[regexp] implement character classes for unicode regexps.

src/regexp/regexp-parser.h

 // Copyright 2016 the V8 project authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #ifndef V8_REGEXP_REGEXP_PARSER_H_
 #define V8_REGEXP_REGEXP_PARSER_H_
 
 #include "src/objects.h"
 #include "src/regexp/regexp-ast.h"
 #include "src/zone.h"
@@ skipping 81 matching lines @@
 
  private:
   ZoneList<T*>* list_;
   T* last_;
 };
 
 
 // Accumulates RegExp atoms and assertions into lists of terms and alternatives.
 class RegExpBuilder : public ZoneObject {
  public:
-  explicit RegExpBuilder(Zone* zone);
+  RegExpBuilder(Zone* zone, JSRegExp::Flags flags);
   void AddCharacter(uc16 character);
   void AddUnicodeCharacter(uc32 character);
   // "Adds" an empty expression. Does nothing except consume a
   // following quantifier
   void AddEmpty();
+  void AddCharacterClass(RegExpCharacterClass* cc);
   void AddAtom(RegExpTree* tree);
+  void AddTerm(RegExpTree* tree);
   void AddAssertion(RegExpTree* tree);
   void NewAlternative();  // '|'
   void AddQuantifierToAtom(int min, int max,
                            RegExpQuantifier::QuantifierType type);
   RegExpTree* ToRegExp();
 
  private:
+  static const uc16 kNoPendingSurrogate = 0;
+  void AddLeadSurrogate(uc16 lead_surrogate);
+  void AddTrailSurrogate(uc16 trail_surrogate);
+  void FlushPendingSurrogate();
   void FlushCharacters();
   void FlushText();
   void FlushTerms();
   Zone* zone() const { return zone_; }
+  bool unicode() const { return (flags_ & JSRegExp::kUnicode) != 0; }
 
   Zone* zone_;
   bool pending_empty_;
+  JSRegExp::Flags flags_;
   ZoneList<uc16>* characters_;
+  uc16 pending_surrogate_;
   BufferedZoneList<RegExpTree, 2> terms_;
   BufferedZoneList<RegExpTree, 2> text_;
   BufferedZoneList<RegExpTree, 2> alternatives_;
 #ifdef DEBUG
   enum { ADD_NONE, ADD_CHAR, ADD_TERM, ADD_ASSERT, ADD_ATOM } last_added_;
 #define LAST(x) last_added_ = x;
 #else
 #define LAST(x)
 #endif
 };
 
 
 class RegExpParser BASE_EMBEDDED {
  public:
-  RegExpParser(FlatStringReader* in, Handle<String>* error, bool multiline_mode,
-               bool unicode, Isolate* isolate, Zone* zone);
+  RegExpParser(FlatStringReader* in, Handle<String>* error,
+               JSRegExp::Flags flags, Isolate* isolate, Zone* zone);
 
   static bool ParseRegExp(Isolate* isolate, Zone* zone, FlatStringReader* input,
-                          bool multiline, bool unicode,
-                          RegExpCompileData* result);
+                          JSRegExp::Flags flags, RegExpCompileData* result);
 
   RegExpTree* ParsePattern();
   RegExpTree* ParseDisjunction();
   RegExpTree* ParseGroup();
   RegExpTree* ParseCharacterClass();
 
   // Parses a {...,...} quantifier and stores the range in the given
   // out parameters.
   bool ParseIntervalQuantifier(int* min_out, int* max_out);
 
@@ skipping 22 matching lines @@
   void Reset(int pos);
 
   // Reports whether the pattern might be used as a literal search string.
   // Only use if the result of the parse is a single atom node.
   bool simple();
   bool contains_anchor() { return contains_anchor_; }
   void set_contains_anchor() { contains_anchor_ = true; }
   int captures_started() { return captures_started_; }
   int position() { return next_pos_ - 1; }
   bool failed() { return failed_; }
+  bool unicode() const { return (flags_ & JSRegExp::kUnicode) != 0; }
+  bool multiline() const { return (flags_ & JSRegExp::kMultiline) != 0; }
 
   static bool IsSyntaxCharacter(uc32 c);
 
   static const int kMaxCaptures = 1 << 16;
   static const uc32 kEndMarker = (1 << 21);
 
  private:
   enum SubexpressionType {
     INITIAL,
     CAPTURE,  // All positive values represent captures.
     POSITIVE_LOOKAROUND,
     NEGATIVE_LOOKAROUND,
     GROUPING
   };
 
   class RegExpParserState : public ZoneObject {
    public:
     RegExpParserState(RegExpParserState* previous_state,
                       SubexpressionType group_type,
                       RegExpLookaround::Type lookaround_type,
-                      int disjunction_capture_index, Zone* zone)
+                      int disjunction_capture_index, JSRegExp::Flags flags,
+                      Zone* zone)
         : previous_state_(previous_state),
-          builder_(new (zone) RegExpBuilder(zone)),
+          builder_(new (zone) RegExpBuilder(zone, flags)),
           group_type_(group_type),
           lookaround_type_(lookaround_type),
           disjunction_capture_index_(disjunction_capture_index) {}
     // Parser state of containing expression, if any.
     RegExpParserState* previous_state() { return previous_state_; }
     bool IsSubexpression() { return previous_state_ != NULL; }
     // RegExpBuilder building this regexp's AST.
     RegExpBuilder* builder() { return builder_; }
     // Type of regexp being parsed (parenthesized group or entire regexp).
     SubexpressionType group_type() { return group_type_; }
@@ skipping 23 matching lines @@
   // Return the 1-indexed RegExpCapture object, allocate if necessary.
   RegExpCapture* GetCapture(int index);
 
   Isolate* isolate() { return isolate_; }
   Zone* zone() const { return zone_; }
 
   uc32 current() { return current_; }
   bool has_more() { return has_more_; }
   bool has_next() { return next_pos_ < in()->length(); }
   uc32 Next();
+  template <bool update_position>
+  uc32 ReadNext();
   FlatStringReader* in() { return in_; }
   void ScanForCaptures();
 
   Isolate* isolate_;
   Zone* zone_;
   Handle<String>* error_;
   ZoneList<RegExpCapture*>* captures_;
   FlatStringReader* in_;
   uc32 current_;
+  JSRegExp::Flags flags_;
   int next_pos_;
   int captures_started_;
   // The capture count is only valid after we have scanned for captures.
   int capture_count_;
   bool has_more_;
-  bool multiline_;
-  bool unicode_;
   bool simple_;
   bool contains_anchor_;
   bool is_scanned_for_captures_;
   bool failed_;
 };
 
 }  // namespace internal
 }  // namespace v8
 
 #endif  // V8_REGEXP_REGEXP_PARSER_H_