antlr4-runtime/atn/OrderedATNConfigSet.h

@@ -0,0 +1,20 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/ATNConfigSet.h"
+#include "atn/ATNConfig.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC OrderedATNConfigSet : public ATNConfigSet {
+  protected:
+    virtual size_t getHash(ATNConfig *c) override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ParseInfo.cpp

@@ -0,0 +1,102 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/ProfilingATNSimulator.h"
+#include "dfa/DFA.h"
+
+#include "atn/ParseInfo.h"
+
+using namespace antlr4::atn;
+
+ParseInfo::ParseInfo(ProfilingATNSimulator *atnSimulator) : _atnSimulator(atnSimulator) {
+}
+
+ParseInfo::~ParseInfo() {
+}
+
+std::vector<DecisionInfo> ParseInfo::getDecisionInfo() {
+  return _atnSimulator->getDecisionInfo();
+}
+
+std::vector<size_t> ParseInfo::getLLDecisions() {
+  std::vector<DecisionInfo> decisions = _atnSimulator->getDecisionInfo();
+  std::vector<size_t> LL;
+  for (size_t i = 0; i < decisions.size(); ++i) {
+    long long fallBack = decisions[i].LL_Fallback;
+    if (fallBack > 0) {
+      LL.push_back(i);
+    }
+  }
+  return LL;
+}
+
+long long ParseInfo::getTotalTimeInPrediction() {
+  std::vector<DecisionInfo> decisions = _atnSimulator->getDecisionInfo();
+  long long t = 0;
+  for (size_t i = 0; i < decisions.size(); ++i) {
+    t += decisions[i].timeInPrediction;
+  }
+  return t;
+}
+
+long long ParseInfo::getTotalSLLLookaheadOps() {
+  std::vector<DecisionInfo> decisions = _atnSimulator->getDecisionInfo();
+  long long k = 0;
+  for (size_t i = 0; i < decisions.size(); ++i) {
+    k += decisions[i].SLL_TotalLook;
+  }
+  return k;
+}
+
+long long ParseInfo::getTotalLLLookaheadOps() {
+  std::vector<DecisionInfo> decisions = _atnSimulator->getDecisionInfo();
+  long long k = 0;
+  for (size_t i = 0; i < decisions.size(); i++) {
+    k += decisions[i].LL_TotalLook;
+  }
+  return k;
+}
+
+long long ParseInfo::getTotalSLLATNLookaheadOps() {
+  std::vector<DecisionInfo> decisions = _atnSimulator->getDecisionInfo();
+  long long k = 0;
+  for (size_t i = 0; i < decisions.size(); ++i) {
+    k += decisions[i].SLL_ATNTransitions;
+  }
+  return k;
+}
+
+long long ParseInfo::getTotalLLATNLookaheadOps() {
+  std::vector<DecisionInfo> decisions = _atnSimulator->getDecisionInfo();
+  long long k = 0;
+  for (size_t i = 0; i < decisions.size(); ++i) {
+    k += decisions[i].LL_ATNTransitions;
+  }
+  return k;
+}
+
+long long ParseInfo::getTotalATNLookaheadOps() {
+  std::vector<DecisionInfo> decisions = _atnSimulator->getDecisionInfo();
+  long long k = 0;
+  for (size_t i = 0; i < decisions.size(); ++i) {
+    k += decisions[i].SLL_ATNTransitions;
+    k += decisions[i].LL_ATNTransitions;
+  }
+  return k;
+}
+
+size_t ParseInfo::getDFASize() {
+  size_t n = 0;
+  std::vector<dfa::DFA> &decisionToDFA = _atnSimulator->decisionToDFA;
+  for (size_t i = 0; i < decisionToDFA.size(); ++i) {
+    n += getDFASize(i);
+  }
+  return n;
+}
+
+size_t ParseInfo::getDFASize(size_t decision) {
+  dfa::DFA &decisionToDFA = _atnSimulator->decisionToDFA[decision];
+  return decisionToDFA.states.size();
+}

runtime-linux/antlr4-runtime/atn/ParseInfo.h

@@ -0,0 +1,102 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionInfo.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ProfilingATNSimulator;
+
+  /// This class provides access to specific and aggregate statistics gathered
+  /// during profiling of a parser.
+  class ANTLR4CPP_PUBLIC ParseInfo {
+  public:
+    ParseInfo(ProfilingATNSimulator *atnSimulator);
+    ParseInfo(ParseInfo const&) = default;
+    virtual ~ParseInfo();
+
+    ParseInfo& operator=(ParseInfo const&) = default;
+
+    /// <summary>
+    /// Gets an array of <seealso cref="DecisionInfo"/> instances containing the profiling
+    /// information gathered for each decision in the ATN.
+    /// </summary>
+    /// <returns> An array of <seealso cref="DecisionInfo"/> instances, indexed by decision
+    /// number. </returns>
+    virtual std::vector<DecisionInfo> getDecisionInfo();
+
+    /// <summary>
+    /// Gets the decision numbers for decisions that required one or more
+    /// full-context predictions during parsing. These are decisions for which
+    /// <seealso cref="DecisionInfo#LL_Fallback"/> is non-zero.
+    /// </summary>
+    /// <returns> A list of decision numbers which required one or more
+    /// full-context predictions during parsing. </returns>
+    virtual std::vector<size_t> getLLDecisions();
+
+    /// <summary>
+    /// Gets the total time spent during prediction across all decisions made
+    /// during parsing. This value is the sum of
+    /// <seealso cref="DecisionInfo#timeInPrediction"/> for all decisions.
+    /// </summary>
+    virtual long long getTotalTimeInPrediction();
+
+    /// <summary>
+    /// Gets the total number of SLL lookahead operations across all decisions
+    /// made during parsing. This value is the sum of
+    /// <seealso cref="DecisionInfo#SLL_TotalLook"/> for all decisions.
+    /// </summary>
+    virtual long long getTotalSLLLookaheadOps();
+
+    /// <summary>
+    /// Gets the total number of LL lookahead operations across all decisions
+    /// made during parsing. This value is the sum of
+    /// <seealso cref="DecisionInfo#LL_TotalLook"/> for all decisions.
+    /// </summary>
+    virtual long long getTotalLLLookaheadOps();
+
+    /// <summary>
+    /// Gets the total number of ATN lookahead operations for SLL prediction
+    /// across all decisions made during parsing.
+    /// </summary>
+    virtual long long getTotalSLLATNLookaheadOps();
+
+    /// <summary>
+    /// Gets the total number of ATN lookahead operations for LL prediction
+    /// across all decisions made during parsing.
+    /// </summary>
+    virtual long long getTotalLLATNLookaheadOps();
+
+    /// <summary>
+    /// Gets the total number of ATN lookahead operations for SLL and LL
+    /// prediction across all decisions made during parsing.
+    ///
+    /// <para>
+    /// This value is the sum of <seealso cref="#getTotalSLLATNLookaheadOps"/> and
+    /// <seealso cref="#getTotalLLATNLookaheadOps"/>.</para>
+    /// </summary>
+    virtual long long getTotalATNLookaheadOps();
+
+    /// <summary>
+    /// Gets the total number of DFA states stored in the DFA cache for all
+    /// decisions in the ATN.
+    /// </summary>
+    virtual size_t getDFASize();
+
+    /// <summary>
+    /// Gets the total number of DFA states stored in the DFA cache for a
+    /// particular decision.
+    /// </summary>
+    virtual size_t getDFASize(size_t decision);
+
+  protected:
+    const ProfilingATNSimulator *_atnSimulator; // non-owning, we are created by this simulator.
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/ParserATNSimulator.h

