Regexp: Refactor ChoiceNode::Emit

src/jsregexp.h

 // Copyright 2012 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_JSREGEXP_H_
 #define V8_JSREGEXP_H_
 
 #include "src/allocation.h"
 #include "src/assembler.h"
 #include "src/zone-inl.h"
@@ skipping 427 matching lines @@
   TextElement(TextType text_type, RegExpTree* tree)
       : cp_offset_(-1), text_type_(text_type), tree_(tree) {}
 
   int cp_offset_;
   TextType text_type_;
   RegExpTree* tree_;
 };
 
 
 class Trace;
-
+struct PreloadState;
+class GreedyLoopState;
+class AlternativeGenerationList;
 
 struct NodeInfo {
   NodeInfo()
       : being_analyzed(false),
         been_analyzed(false),
         follows_word_interest(false),
         follows_newline_interest(false),
         follows_start_interest(false),
         at_end(false),
         visited(false),
@@ skipping 610 matching lines @@
                                     bool not_at_start);
   virtual void FillInBMInfo(int offset,
                             int budget,
                             BoyerMooreLookahead* bm,
                             bool not_at_start);
 
   bool being_calculated() { return being_calculated_; }
   bool not_at_start() { return not_at_start_; }
   void set_not_at_start() { not_at_start_ = true; }
   void set_being_calculated(bool b) { being_calculated_ = b; }
-  virtual bool try_to_emit_quick_check_for_alternative(int i) { return true; }
+  virtual bool try_to_emit_quick_check_for_alternative(bool is_first) {
+    return true;
+  }
   virtual RegExpNode* FilterASCII(int depth, bool ignore_case);
 
  protected:
   int GreedyLoopTextLengthForAlternative(GuardedAlternative* alternative);
   ZoneList<GuardedAlternative>* alternatives_;
 
  private:
   friend class DispatchTableConstructor;
   friend class Analysis;
   void GenerateGuard(RegExpMacroAssembler* macro_assembler,
                      Guard* guard,
                      Trace* trace);
   int CalculatePreloadCharacters(RegExpCompiler* compiler, int eats_at_least);
   void EmitOutOfLineContinuation(RegExpCompiler* compiler,
                                  Trace* trace,
                                  GuardedAlternative alternative,
                                  AlternativeGeneration* alt_gen,
                                  int preload_characters,
                                  bool next_expects_preload);
+  void SetUpPreLoad(RegExpCompiler* compiler,
+                    Trace* current_trace,
+                    PreloadState* preloads);
+  void AssertGuardsMentionRegisters(Trace* trace);
+  int EmitOptimizedUnanchoredSearch(RegExpCompiler* compiler, Trace* trace);
+  Trace* EmitGreedyLoop(RegExpCompiler* compiler,
+                        Trace* trace,
+                        AlternativeGenerationList* alt_gens,
+                        PreloadState* preloads,
+                        GreedyLoopState* greedy_loop_state,
+                        int text_length);
+  void EmitChoices(RegExpCompiler* compiler,
+                   AlternativeGenerationList* alt_gens,
+                   int first_choice,
+                   Trace* trace,
+                   bool is_greedy_loop,
+                   PreloadState* preloads);
   DispatchTable* table_;
   // If true, this node is never checked at the start of the input.
   // Allows a new trace to start with at_start() set to false.
   bool not_at_start_;
   bool being_calculated_;
 };
 
 
 class NegativeLookaheadChoiceNode: public ChoiceNode {
  public:
@@ skipping 15 matching lines @@
                             bool not_at_start) {
     alternatives_->at(1).node()->FillInBMInfo(
         offset, budget - 1, bm, not_at_start);
     if (offset == 0) set_bm_info(not_at_start, bm);
   }
   // For a negative lookahead we don't emit the quick check for the
   // alternative that is expected to fail.  This is because quick check code
   // starts by loading enough characters for the alternative that takes fewest
   // characters, but on a negative lookahead the negative branch did not take
   // part in that calculation (EatsAtLeast) so the assumptions don't hold.
-  virtual bool try_to_emit_quick_check_for_alternative(int i) { return i != 0; }
+  virtual bool try_to_emit_quick_check_for_alternative(bool is_first) {
+    return !is_first;
+  }
   virtual RegExpNode* FilterASCII(int depth, bool ignore_case);
 };
 
