Ensure that scanner state is correctly reset when an error is encountered.

src/scanner.cc

 // Copyright 2006-2008 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 301 matching lines @@
       if (MatchKeywordStart(input, "with", 1, Token::WITH)) return;
       break;
     default:
       UNREACHABLE();
   }
   // On fallthrough, it's a failure.
   state_ = UNMATCHABLE;
 }
 
 
+
+// ----------------------------------------------------------------------------
+// Scanner::LiteralScope
+
+Scanner::LiteralScope::LiteralScope(Scanner* self)
+    : scanner_(self), complete_(false) {
+  self->StartLiteral();
+}
+
+
+Scanner::LiteralScope::~LiteralScope() {
+  if (!complete_) scanner_->DropLiteral();
+}
+
+
+void Scanner::LiteralScope::Complete() {
+  scanner_->TerminateLiteral();
+  complete_ = true;
+}
+
 // ----------------------------------------------------------------------------
 // Scanner
 
 Scanner::Scanner(ParserMode pre)
     : is_pre_parsing_(pre == PREPARSE), stack_overflow_(false) { }
 
 
 void Scanner::Initialize(Handle<String> source,
                          ParserLanguage language) {
   Init(source, NULL, 0, source->length(), language);
@@ skipping 47 matching lines @@
                                    start_position,
                                    end_position);
     source_ = &char_stream_buffer_;
   }
 
   is_parsing_json_ = (language == JSON);
 
   // Set c0_ (one character ahead)
   ASSERT(kCharacterLookaheadBufferSize == 1);
   Advance();
-  // Initialise current_ to not refer to a literal.
+  // Initialize current_ to not refer to a literal.
   current_.literal_chars = Vector<const char>();
+  // Reset literal buffer.
+  literal_buffer_.Reset();
 
   // Skip initial whitespace allowing HTML comment ends just like
   // after a newline and scan first token.
   has_line_terminator_before_next_ = true;
   SkipWhiteSpace();
   Scan();
 }
 
 
 Token::Value Scanner::Next() {
@@ skipping 15 matching lines @@
 
 void Scanner::StartLiteral() {
   literal_buffer_.StartLiteral();
 }
 
 
 void Scanner::AddChar(uc32 c) {
   literal_buffer_.AddChar(c);
 }
 
+
 void Scanner::TerminateLiteral() {
   next_.literal_chars = literal_buffer_.EndLiteral();
 }
 
 
+void Scanner::DropLiteral() {
+  literal_buffer_.DropLiteral();
+}
+
+
 void Scanner::AddCharAdvance() {
   AddChar(c0_);
   Advance();
 }
 
 
 static inline bool IsByteOrderMark(uc32 c) {
   // The Unicode value U+FFFE is guaranteed never to be assigned as a
   // Unicode character; this implies that in a Unicode context the
   // 0xFF, 0xFE byte pattern can only be interpreted as the U+FEFF
@@ skipping 188 matching lines @@
   } while (token == Token::WHITESPACE);
 
   next_.location.end_pos = source_pos();
   next_.token = token;
 }
 
 
 Token::Value Scanner::ScanJsonString() {
   ASSERT_EQ('"', c0_);
   Advance();
-  StartLiteral();
+  LiteralScope literal(this);
   while (c0_ != '"' && c0_ > 0) {
     // Check for control character (0x00-0x1f) or unterminated string (<0).
     if (c0_ < 0x20) return Token::ILLEGAL;
     if (c0_ != '\\') {
       AddCharAdvance();
     } else {
       Advance();
       switch (c0_) {
         case '"':
         case '\\':
@@ skipping 13 matching lines @@
           AddChar('\x0d');
           break;
         case 't':
           AddChar('\x09');
           break;
         case 'u': {
           uc32 value = 0;
           for (int i = 0; i < 4; i++) {
             Advance();
             int digit = HexValue(c0_);
-            if (digit < 0) return Token::ILLEGAL;
+            if (digit < 0) {
+              return Token::ILLEGAL;
+            }
             value = value * 16 + digit;
           }
           AddChar(value);
           break;
         }
         default:
           return Token::ILLEGAL;
       }
       Advance();
     }
   }
   if (c0_ != '"') {
     return Token::ILLEGAL;
   }
-  TerminateLiteral();
+  literal.Complete();
   Advance();
   return Token::STRING;
 }
 