@@ -0,0 +1,904 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "PredictionMode.h"
+#include "dfa/DFAState.h"
+#include "atn/ATNSimulator.h"
+#include "atn/PredictionContext.h"
+#include "SemanticContext.h"
+#include "atn/ATNConfig.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /**
+   * The embodiment of the adaptive LL(*), ALL(*), parsing strategy.
+   *
+   * <p>
+   * The basic complexity of the adaptive strategy makes it harder to understand.
+   * We begin with ATN simulation to build paths in a DFA. Subsequent prediction
+   * requests go through the DFA first. If they reach a state without an edge for
+   * the current symbol, the algorithm fails over to the ATN simulation to
+   * complete the DFA path for the current input (until it finds a conflict state
+   * or uniquely predicting state).</p>
+   *
+   * <p>
+   * All of that is done without using the outer context because we want to create
+   * a DFA that is not dependent upon the rule invocation stack when we do a
+   * prediction. One DFA works in all contexts. We avoid using context not
+   * necessarily because it's slower, although it can be, but because of the DFA
+   * caching problem. The closure routine only considers the rule invocation stack
+   * created during prediction beginning in the decision rule. For example, if
+   * prediction occurs without invoking another rule's ATN, there are no context
+   * stacks in the configurations. When lack of context leads to a conflict, we
+   * don't know if it's an ambiguity or a weakness in the strong LL(*) parsing
+   * strategy (versus full LL(*)).</p>
+   *
+   * <p>
+   * When SLL yields a configuration set with conflict, we rewind the input and
+   * retry the ATN simulation, this time using full outer context without adding
+   * to the DFA. Configuration context stacks will be the full invocation stacks
+   * from the start rule. If we get a conflict using full context, then we can
+   * definitively say we have a true ambiguity for that input sequence. If we
+   * don't get a conflict, it implies that the decision is sensitive to the outer
+   * context. (It is not context-sensitive in the sense of context-sensitive
+   * grammars.)</p>
+   *
+   * <p>
+   * The next time we reach this DFA state with an SLL conflict, through DFA
+   * simulation, we will again retry the ATN simulation using full context mode.
+   * This is slow because we can't save the results and have to "interpret" the
+   * ATN each time we get that input.</p>
+   *
+   * <p>
+   * <strong>CACHING FULL CONTEXT PREDICTIONS</strong></p>
+   *
+   * <p>
+   * We could cache results from full context to predicted alternative easily and
+   * that saves a lot of time but doesn't work in presence of predicates. The set
+   * of visible predicates from the ATN start state changes depending on the
+   * context, because closure can fall off the end of a rule. I tried to cache
+   * tuples (stack context, semantic context, predicted alt) but it was slower
+   * than interpreting and much more complicated. Also required a huge amount of
+   * memory. The goal is not to create the world's fastest parser anyway. I'd like
+   * to keep this algorithm simple. By launching multiple threads, we can improve
+   * the speed of parsing across a large number of files.</p>
+   *
+   * <p>
+   * There is no strict ordering between the amount of input used by SLL vs LL,
+   * which makes it really hard to build a cache for full context. Let's say that
+   * we have input A B C that leads to an SLL conflict with full context X. That
+   * implies that using X we might only use A B but we could also use A B C D to
+   * resolve conflict. Input A B C D could predict alternative 1 in one position
+   * in the input and A B C E could predict alternative 2 in another position in
+   * input. The conflicting SLL configurations could still be non-unique in the
+   * full context prediction, which would lead us to requiring more input than the
+   * original A B C.	To make a	prediction cache work, we have to track	the exact
+   * input	used during the previous prediction. That amounts to a cache that maps
+   * X to a specific DFA for that context.</p>
+   *
+   * <p>
+   * Something should be done for left-recursive expression predictions. They are
+   * likely LL(1) + pred eval. Easier to do the whole SLL unless error and retry
+   * with full LL thing Sam does.</p>
+   *
+   * <p>
+   * <strong>AVOIDING FULL CONTEXT PREDICTION</strong></p>
+   *
+   * <p>
+   * We avoid doing full context retry when the outer context is empty, we did not
+   * dip into the outer context by falling off the end of the decision state rule,
+   * or when we force SLL mode.</p>
+   *
+   * <p>
+   * As an example of the not dip into outer context case, consider as super
+   * constructor calls versus function calls. One grammar might look like
+   * this:</p>
+   *
+   * <pre>
+   * ctorBody
+   *   : '{' superCall? stat* '}'
+   *   ;
+   * </pre>
+   *
+   * <p>
+   * Or, you might see something like</p>
+   *
+   * <pre>
+   * stat
+   *   : superCall ';'
+   *   | expression ';'
+   *   | ...
+   *   ;
+   * </pre>
+   *
+   * <p>
+   * In both cases I believe that no closure operations will dip into the outer
+   * context. In the first case ctorBody in the worst case will stop at the '}'.
+   * In the 2nd case it should stop at the ';'. Both cases should stay within the
+   * entry rule and not dip into the outer context.</p>
+   *
+   * <p>
+   * <strong>PREDICATES</strong></p>
+   *
+   * <p>
+   * Predicates are always evaluated if present in either SLL or LL both. SLL and
+   * LL simulation deals with predicates differently. SLL collects predicates as
+   * it performs closure operations like ANTLR v3 did. It delays predicate
+   * evaluation until it reaches and accept state. This allows us to cache the SLL
+   * ATN simulation whereas, if we had evaluated predicates on-the-fly during
+   * closure, the DFA state configuration sets would be different and we couldn't
+   * build up a suitable DFA.</p>
+   *
+   * <p>
+   * When building a DFA accept state during ATN simulation, we evaluate any
+   * predicates and return the sole semantically valid alternative. If there is
+   * more than 1 alternative, we report an ambiguity. If there are 0 alternatives,
+   * we throw an exception. Alternatives without predicates act like they have
+   * true predicates. The simple way to think about it is to strip away all
+   * alternatives with false predicates and choose the minimum alternative that
+   * remains.</p>
+   *
+   * <p>
+   * When we start in the DFA and reach an accept state that's predicated, we test
+   * those and return the minimum semantically viable alternative. If no
+   * alternatives are viable, we throw an exception.</p>
+   *
+   * <p>
+   * During full LL ATN simulation, closure always evaluates predicates and
+   * on-the-fly. This is crucial to reducing the configuration set size during
+   * closure. It hits a landmine when parsing with the Java grammar, for example,
+   * without this on-the-fly evaluation.</p>
+   *
+   * <p>
+   * <strong>SHARING DFA</strong></p>
+   *
+   * <p>
+   * All instances of the same parser share the same decision DFAs through a
+   * static field. Each instance gets its own ATN simulator but they share the
+   * same {@link #decisionToDFA} field. They also share a
+   * {@link PredictionContextCache} object that makes sure that all
+   * {@link PredictionContext} objects are shared among the DFA states. This makes
+   * a big size difference.</p>
+   *
+   * <p>
+   * <strong>THREAD SAFETY</strong></p>
+   *
+   * <p>