 
 class LoopChoiceNode: public ChoiceNode {
  public:
   explicit LoopChoiceNode(bool body_can_be_zero_length, Zone* zone)
       : ChoiceNode(2, zone),
         loop_node_(NULL),
         continue_node_(NULL),
-        body_can_be_zero_length_(body_can_be_zero_length) { }
+        body_can_be_zero_length_(body_can_be_zero_length)
+        { }
   void AddLoopAlternative(GuardedAlternative alt);
   void AddContinueAlternative(GuardedAlternative alt);
   virtual void Emit(RegExpCompiler* compiler, Trace* trace);
   virtual int EatsAtLeast(int still_to_find,  int budget, bool not_at_start);
   virtual void GetQuickCheckDetails(QuickCheckDetails* details,
                                     RegExpCompiler* compiler,
                                     int characters_filled_in,
                                     bool not_at_start);
   virtual void FillInBMInfo(int offset,
                             int budget,
@@ skipping 130 matching lines @@
     }
   }
 
   void SetAll(int map_number) {
     bitmaps_->at(map_number)->SetAll();
   }
 
   void SetRest(int from_map) {
     for (int i = from_map; i < length_; i++) SetAll(i);
   }
-  bool EmitSkipInstructions(RegExpMacroAssembler* masm);
+  void EmitSkipInstructions(RegExpMacroAssembler* masm);
 
  private:
   // This is the value obtained by EatsAtLeast.  If we do not have at least this
   // many characters left in the sample string then the match is bound to fail.
   // Therefore it is OK to read a character this far ahead of the current match
   // point.
   int length_;
   RegExpCompiler* compiler_;
   // 0x7f for ASCII, 0xffff for UTF-16.
   int max_char_;
@@ skipping 172 matching lines @@
   RegExpNode* stop_node_;
   Label* loop_label_;
   int characters_preloaded_;
   int bound_checked_up_to_;
   QuickCheckDetails quick_check_performed_;
   int flush_budget_;
   TriBool at_start_;
 };
 
 
+class GreedyLoopState {
+ public:
+  explicit GreedyLoopState(bool not_at_start);
+
+  Label* label() { return &label_; }
+  Trace* counter_backtrack_trace() { return &counter_backtrack_trace_; }
+
+ private:
+  Label label_;
+  Trace counter_backtrack_trace_;
+};
+
+
+struct PreloadState {
+  static const int kEatsAtLeastNotYetInitialized = -1;
+  bool preload_is_current_;
+  bool preload_has_checked_bounds_;
+  int preload_characters_;
+  int eats_at_least_;
+  void init() {
+    eats_at_least_ = kEatsAtLeastNotYetInitialized;
+  }
+};
+
+
 class NodeVisitor {
  public:
   virtual ~NodeVisitor() { }
 #define DECLARE_VISIT(Type)                                          \
   virtual void Visit##Type(Type##Node* that) = 0;
 FOR_EACH_NODE_TYPE(DECLARE_VISIT)
 #undef DECLARE_VISIT
   virtual void VisitLoopChoice(LoopChoiceNode* that) { VisitChoice(that); }
 };
 
@@ skipping 117 matching lines @@
                                    Handle<String> sample_subject,
                                    bool is_ascii, Zone* zone);
 
   static void DotPrint(const char* label, RegExpNode* node, bool ignore_case);
 };
 
 
 } }  // namespace v8::internal
 
 #endif  // V8_JSREGEXP_H_