 
 Token::Value Scanner::ScanJsonNumber() {
-  StartLiteral();
+  LiteralScope literal(this);
   if (c0_ == '-') AddCharAdvance();
   if (c0_ == '0') {
     AddCharAdvance();
     // Prefix zero is only allowed if it's the only digit before
     // a decimal point or exponent.
     if ('0' <= c0_ && c0_ <= '9') return Token::ILLEGAL;
   } else {
     if (c0_ < '1' || c0_ > '9') return Token::ILLEGAL;
     do {
       AddCharAdvance();
     } while (c0_ >= '0' && c0_ <= '9');
   }
   if (c0_ == '.') {
     AddCharAdvance();
     if (c0_ < '0' || c0_ > '9') return Token::ILLEGAL;
     do {
       AddCharAdvance();
     } while (c0_ >= '0' && c0_ <= '9');
   }
   if ((c0_ | 0x20) == 'e') {
     AddCharAdvance();
     if (c0_ == '-' || c0_ == '+') AddCharAdvance();
     if (c0_ < '0' || c0_ > '9') return Token::ILLEGAL;
     do {
       AddCharAdvance();
     } while (c0_ >= '0' && c0_ <= '9');
   }
-  TerminateLiteral();
+  literal.Complete();
   return Token::NUMBER;
 }
 
 
 Token::Value Scanner::ScanJsonIdentifier(const char* text,
                                          Token::Value token) {
-  StartLiteral();
+  LiteralScope literal(this);
   while (*text != '\0') {
     if (c0_ != *text) return Token::ILLEGAL;
     Advance();
     text++;
   }
   if (kIsIdentifierPart.get(c0_)) return Token::ILLEGAL;
-  TerminateLiteral();
+  literal.Complete();
   return token;
 }
 
 
 void Scanner::ScanJavaScript() {
   next_.literal_chars = Vector<const char>();
   Token::Value token;
   has_line_terminator_before_next_ = false;
   do {
     // Remember the position of the next token
@@ skipping 322 matching lines @@
   // should be illegal, but they are commonly handled
   // as non-escaped characters by JS VMs.
   AddChar(c);
 }
 
 
 Token::Value Scanner::ScanString() {
   uc32 quote = c0_;
   Advance();  // consume quote
 
-  StartLiteral();
+  LiteralScope literal(this);
   while (c0_ != quote && c0_ >= 0 && !kIsLineTerminator.get(c0_)) {
     uc32 c = c0_;
     Advance();
     if (c == '\\') {
       if (c0_ < 0) return Token::ILLEGAL;
       ScanEscape();
     } else {
       AddChar(c);
     }
   }
-  if (c0_ != quote) {
-    return Token::ILLEGAL;
-  }
-  TerminateLiteral();
+  if (c0_ != quote) return Token::ILLEGAL;
+  literal.Complete();
 
   Advance();  // consume quote
   return Token::STRING;
 }
 
 
 Token::Value Scanner::Select(Token::Value tok) {
   Advance();
   return tok;
 }
@@ skipping 15 matching lines @@
   while (IsDecimalDigit(c0_))
     AddCharAdvance();
 }
 
 
 Token::Value Scanner::ScanNumber(bool seen_period) {
   ASSERT(IsDecimalDigit(c0_));  // the first digit of the number or the fraction
 
   enum { DECIMAL, HEX, OCTAL } kind = DECIMAL;
 
-  StartLiteral();
+  LiteralScope literal(this);
   if (seen_period) {
     // we have already seen a decimal point of the float
     AddChar('.');
     ScanDecimalDigits();  // we know we have at least one digit
 
   } else {
     // if the first character is '0' we must check for octals and hex
     if (c0_ == '0') {
       AddCharAdvance();
 