+   * The {@link ParserATNSimulator} locks on the {@link #decisionToDFA} field when
+   * it adds a new DFA object to that array. {@link #addDFAEdge}
+   * locks on the DFA for the current decision when setting the
+   * {@link DFAState#edges} field. {@link #addDFAState} locks on
+   * the DFA for the current decision when looking up a DFA state to see if it
+   * already exists. We must make sure that all requests to add DFA states that
+   * are equivalent result in the same shared DFA object. This is because lots of
+   * threads will be trying to update the DFA at once. The
+   * {@link #addDFAState} method also locks inside the DFA lock
+   * but this time on the shared context cache when it rebuilds the
+   * configurations' {@link PredictionContext} objects using cached
+   * subgraphs/nodes. No other locking occurs, even during DFA simulation. This is
+   * safe as long as we can guarantee that all threads referencing
+   * {@code s.edge[t]} get the same physical target {@link DFAState}, or
+   * {@code null}. Once into the DFA, the DFA simulation does not reference the
+   * {@link DFA#states} map. It follows the {@link DFAState#edges} field to new
+   * targets. The DFA simulator will either find {@link DFAState#edges} to be
+   * {@code null}, to be non-{@code null} and {@code dfa.edges[t]} null, or
+   * {@code dfa.edges[t]} to be non-null. The
+   * {@link #addDFAEdge} method could be racing to set the field
+   * but in either case the DFA simulator works; if {@code null}, and requests ATN
+   * simulation. It could also race trying to get {@code dfa.edges[t]}, but either
+   * way it will work because it's not doing a test and set operation.</p>
+   *
+   * <p>
+   * <strong>Starting with SLL then failing to combined SLL/LL (Two-Stage
+   * Parsing)</strong></p>
+   *
+   * <p>
+   * Sam pointed out that if SLL does not give a syntax error, then there is no
+   * point in doing full LL, which is slower. We only have to try LL if we get a
+   * syntax error. For maximum speed, Sam starts the parser set to pure SLL
+   * mode with the {@link BailErrorStrategy}:</p>
+   *
+   * <pre>
+   * parser.{@link Parser#getInterpreter() getInterpreter()}.{@link #setPredictionMode setPredictionMode}{@code (}{@link PredictionMode#SLL}{@code )};
+   * parser.{@link Parser#setErrorHandler setErrorHandler}(new {@link BailErrorStrategy}());
+   * </pre>
+   *
+   * <p>
+   * If it does not get a syntax error, then we're done. If it does get a syntax
+   * error, we need to retry with the combined SLL/LL strategy.</p>
+   *
+   * <p>
+   * The reason this works is as follows. If there are no SLL conflicts, then the
+   * grammar is SLL (at least for that input set). If there is an SLL conflict,
+   * the full LL analysis must yield a set of viable alternatives which is a
+   * subset of the alternatives reported by SLL. If the LL set is a singleton,
+   * then the grammar is LL but not SLL. If the LL set is the same size as the SLL
+   * set, the decision is SLL. If the LL set has size &gt; 1, then that decision
+   * is truly ambiguous on the current input. If the LL set is smaller, then the
+   * SLL conflict resolution might choose an alternative that the full LL would
+   * rule out as a possibility based upon better context information. If that's
+   * the case, then the SLL parse will definitely get an error because the full LL
+   * analysis says it's not viable. If SLL conflict resolution chooses an
+   * alternative within the LL set, them both SLL and LL would choose the same
+   * alternative because they both choose the minimum of multiple conflicting
+   * alternatives.</p>
+   *
+   * <p>
+   * Let's say we have a set of SLL conflicting alternatives {@code {1, 2, 3}} and
+   * a smaller LL set called <em>s</em>. If <em>s</em> is {@code {2, 3}}, then SLL
+   * parsing will get an error because SLL will pursue alternative 1. If
+   * <em>s</em> is {@code {1, 2}} or {@code {1, 3}} then both SLL and LL will
+   * choose the same alternative because alternative one is the minimum of either
+   * set. If <em>s</em> is {@code {2}} or {@code {3}} then SLL will get a syntax
+   * error. If <em>s</em> is {@code {1}} then SLL will succeed.</p>
+   *
+   * <p>
+   * Of course, if the input is invalid, then we will get an error for sure in
+   * both SLL and LL parsing. Erroneous input will therefore require 2 passes over
+   * the input.</p>
+   */
+  class ANTLR4CPP_PUBLIC ParserATNSimulator : public ATNSimulator {
+  public:
+    /// Testing only!
+    ParserATNSimulator(const ATN &atn, std::vector<dfa::DFA> &decisionToDFA,
+                       PredictionContextCache &sharedContextCache);
+
+    ParserATNSimulator(Parser *parser, const ATN &atn, std::vector<dfa::DFA> &decisionToDFA,
+                       PredictionContextCache &sharedContextCache);
+
+    virtual void reset() override;
+    virtual void clearDFA() override;
+    virtual size_t adaptivePredict(TokenStream *input, size_t decision, ParserRuleContext *outerContext);
+    
+    static const bool TURN_OFF_LR_LOOP_ENTRY_BRANCH_OPT;
+
+    std::vector<dfa::DFA> &decisionToDFA;
+    
+    /** Implements first-edge (loop entry) elimination as an optimization
+     *  during closure operations.  See antlr/antlr4#1398.
+     *
+     * The optimization is to avoid adding the loop entry config when
+     * the exit path can only lead back to the same
+     * StarLoopEntryState after popping context at the rule end state
+     * (traversing only epsilon edges, so we're still in closure, in
+     * this same rule).
+     *
+     * We need to detect any state that can reach loop entry on
+     * epsilon w/o exiting rule. We don't have to look at FOLLOW
+     * links, just ensure that all stack tops for config refer to key
+     * states in LR rule.
+     *
+     * To verify we are in the right situation we must first check
+     * closure is at a StarLoopEntryState generated during LR removal.
+     * Then we check that each stack top of context is a return state
+     * from one of these cases:
+     *
+     *   1. 'not' expr, '(' type ')' expr. The return state points at loop entry state
+     *   2. expr op expr. The return state is the block end of internal block of (...)*
+     *   3. 'between' expr 'and' expr. The return state of 2nd expr reference.
+     *      That state points at block end of internal block of (...)*.
+     *   4. expr '?' expr ':' expr. The return state points at block end,
+     *      which points at loop entry state.
+     *
+     * If any is true for each stack top, then closure does not add a
+     * config to the current config set for edge[0], the loop entry branch.
+     *
+     *  Conditions fail if any context for the current config is:
+     *
+     *   a. empty (we'd fall out of expr to do a global FOLLOW which could
+     *      even be to some weird spot in expr) or,
+     *   b. lies outside of expr or,
+     *   c. lies within expr but at a state not the BlockEndState
+     *   generated during LR removal
+     *
+     * Do we need to evaluate predicates ever in closure for this case?
+     *
+     * No. Predicates, including precedence predicates, are only
+     * evaluated when computing a DFA start state. I.e., only before
+     * the lookahead (but not parser) consumes a token.
+     *
+     * There are no epsilon edges allowed in LR rule alt blocks or in
+     * the "primary" part (ID here). If closure is in
+     * StarLoopEntryState any lookahead operation will have consumed a
+     * token as there are no epsilon-paths that lead to
+     * StarLoopEntryState. We do not have to evaluate predicates
