[regexp] back refs must not start/end in the middle of a surrogate pair

src/regexp/regexp-macro-assembler.h

Index: src/regexp/regexp-macro-assembler.h
diff --git a/src/regexp/regexp-macro-assembler.h b/src/regexp/regexp-macro-assembler.h
index 20599334cd3f7450e9991e188a29bab2b1fe0f55..a80d961b1bd374e20e3527d4bc2294803ec36eb6 100644
--- a/src/regexp/regexp-macro-assembler.h
+++ b/src/regexp/regexp-macro-assembler.h
@@ -11,6 +11,13 @@
 namespace v8 {
 namespace internal {
 
+static const uc32 kLeadSurrogateStart = 0xd800;
+static const uc32 kLeadSurrogateEnd = 0xdbff;
+static const uc32 kTrailSurrogateStart = 0xdc00;
+static const uc32 kTrailSurrogateEnd = 0xdfff;
+static const uc32 kNonBmpStart = 0x10000;
+static const uc32 kNonBmpEnd = 0x10ffff;
+
 struct DisjunctDecisionRow {
   RegExpCharacterClass cc;
   Label* on_match;
@@ -146,6 +153,9 @@ class RegExpMacroAssembler {
   virtual void ClearRegisters(int reg_from, int reg_to) = 0;
   virtual void WriteStackPointerToRegister(int reg) = 0;
 
+  // Check that we are not in the middle of a surrogate pair.
+  void CheckNotInSurrogatePair(int cp_offset, Label* on_failure);
+
   // Controls the generation of large inlined constants in the code.
   void set_slow_safe(bool ssc) { slow_safe_compiler_ = ssc; }
   bool slow_safe() { return slow_safe_compiler_; }