Implement ES6 unicode escapes.

src/scanner.cc

Index: src/scanner.cc
diff --git a/src/scanner.cc b/src/scanner.cc
index ddcd937584d50fc45fa921a25e2a5d21fcf5697e..7289c7a7d08b233a9669fa4c729123b1a1a2fe58 100644
--- a/src/scanner.cc
+++ b/src/scanner.cc
@@ -71,6 +71,22 @@ uc32 Scanner::ScanHexNumber(int expected_length) {
 }
 
 
+uc32 Scanner::ScanUnlimitedLengthHexNumber(int max_value) {
+  uc32 x = 0;
+  int d = HexValue(c0_);
+  if (d < 0) {
+    return -1;
+  }
+  while (d >= 0) {
+    x = x * 16 + d;
+    if (x > max_value) return -1;
+    Advance();
+    d = HexValue(c0_);
+  }
+  return x;
+}
+
+
 // Ensure that tokens can be stored in a byte.
 STATIC_ASSERT(Token::NUM_TOKENS <= 0x100);
 
@@ -673,6 +689,7 @@ void Scanner::SeekForward(int pos) {
 
 bool Scanner::ScanEscape() {
   uc32 c = c0_;
+  uc32 c2 = -1;
   Advance();
 
   // Skip escaped newlines.
@@ -694,7 +711,7 @@ bool Scanner::ScanEscape() {
     case 'r' : c = '\r'; break;
     case 't' : c = '\t'; break;
     case 'u' : {
-      c = ScanHexNumber(4);
+      c = ScanUnicodeEscape();
       if (c < 0) return false;
       break;
     }
@@ -718,6 +735,9 @@ bool Scanner::ScanEscape() {
   // above cases should be illegal, but they are commonly handled as
   // non-escaped characters by JS VMs.
   AddLiteralChar(c);
+  if (c2 >= 0) {
+    AddLiteralChar(c2);
+  }
   return true;
 }
 
@@ -885,8 +905,27 @@ Token::Value Scanner::ScanNumber(bool seen_period) {
 
 uc32 Scanner::ScanIdentifierUnicodeEscape() {
   Advance();
-  if (c0_ != 'u') return -1;
+  if (c0_ != 'u') return false;

[caitp (gmail), 2014/11/13 03:21:42]

Will cause `var \x` to look like `var \u{0}x`, if I'm not mistaken

[marja, 2014/11/13 10:06:23]

Oh, yes, this was a pure editing error.

   Advance();
+  return ScanUnicodeEscape();
+}
+
+
+uc32 Scanner::ScanUnicodeEscape() {
+  // Accept both \uxxxx and \u{xxxxxx}. In the latter case, the number of hex
+  // digits between { } is arbitrary. \ and u have already been read.
+  if (c0_ == '{') {
+    Advance();
+    uc32 cp = ScanUnlimitedLengthHexNumber(0x10ffff);
+    if (cp < 0) {
+      return -1;
+    }
+    if (c0_ != '}') {
+      return -1;
+    }
+    Advance();
+    return cp;
+  }
   return ScanHexNumber(4);
 }