+     * therefore if we are in the generated StarLoopEntryState of a LR
+     * rule. Note that when making a prediction starting at that
+     * decision point, decision d=2, compute-start-state performs
+     * closure starting at edges[0], edges[1] emanating from
+     * StarLoopEntryState. That means it is not performing closure on
+     * StarLoopEntryState during compute-start-state.
+     *
+     * How do we know this always gives same prediction answer?
+     *
+     * Without predicates, loop entry and exit paths are ambiguous
+     * upon remaining input +b (in, say, a+b). Either paths lead to
+     * valid parses. Closure can lead to consuming + immediately or by
+     * falling out of this call to expr back into expr and loop back
+     * again to StarLoopEntryState to match +b. In this special case,
+     * we choose the more efficient path, which is to take the bypass
+     * path.
+     *
+     * The lookahead language has not changed because closure chooses
+     * one path over the other. Both paths lead to consuming the same
+     * remaining input during a lookahead operation. If the next token
+     * is an operator, lookahead will enter the choice block with
+     * operators. If it is not, lookahead will exit expr. Same as if
+     * closure had chosen to enter the choice block immediately.
+     *
+     * Closure is examining one config (some loopentrystate, some alt,
+     * context) which means it is considering exactly one alt. Closure
+     * always copies the same alt to any derived configs.
+     *
+     * How do we know this optimization doesn't mess up precedence in
+     * our parse trees?
+     *
+     * Looking through expr from left edge of stat only has to confirm
+     * that an input, say, a+b+c; begins with any valid interpretation
+     * of an expression. The precedence actually doesn't matter when
+     * making a decision in stat seeing through expr. It is only when
+     * parsing rule expr that we must use the precedence to get the
+     * right interpretation and, hence, parse tree.
+     */
+    bool canDropLoopEntryEdgeInLeftRecursiveRule(ATNConfig *config) const;
+    virtual std::string getRuleName(size_t index);
+
+    virtual Ref<ATNConfig> precedenceTransition(Ref<ATNConfig> const& config, PrecedencePredicateTransition *pt,
+                                                bool collectPredicates, bool inContext, bool fullCtx);
+
+    void setPredictionMode(PredictionMode newMode);
+    PredictionMode getPredictionMode();
+
+    Parser* getParser();
+    
+    virtual std::string getTokenName(size_t t);
+
+    virtual std::string getLookaheadName(TokenStream *input);
+
+    /// <summary>
+    /// Used for debugging in adaptivePredict around execATN but I cut
+    ///  it out for clarity now that alg. works well. We can leave this
+    ///  "dead" code for a bit.
+    /// </summary>
+    virtual void dumpDeadEndConfigs(NoViableAltException &nvae);
+    
+  protected:
+    Parser *const parser;
+
+    /// <summary>
+    /// Each prediction operation uses a cache for merge of prediction contexts.
+    /// Don't keep around as it wastes huge amounts of memory. The merge cache
+    /// isn't synchronized but we're ok since two threads shouldn't reuse same
+    /// parser/atnsim object because it can only handle one input at a time.
+    /// This maps graphs a and b to merged result c. (a,b)->c. We can avoid
+    /// the merge if we ever see a and b again.  Note that (b,a)->c should
+    /// also be examined during cache lookup.
+    /// </summary>
+    PredictionContextMergeCache mergeCache;
+
+    // LAME globals to avoid parameters!!!!! I need these down deep in predTransition
+    TokenStream *_input;
+    size_t _startIndex;
+    ParserRuleContext *_outerContext;
+    dfa::DFA *_dfa; // Reference into the decisionToDFA vector.
+    
+    /// <summary>
+    /// Performs ATN simulation to compute a predicted alternative based
+    ///  upon the remaining input, but also updates the DFA cache to avoid
+    ///  having to traverse the ATN again for the same input sequence.
+    ///
+    /// There are some key conditions we're looking for after computing a new
+    /// set of ATN configs (proposed DFA state):
+    /// if the set is empty, there is no viable alternative for current symbol
+    /// does the state uniquely predict an alternative?
+    /// does the state have a conflict that would prevent us from
+    ///         putting it on the work list?
+    ///
+    /// We also have some key operations to do:
+    /// add an edge from previous DFA state to potentially new DFA state, D,
+    ///         upon current symbol but only if adding to work list, which means in all
+    ///         cases except no viable alternative (and possibly non-greedy decisions?)
+    /// collecting predicates and adding semantic context to DFA accept states
+    /// adding rule context to context-sensitive DFA accept states
+    /// consuming an input symbol
+    /// reporting a conflict
+    /// reporting an ambiguity
+    /// reporting a context sensitivity
+    /// reporting insufficient predicates
+    ///
+    /// cover these cases:
+    ///    dead end
+    ///    single alt
+    ///    single alt + preds
+    ///    conflict
+    ///    conflict + preds
+    /// </summary>
+    virtual size_t execATN(dfa::DFA &dfa, dfa::DFAState *s0, TokenStream *input, size_t startIndex,
+                           ParserRuleContext *outerContext);
+
+    /// <summary>
+    /// Get an existing target state for an edge in the DFA. If the target state
+    /// for the edge has not yet been computed or is otherwise not available,
+    /// this method returns {@code null}.
+    /// </summary>
+    /// <param name="previousD"> The current DFA state </param>
+    /// <param name="t"> The next input symbol </param>
+    /// <returns> The existing target DFA state for the given input symbol
+    /// {@code t}, or {@code null} if the target state for this edge is not
+    /// already cached </returns>
+    virtual dfa::DFAState* getExistingTargetState(dfa::DFAState *previousD, size_t t);
+
+    /// <summary>
+    /// Compute a target state for an edge in the DFA, and attempt to add the
+    /// computed state and corresponding edge to the DFA.
+    /// </summary>
+    /// <param name="dfa"> The DFA </param>
+    /// <param name="previousD"> The current DFA state </param>
+    /// <param name="t"> The next input symbol
+    /// </param>
+    /// <returns> The computed target DFA state for the given input symbol
+    /// {@code t}. If {@code t} does not lead to a valid DFA state, this method
+    /// returns <seealso cref="#ERROR"/>. </returns>
+    virtual dfa::DFAState *computeTargetState(dfa::DFA &dfa, dfa::DFAState *previousD, size_t t);
+
+    virtual void predicateDFAState(dfa::DFAState *dfaState, DecisionState *decisionState);
+
+    // comes back with reach.uniqueAlt set to a valid alt
+    virtual size_t execATNWithFullContext(dfa::DFA &dfa, dfa::DFAState *D, ATNConfigSet *s0,
+                                          TokenStream *input, size_t startIndex, ParserRuleContext *outerContext); // how far we got before failing over
+
+    virtual std::unique_ptr<ATNConfigSet> computeReachSet(ATNConfigSet *closure, size_t t, bool fullCtx);
+
+    /// <summary>
+    /// Return a configuration set containing only the configurations from
+    /// {@code configs} which are in a <seealso cref="RuleStopState"/>. If all
+    /// configurations in {@code configs} are already in a rule stop state, this
+    /// method simply returns {@code configs}.
+    /// <p/>
+    /// When {@code lookToEndOfRule} is true, this method uses
