[regexp] quantifier refers to the surrogate pair in unicode regexp.

src/regexp/regexp-parser.cc

Index: src/regexp/regexp-parser.cc
diff --git a/src/regexp/regexp-parser.cc b/src/regexp/regexp-parser.cc
index 5cdad974b3886254a50073798a24df09754182b5..cb6073ee4a0bec92e908d98118599ee02a996d31 100644
--- a/src/regexp/regexp-parser.cc
+++ b/src/regexp/regexp-parser.cc
@@ -56,6 +56,16 @@ void RegExpParser::Advance() {
     } else {
       current_ = in()->Get(next_pos_);
       next_pos_++;
+      // Read the whole surrogate pair in case of unicode flag, if possible.
+      if (unicode_ && next_pos_ < in()->length() &&
+          unibrow::Utf16::IsLeadSurrogate(static_cast<uc16>(current_))) {
+        uc16 trail = in()->Get(next_pos_);
+        if (unibrow::Utf16::IsTrailSurrogate(trail)) {
+          current_ = unibrow::Utf16::CombineSurrogatePair(
+              static_cast<uc16>(current_), trail);
+          next_pos_++;
+        }
+      }
     }
   } else {
     current_ = kEndMarker;
@@ -417,12 +427,7 @@ RegExpTree* RegExpParser::ParseDisjunction() {
             Advance(2);
             uc32 value;
             if (ParseUnicodeEscape(&value)) {
-              if (value > unibrow::Utf16::kMaxNonSurrogateCharCode) {
-                builder->AddCharacter(unibrow::Utf16::LeadSurrogate(value));
-                builder->AddCharacter(unibrow::Utf16::TrailSurrogate(value));
-              } else {
-                builder->AddCharacter(static_cast<uc16>(value));
-              }
+              builder->AddUnicodeCharacter(value);
             } else if (!FLAG_harmony_unicode_regexps || !unicode_) {
               builder->AddCharacter('u');
             } else {
@@ -457,7 +462,11 @@ RegExpTree* RegExpParser::ParseDisjunction() {
         // fallthrough
       }
       default:
-        builder->AddCharacter(current());
+        if (unicode_) {

[rossberg, 2016/01/11 12:25:41]

Nit: is this if necessary? Can't you always use AddUnicode?

[Yang, 2016/01/11 14:41:44]

Good point. Done.

+          builder->AddUnicodeCharacter(current());
+        } else {
+          builder->AddCharacter(current());
+        }
         Advance();
         break;
     }  // end switch(current())
@@ -1059,6 +1068,19 @@ void RegExpBuilder::AddCharacter(uc16 c) {
 }
 
 
+void RegExpBuilder::AddUnicodeCharacter(uc32 c) {
+  if (c > unibrow::Utf16::kMaxNonSurrogateCharCode) {
+    ZoneList<uc16> surrogate_pair(2, zone());
+    surrogate_pair.Add(unibrow::Utf16::LeadSurrogate(c), zone());
+    surrogate_pair.Add(unibrow::Utf16::TrailSurrogate(c), zone());
+    RegExpAtom* atom = new (zone()) RegExpAtom(surrogate_pair.ToConstVector());
+    AddAtom(atom);
+  } else {
+    AddCharacter(static_cast<uc16>(c));
+  }
+}
+
+
 void RegExpBuilder::AddEmpty() { pending_empty_ = true; }