+    /// <seealso cref="ATN#nextTokens"/> for each configuration in {@code configs} which is
+    /// not already in a rule stop state to see if a rule stop state is reachable
+    /// from the configuration via epsilon-only transitions.
+    /// </summary>
+    /// <param name="configs"> the configuration set to update </param>
+    /// <param name="lookToEndOfRule"> when true, this method checks for rule stop states
+    /// reachable by epsilon-only transitions from each configuration in
+    /// {@code configs}.
+    /// </param>
+    /// <returns> {@code configs} if all configurations in {@code configs} are in a
+    /// rule stop state, otherwise return a new configuration set containing only
+    /// the configurations from {@code configs} which are in a rule stop state </returns>
+    virtual ATNConfigSet* removeAllConfigsNotInRuleStopState(ATNConfigSet *configs, bool lookToEndOfRule);
+
+    virtual std::unique_ptr<ATNConfigSet> computeStartState(ATNState *p, RuleContext *ctx, bool fullCtx);
+
+    /* parrt internal source braindump that doesn't mess up
+     * external API spec.
+
+     applyPrecedenceFilter is an optimization to avoid highly
+     nonlinear prediction of expressions and other left recursive
+     rules. The precedence predicates such as {3>=prec}? Are highly
+     context-sensitive in that they can only be properly evaluated
+     in the context of the proper prec argument. Without pruning,
+     these predicates are normal predicates evaluated when we reach
+     conflict state (or unique prediction). As we cannot evaluate
+     these predicates out of context, the resulting conflict leads
+     to full LL evaluation and nonlinear prediction which shows up
+     very clearly with fairly large expressions.
+
+     Example grammar:
+
+     e : e '*' e
+     | e '+' e
+     | INT
+     ;
+
+     We convert that to the following:
+
+     e[int prec]
+     :   INT
+     ( {3>=prec}? '*' e[4]
+     | {2>=prec}? '+' e[3]
+     )*
+     ;
+
+     The (..)* loop has a decision for the inner block as well as
+     an enter or exit decision, which is what concerns us here. At
+     the 1st + of input 1+2+3, the loop entry sees both predicates
+     and the loop exit also sees both predicates by falling off the
+     edge of e.  This is because we have no stack information with
+     SLL and find the follow of e, which will hit the return states
+     inside the loop after e[4] and e[3], which brings it back to
+     the enter or exit decision. In this case, we know that we
+     cannot evaluate those predicates because we have fallen off
+     the edge of the stack and will in general not know which prec
+     parameter is the right one to use in the predicate.
+
+     Because we have special information, that these are precedence
+     predicates, we can resolve them without failing over to full
+     LL despite their context sensitive nature. We make an
+     assumption that prec[-1] <= prec[0], meaning that the current
+     precedence level is greater than or equal to the precedence
+     level of recursive invocations above us in the stack. For
+     example, if predicate {3>=prec}? is true of the current prec,
+     then one option is to enter the loop to match it now. The
+     other option is to exit the loop and the left recursive rule
+     to match the current operator in rule invocation further up
+     the stack. But, we know that all of those prec are lower or
+     the same value and so we can decide to enter the loop instead
+     of matching it later. That means we can strip out the other
+     configuration for the exit branch.
+
+     So imagine we have (14,1,$,{2>=prec}?) and then
+     (14,2,$-dipsIntoOuterContext,{2>=prec}?). The optimization
+     allows us to collapse these two configurations. We know that
+     if {2>=prec}? is true for the current prec parameter, it will
+     also be true for any prec from an invoking e call, indicated
+     by dipsIntoOuterContext. As the predicates are both true, we
+     have the option to evaluate them early in the decision start
+     state. We do this by stripping both predicates and choosing to
+     enter the loop as it is consistent with the notion of operator
+     precedence. It's also how the full LL conflict resolution
+     would work.
+
+     The solution requires a different DFA start state for each
+     precedence level.
+
+     The basic filter mechanism is to remove configurations of the
+     form (p, 2, pi) if (p, 1, pi) exists for the same p and pi. In
+     other words, for the same ATN state and predicate context,
+     remove any configuration associated with an exit branch if
+     there is a configuration associated with the enter branch.
+
+     It's also the case that the filter evaluates precedence
+     predicates and resolves conflicts according to precedence
+     levels. For example, for input 1+2+3 at the first +, we see
+     prediction filtering
+
+     [(11,1,[$],{3>=prec}?), (14,1,[$],{2>=prec}?), (5,2,[$],up=1),
+     (11,2,[$],up=1), (14,2,[$],up=1)],hasSemanticContext=true,dipsIntoOuterContext
+
+     to
+
+     [(11,1,[$]), (14,1,[$]), (5,2,[$],up=1)],dipsIntoOuterContext
+
+     This filters because {3>=prec}? evals to true and collapses
+     (11,1,[$],{3>=prec}?) and (11,2,[$],up=1) since early conflict
+     resolution based upon rules of operator precedence fits with
+     our usual match first alt upon conflict.
+
+     We noticed a problem where a recursive call resets precedence
+     to 0. Sam's fix: each config has flag indicating if it has
+     returned from an expr[0] call. then just don't filter any
+     config with that flag set. flag is carried along in
+     closure(). so to avoid adding field, set bit just under sign
+     bit of dipsIntoOuterContext (SUPPRESS_PRECEDENCE_FILTER).
+     With the change you filter "unless (p, 2, pi) was reached
+     after leaving the rule stop state of the LR rule containing
+     state p, corresponding to a rule invocation with precedence
+     level 0"
+     */
+
+    /**
+     * This method transforms the start state computed by
+     * {@link #computeStartState} to the special start state used by a
+     * precedence DFA for a particular precedence value. The transformation
+     * process applies the following changes to the start state's configuration
+     * set.
+     *
+     * <ol>
+     * <li>Evaluate the precedence predicates for each configuration using
+     * {@link SemanticContext#evalPrecedence}.</li>
+     * <li>When {@link ATNConfig#isPrecedenceFilterSuppressed} is {@code false},
+     * remove all configurations which predict an alternative greater than 1,
+     * for which another configuration that predicts alternative 1 is in the
+     * same ATN state with the same prediction context. This transformation is
+     * valid for the following reasons:
+     * <ul>
+     * <li>The closure block cannot contain any epsilon transitions which bypass
+     * the body of the closure, so all states reachable via alternative 1 are
+     * part of the precedence alternatives of the transformed left-recursive
+     * rule.</li>
+     * <li>The "primary" portion of a left recursive rule cannot contain an
+     * epsilon transition, so the only way an alternative other than 1 can exist
+     * in a state that is also reachable via alternative 1 is by nesting calls
+     * to the left-recursive rule, with the outer calls not being at the
+     * preferred precedence level. The
+     * {@link ATNConfig#isPrecedenceFilterSuppressed} property marks ATN
+     * configurations which do not meet this condition, and therefore are not
+     * eligible for elimination during the filtering process.</li>
+     * </ul>
+     * </li>
+     * </ol>
+     *
+     * <p>
+     * The prediction context must be considered by this filter to address
+     * situations like the following.
+     * </p>
+     * <code>
+     * <pre>
+     * grammar TA;
+     * prog: statement* EOF;
+     * statement: letterA | statement letterA 'b' ;
+     * letterA: 'a';
+     * </pre>
+     * </code>
+     * <p>
+     * If the above grammar, the ATN state immediately before the token
+     * reference {@code 'a'} in {@code letterA} is reachable from the left edge
+     * of both the primary and closure blocks of the left-recursive rule
+     * {@code statement}. The prediction context associated with each of these
+     * configurations distinguishes between them, and prevents the alternative
+     * which stepped out to {@code prog} (and then back in to {@code statement}
+     * from being eliminated by the filter.
+     * </p>
+     *
+     * @param configs The configuration set computed by
+     * {@link #computeStartState} as the start state for the DFA.
+     * @return The transformed configuration set representing the start state
+     * for a precedence DFA at a particular precedence level (determined by
+     * calling {@link Parser#getPrecedence}).
+     */
+    std::unique_ptr<ATNConfigSet> applyPrecedenceFilter(ATNConfigSet *configs);
+
+    virtual ATNState *getReachableTarget(Transition *trans, size_t ttype);
+
+    virtual std::vector<Ref<SemanticContext>> getPredsForAmbigAlts(const antlrcpp::BitSet &ambigAlts,
+                                                                   ATNConfigSet *configs, size_t nalts);
+
+    virtual std::vector<dfa::DFAState::PredPrediction*> getPredicatePredictions(const antlrcpp::BitSet &ambigAlts,
+                                                                                std::vector<Ref<SemanticContext>> altToPred);
+
+    /**
+     * This method is used to improve the localization of error messages by
+     * choosing an alternative rather than throwing a
+     * {@link NoViableAltException} in particular prediction scenarios where the
+     * {@link #ERROR} state was reached during ATN simulation.
+     *
+     * <p>
+     * The default implementation of this method uses the following
+     * algorithm to identify an ATN configuration which successfully parsed the
+     * decision entry rule. Choosing such an alternative ensures that the
+     * {@link ParserRuleContext} returned by the calling rule will be complete
+     * and valid, and the syntax error will be reported later at a more
+     * localized location.</p>
+     *
+     * <ul>
+     * <li>If a syntactically valid path or paths reach the end of the decision rule and
+     * they are semantically valid if predicated, return the min associated alt.</li>
+     * <li>Else, if a semantically invalid but syntactically valid path exist
+     * or paths exist, return the minimum associated alt.
+     * </li>
+     * <li>Otherwise, return {@link ATN#INVALID_ALT_NUMBER}.</li>
+     * </ul>
+     *
+     * <p>
+     * In some scenarios, the algorithm described above could predict an
+     * alternative which will result in a {@link FailedPredicateException} in
+     * the parser. Specifically, this could occur if the <em>only</em> configuration
+     * capable of successfully parsing to the end of the decision rule is
+     * blocked by a semantic predicate. By choosing this alternative within
+     * {@link #adaptivePredict} instead of throwing a
+     * {@link NoViableAltException}, the resulting
+     * {@link FailedPredicateException} in the parser will identify the specific
+     * predicate which is preventing the parser from successfully parsing the
+     * decision rule, which helps developers identify and correct logic errors
+     * in semantic predicates.
+     * </p>
+     *
+     * @param configs The ATN configurations which were valid immediately before
+     * the {@link #ERROR} state was reached
+     * @param outerContext The is the \gamma_0 initial parser context from the paper
+     * or the parser stack at the instant before prediction commences.
+     *
+     * @return The value to return from {@link #adaptivePredict}, or
+     * {@link ATN#INVALID_ALT_NUMBER} if a suitable alternative was not
+     * identified and {@link #adaptivePredict} should report an error instead.
+     */
+    size_t getSynValidOrSemInvalidAltThatFinishedDecisionEntryRule(ATNConfigSet *configs,
+                                                                   ParserRuleContext *outerContext);
+
+    virtual size_t getAltThatFinishedDecisionEntryRule(ATNConfigSet *configs);
+
+    /** Walk the list of configurations and split them according to
+     *  those that have preds evaluating to true/false.  If no pred, assume
+     *  true pred and include in succeeded set.  Returns Pair of sets.
+     *
+     *  Create a new set so as not to alter the incoming parameter.
+     *
+     *  Assumption: the input stream has been restored to the starting point
+     *  prediction, which is where predicates need to evaluate.
+     */
+    std::pair<ATNConfigSet *, ATNConfigSet *> splitAccordingToSemanticValidity(ATNConfigSet *configs,
+                                                                               ParserRuleContext *outerContext);
+
+    /// <summary>
+    /// Look through a list of predicate/alt pairs, returning alts for the
+    ///  pairs that win. A {@code NONE} predicate indicates an alt containing an
+    ///  unpredicated config which behaves as "always true." If !complete
+    ///  then we stop at the first predicate that evaluates to true. This
+    ///  includes pairs with null predicates.
+    /// </summary>
+    virtual antlrcpp::BitSet evalSemanticContext(std::vector<dfa::DFAState::PredPrediction*> predPredictions,
+                                                 ParserRuleContext *outerContext, bool complete);
+
+    /**
+     * Evaluate a semantic context within a specific parser context.
+     *
+     * <p>
+     * This method might not be called for every semantic context evaluated
+     * during the prediction process. In particular, we currently do not
+     * evaluate the following but it may change in the future:</p>
+     *
+     * <ul>
+     * <li>Precedence predicates (represented by
+     * {@link SemanticContext.PrecedencePredicate}) are not currently evaluated
+     * through this method.</li>
+     * <li>Operator predicates (represented by {@link SemanticContext.AND} and
+     * {@link SemanticContext.OR}) are evaluated as a single semantic
+     * context, rather than evaluating the operands individually.
+     * Implementations which require evaluation results from individual
+     * predicates should override this method to explicitly handle evaluation of
+     * the operands within operator predicates.</li>
+     * </ul>
+     *
+     * @param pred The semantic context to evaluate
+     * @param parserCallStack The parser context in which to evaluate the
+     * semantic context
+     * @param alt The alternative which is guarded by {@code pred}
+     * @param fullCtx {@code true} if the evaluation is occurring during LL
+     * prediction; otherwise, {@code false} if the evaluation is occurring
+     * during SLL prediction
+     *
+     * @since 4.3
+     */
+    virtual bool evalSemanticContext(Ref<SemanticContext> const& pred, ParserRuleContext *parserCallStack,
+                                     size_t alt, bool fullCtx);
+
+    /* TO_DO: If we are doing predicates, there is no point in pursuing
+     closure operations if we reach a DFA state that uniquely predicts
+     alternative. We will not be caching that DFA state and it is a
+     waste to pursue the closure. Might have to advance when we do
+     ambig detection thought :(
+     */
+    virtual void closure(Ref<ATNConfig> const& config, ATNConfigSet *configs, ATNConfig::Set &closureBusy,
+                         bool collectPredicates, bool fullCtx, bool treatEofAsEpsilon);
+
+    virtual void closureCheckingStopState(Ref<ATNConfig> const& config, ATNConfigSet *configs, ATNConfig::Set &closureBusy,
+                                          bool collectPredicates, bool fullCtx, int depth, bool treatEofAsEpsilon);
+    
+    /// Do the actual work of walking epsilon edges.
+    virtual void closure_(Ref<ATNConfig> const& config, ATNConfigSet *configs, ATNConfig::Set &closureBusy,
+                          bool collectPredicates, bool fullCtx, int depth, bool treatEofAsEpsilon);
+    
+    virtual Ref<ATNConfig> getEpsilonTarget(Ref<ATNConfig> const& config, Transition *t, bool collectPredicates,
+                                            bool inContext, bool fullCtx, bool treatEofAsEpsilon);
+    virtual Ref<ATNConfig> actionTransition(Ref<ATNConfig> const& config, ActionTransition *t);
+
+    virtual Ref<ATNConfig> predTransition(Ref<ATNConfig> const& config, PredicateTransition *pt, bool collectPredicates,
+                                          bool inContext, bool fullCtx);
+
+    virtual Ref<ATNConfig> ruleTransition(Ref<ATNConfig> const& config, RuleTransition *t);
+
+    /**
+     * Gets a {@link BitSet} containing the alternatives in {@code configs}
+     * which are part of one or more conflicting alternative subsets.
+     *
+     * @param configs The {@link ATNConfigSet} to analyze.
+     * @return The alternatives in {@code configs} which are part of one or more
+     * conflicting alternative subsets. If {@code configs} does not contain any
+     * conflicting subsets, this method returns an empty {@link BitSet}.
+     */
+    virtual antlrcpp::BitSet getConflictingAlts(ATNConfigSet *configs);
+
+    /// <summary>
+    /// Sam pointed out a problem with the previous definition, v3, of
+    /// ambiguous states. If we have another state associated with conflicting
+    /// alternatives, we should keep going. For example, the following grammar
+    ///
+    /// s : (ID | ID ID?) ';' ;
+    ///
+    /// When the ATN simulation reaches the state before ';', it has a DFA
+    /// state that looks like: [12|1|[], 6|2|[], 12|2|[]]. Naturally
+    /// 12|1|[] and 12|2|[] conflict, but we cannot stop processing this node
+    /// because alternative to has another way to continue, via [6|2|[]].
+    /// The key is that we have a single state that has config's only associated
+    /// with a single alternative, 2, and crucially the state transitions
+    /// among the configurations are all non-epsilon transitions. That means
+    /// we don't consider any conflicts that include alternative 2. So, we
+    /// ignore the conflict between alts 1 and 2. We ignore a set of
+    /// conflicting alts when there is an intersection with an alternative
+    /// associated with a single alt state in the state->config-list map.
+    ///
+    /// It's also the case that we might have two conflicting configurations but
+    /// also a 3rd nonconflicting configuration for a different alternative:
+    /// [1|1|[], 1|2|[], 8|3|[]]. This can come about from grammar:
+    ///
+    /// a : A | A | A B ;
+    ///
+    /// After matching input A, we reach the stop state for rule A, state 1.
+    /// State 8 is the state right before B. Clearly alternatives 1 and 2
+    /// conflict and no amount of further lookahead will separate the two.
+    /// However, alternative 3 will be able to continue and so we do not
+    /// stop working on this state. In the previous example, we're concerned
+    /// with states associated with the conflicting alternatives. Here alt
+    /// 3 is not associated with the conflicting configs, but since we can continue
+    /// looking for input reasonably, I don't declare the state done. We
+    /// ignore a set of conflicting alts when we have an alternative
+    /// that we still need to pursue.
+    /// </summary>
+
+    virtual antlrcpp::BitSet getConflictingAltsOrUniqueAlt(ATNConfigSet *configs);
+
+    virtual NoViableAltException noViableAlt(TokenStream *input, ParserRuleContext *outerContext,
+                                              ATNConfigSet *configs, size_t startIndex);
+
+    static size_t getUniqueAlt(ATNConfigSet *configs);
+
+    /// <summary>
+    /// Add an edge to the DFA, if possible. This method calls
+    /// <seealso cref="#addDFAState"/> to ensure the {@code to} state is present in the
+    /// DFA. If {@code from} is {@code null}, or if {@code t} is outside the
+    /// range of edges that can be represented in the DFA tables, this method
+    /// returns without adding the edge to the DFA.
+    /// <p/>
+    /// If {@code to} is {@code null}, this method returns {@code null}.
+    /// Otherwise, this method returns the <seealso cref="DFAState"/> returned by calling
+    /// <seealso cref="#addDFAState"/> for the {@code to} state.
+    /// </summary>
+    /// <param name="dfa"> The DFA </param>
+    /// <param name="from"> The source state for the edge </param>
+    /// <param name="t"> The input symbol </param>
+    /// <param name="to"> The target state for the edge
+    /// </param>
+    /// <returns> If {@code to} is {@code null}, this method returns {@code null};
+    /// otherwise this method returns the result of calling <seealso cref="#addDFAState"/>
+    /// on {@code to} </returns>
+    virtual dfa::DFAState *addDFAEdge(dfa::DFA &dfa, dfa::DFAState *from, ssize_t t, dfa::DFAState *to);
+
+    /// <summary>
+    /// Add state {@code D} to the DFA if it is not already present, and return
+    /// the actual instance stored in the DFA. If a state equivalent to {@code D}
+    /// is already in the DFA, the existing state is returned. Otherwise this
+    /// method returns {@code D} after adding it to the DFA.
+    /// <p/>
+    /// If {@code D} is <seealso cref="#ERROR"/>, this method returns <seealso cref="#ERROR"/> and
+    /// does not change the DFA.
+    /// </summary>
+    /// <param name="dfa"> The dfa </param>
+    /// <param name="D"> The DFA state to add </param>
+    /// <returns> The state stored in the DFA. This will be either the existing
+    /// state if {@code D} is already in the DFA, or {@code D} itself if the
+    /// state was not already present. </returns>
+    virtual dfa::DFAState *addDFAState(dfa::DFA &dfa, dfa::DFAState *D);
+
+    virtual void reportAttemptingFullContext(dfa::DFA &dfa, const antlrcpp::BitSet &conflictingAlts,
+      ATNConfigSet *configs, size_t startIndex, size_t stopIndex);
+
+    virtual void reportContextSensitivity(dfa::DFA &dfa, size_t prediction, ATNConfigSet *configs,
+                                          size_t startIndex, size_t stopIndex);
+
+    /// If context sensitive parsing, we know it's ambiguity not conflict.
+    virtual void reportAmbiguity(dfa::DFA &dfa,
+                                 dfa::DFAState *D, // the DFA state from execATN() that had SLL conflicts
+                                 size_t startIndex, size_t stopIndex,
+                                 bool exact,
+                                 const antlrcpp::BitSet &ambigAlts,
+                                 ATNConfigSet *configs); // configs that LL not SLL considered conflicting
+
+  private:
+    // SLL, LL, or LL + exact ambig detection?
+    PredictionMode _mode;
+
+    static bool getLrLoopSetting();
+    void InitializeInstanceFields();
+  };
+
+} // namespace atn
+} // namespace antlr4
+

runtime-linux/antlr4-runtime/atn/PlusBlockStartState.cpp

@@ -0,0 +1,12 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/PlusBlockStartState.h"
+
+using namespace antlr4::atn;
+
+size_t PlusBlockStartState::getStateType() {
+  return PLUS_BLOCK_START;
+}

runtime-linux/antlr4-runtime/atn/PlusBlockStartState.h

@@ -0,0 +1,25 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/BlockStartState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Start of {@code (A|B|...)+} loop. Technically a decision state, but
+  /// we don't use for code generation; somebody might need it, so I'm defining
+  /// it for completeness. In reality, the <seealso cref="PlusLoopbackState"/> node is the
+  /// real decision-making note for {@code A+}.
+  class ANTLR4CPP_PUBLIC PlusBlockStartState final : public BlockStartState {
+  public:
+    PlusLoopbackState *loopBackState = nullptr;
+
+    virtual size_t getStateType() override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/PlusLoopbackState.cpp

@@ -0,0 +1,12 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/PlusLoopbackState.h"
+
+using namespace antlr4::atn;
+
+size_t PlusLoopbackState::getStateType() {
+  return PLUS_LOOP_BACK;
+}

runtime-linux/antlr4-runtime/atn/PlusLoopbackState.h

@@ -0,0 +1,22 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionState.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// Decision state for {@code A+} and {@code (A|B)+}. It has two transitions:
+  /// one to the loop back to start of the block and one to exit.
+  class ANTLR4CPP_PUBLIC PlusLoopbackState final : public DecisionState {
+
+  public:
+    virtual size_t getStateType() override;
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/PrecedencePredicateTransition.cpp

@@ -0,0 +1,32 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/PrecedencePredicateTransition.h"
+
+using namespace antlr4::atn;
+
+PrecedencePredicateTransition::PrecedencePredicateTransition(ATNState *target, int precedence)
+  : AbstractPredicateTransition(target), precedence(precedence) {
+}
+
+Transition::SerializationType PrecedencePredicateTransition::getSerializationType() const {
+  return PRECEDENCE;
+}
+
+bool PrecedencePredicateTransition::isEpsilon() const {
+  return true;
+}
+
+bool PrecedencePredicateTransition::matches(size_t /*symbol*/, size_t /*minVocabSymbol*/, size_t /*maxVocabSymbol*/) const {
+  return false;
+}
+
+Ref<SemanticContext::PrecedencePredicate> PrecedencePredicateTransition::getPredicate() const {
+  return std::make_shared<SemanticContext::PrecedencePredicate>(precedence);
+}
+
+std::string PrecedencePredicateTransition::toString() const {
+  return "PRECEDENCE " + Transition::toString() + " { precedence: " + std::to_string(precedence) + " }";
+}

runtime-linux/antlr4-runtime/atn/PrecedencePredicateTransition.h

@@ -0,0 +1,29 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/AbstractPredicateTransition.h"
+#include "SemanticContext.h"
+
+namespace antlr4 {
+namespace atn {
+
+  class ANTLR4CPP_PUBLIC PrecedencePredicateTransition final : public AbstractPredicateTransition {
+  public:
+    const int precedence;
+
+    PrecedencePredicateTransition(ATNState *target, int precedence);
+
+    virtual SerializationType getSerializationType() const override;
+    virtual bool isEpsilon() const override;
+    virtual bool matches(size_t symbol, size_t minVocabSymbol, size_t maxVocabSymbol) const override;
+    Ref<SemanticContext::PrecedencePredicate> getPredicate() const;
+    virtual std::string toString() const override;
+
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/PredicateEvalInfo.cpp

@@ -0,0 +1,17 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "SemanticContext.h"
+
+#include "atn/PredicateEvalInfo.h"
+
+using namespace antlr4;
+using namespace antlr4::atn;
+
+PredicateEvalInfo::PredicateEvalInfo(size_t decision, TokenStream *input, size_t startIndex, size_t stopIndex,
+  Ref<SemanticContext> const& semctx, bool evalResult, size_t predictedAlt, bool fullCtx)
+  : DecisionEventInfo(decision, nullptr, input, startIndex, stopIndex, fullCtx),
+    semctx(semctx), predictedAlt(predictedAlt), evalResult(evalResult) {
+}

runtime-linux/antlr4-runtime/atn/PredicateEvalInfo.h

@@ -0,0 +1,62 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#pragma once
+
+#include "atn/DecisionEventInfo.h"
+
+namespace antlr4 {
+namespace atn {
+
+  /// <summary>
+  /// This class represents profiling event information for semantic predicate
+  /// evaluations which occur during prediction.
+  /// </summary>
+  /// <seealso cref= ParserATNSimulator#evalSemanticContext
+  ///
+  /// @since 4.3 </seealso>
+  class ANTLR4CPP_PUBLIC PredicateEvalInfo : public DecisionEventInfo {
+  public:
+    /// The semantic context which was evaluated.
+    const Ref<SemanticContext> semctx;
+
+    /// <summary>
+    /// The alternative number for the decision which is guarded by the semantic
+    /// context <seealso cref="#semctx"/>. Note that other ATN
+    /// configurations may predict the same alternative which are guarded by
+    /// other semantic contexts and/or <seealso cref="SemanticContext#NONE"/>.
+    /// </summary>
+    const size_t predictedAlt;
+
+    /// The result of evaluating the semantic context <seealso cref="#semctx"/>.
+    const bool evalResult;
+
+    /// <summary>
+    /// Constructs a new instance of the <seealso cref="PredicateEvalInfo"/> class with the
+    /// specified detailed predicate evaluation information.
+    /// </summary>
+    /// <param name="decision"> The decision number </param>
+    /// <param name="input"> The input token stream </param>
+    /// <param name="startIndex"> The start index for the current prediction </param>
+    /// <param name="stopIndex"> The index at which the predicate evaluation was
+    /// triggered. Note that the input stream may be reset to other positions for
+    /// the actual evaluation of individual predicates. </param>
+    /// <param name="semctx"> The semantic context which was evaluated </param>
+    /// <param name="evalResult"> The results of evaluating the semantic context </param>
+    /// <param name="predictedAlt"> The alternative number for the decision which is
+    /// guarded by the semantic context {@code semctx}. See <seealso cref="#predictedAlt"/>
+    /// for more information. </param>
+    /// <param name="fullCtx"> {@code true} if the semantic context was
+    /// evaluated during LL prediction; otherwise, {@code false} if the semantic
+    /// context was evaluated during SLL prediction
+    /// </param>
+    /// <seealso cref= ParserATNSimulator#evalSemanticContext(SemanticContext, ParserRuleContext, int, boolean) </seealso>
+    /// <seealso cref= SemanticContext#eval(Recognizer, RuleContext) </seealso>
+    PredicateEvalInfo(size_t decision, TokenStream *input, size_t startIndex, size_t stopIndex,
+                      Ref<SemanticContext> const& semctx, bool evalResult, size_t predictedAlt, bool fullCtx);
+  };
+
+} // namespace atn
+} // namespace antlr4

runtime-linux/antlr4-runtime/atn/PredicateTransition.cpp

@@ -0,0 +1,34 @@
+/* Copyright (c) 2012-2017 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.
+ */
+
+#include "atn/PredicateTransition.h"
+
+using namespace antlr4::atn;
+
+PredicateTransition::PredicateTransition(ATNState *target, size_t ruleIndex, size_t predIndex, bool isCtxDependent) : AbstractPredicateTransition(target), ruleIndex(ruleIndex), predIndex(predIndex), isCtxDependent(isCtxDependent) {
+}
+
+Transition::SerializationType PredicateTransition::getSerializationType() const {
+  return PREDICATE;
+}
+
+bool PredicateTransition::isEpsilon() const {
+  return true;
+}
+
+bool PredicateTransition::matches(size_t /*symbol*/, size_t /*minVocabSymbol*/, size_t /*maxVocabSymbol*/) const {
+  return false;
+}
+
+Ref<SemanticContext::Predicate> PredicateTransition::getPredicate() const {
+  return std::make_shared<SemanticContext::Predicate>(ruleIndex, predIndex, isCtxDependent);
+}
+
+std::string PredicateTransition::toString() const {
+  return "PREDICATE " + Transition::toString() + " { ruleIndex: " + std::to_string(ruleIndex) +
+    ", predIndex: " + std::to_string(predIndex) + ", isCtxDependent: " + std::to_string(isCtxDependent) + " }";
+
+  // Generate and add a predicate context here?
+}