       // either 0, 0exxx, 0Exxx, 0.xxx, an octal number, or a hex number
       if (c0_ == 'x' || c0_ == 'X') {
         // hex number
         kind = HEX;
         AddCharAdvance();
-        if (!IsHexDigit(c0_))
+        if (!IsHexDigit(c0_)) {
           // we must have at least one hex digit after 'x'/'X'
           return Token::ILLEGAL;
-        while (IsHexDigit(c0_))
+        }
+        while (IsHexDigit(c0_)) {
           AddCharAdvance();
-
+        }
       } else if ('0' <= c0_ && c0_ <= '7') {
         // (possible) octal number
         kind = OCTAL;
         while (true) {
           if (c0_ == '8' || c0_ == '9') {
             kind = DECIMAL;
             break;
           }
           if (c0_  < '0' || '7'  < c0_) break;
           AddCharAdvance();
@@ skipping 12 matching lines @@
   }
 
   // scan exponent, if any
   if (c0_ == 'e' || c0_ == 'E') {
     ASSERT(kind != HEX);  // 'e'/'E' must be scanned as part of the hex number
     if (kind == OCTAL) return Token::ILLEGAL;  // no exponent for octals allowed
     // scan exponent
     AddCharAdvance();
     if (c0_ == '+' || c0_ == '-')
       AddCharAdvance();
-    if (!IsDecimalDigit(c0_))
+    if (!IsDecimalDigit(c0_)) {
       // we must have at least one decimal digit after 'e'/'E'
       return Token::ILLEGAL;
+    }
     ScanDecimalDigits();
   }
-  TerminateLiteral();
 
   // The source character immediately following a numeric literal must
   // not be an identifier start or a decimal digit; see ECMA-262
   // section 7.8.3, page 17 (note that we read only one decimal digit
   // if the value is 0).
   if (IsDecimalDigit(c0_) || kIsIdentifierStart.get(c0_))
     return Token::ILLEGAL;
 
+  literal.Complete();
+
   return Token::NUMBER;
 }
 
 
 uc32 Scanner::ScanIdentifierUnicodeEscape() {
   Advance();
   if (c0_ != 'u') return unibrow::Utf8::kBadChar;
   Advance();
   uc32 c = ScanHexEscape('u', 4);
   // We do not allow a unicode escape sequence to start another
   // unicode escape sequence.
   if (c == '\\') return unibrow::Utf8::kBadChar;
   return c;
 }
 
 
 Token::Value Scanner::ScanIdentifier() {
   ASSERT(kIsIdentifierStart.get(c0_));
 
-  StartLiteral();
+  LiteralScope literal(this);
   KeywordMatcher keyword_match;
 
   // Scan identifier start character.
   if (c0_ == '\\') {
     uc32 c = ScanIdentifierUnicodeEscape();
     // Only allow legal identifier start characters.
     if (!kIsIdentifierStart.get(c)) return Token::ILLEGAL;
     AddChar(c);
     keyword_match.Fail();
   } else {
     AddChar(c0_);
     keyword_match.AddChar(c0_);
     Advance();
   }
 
   // Scan the rest of the identifier characters.
   while (kIsIdentifierPart.get(c0_)) {
     if (c0_ == '\\') {
       uc32 c = ScanIdentifierUnicodeEscape();
       // Only allow legal identifier part characters.
       if (!kIsIdentifierPart.get(c)) return Token::ILLEGAL;
       AddChar(c);
       keyword_match.Fail();
     } else {
       AddChar(c0_);
       keyword_match.AddChar(c0_);
       Advance();
     }
   }
-  TerminateLiteral();
+  literal.Complete();
 
   return keyword_match.token();
 }
 
 
 
 bool Scanner::IsIdentifier(unibrow::CharacterStream* buffer) {
   // Checks whether the buffer contains an identifier (no escape).
   if (!buffer->has_more()) return false;
   if (!kIsIdentifierStart.get(buffer->GetNext())) return false;
   while (buffer->has_more()) {
     if (!kIsIdentifierPart.get(buffer->GetNext())) return false;
   }
   return true;
 }
 
 
 bool Scanner::ScanRegExpPattern(bool seen_equal) {
   // Scan: ('/' | '/=') RegularExpressionBody '/' RegularExpressionFlags
   bool in_character_class = false;
 
   // Previous token is either '/' or '/=', in the second case, the
   // pattern starts at =.
   next_.location.beg_pos = source_pos() - (seen_equal ? 2 : 1);
   next_.location.end_pos = source_pos() - (seen_equal ? 1 : 0);
 
   // Scan regular expression body: According to ECMA-262, 3rd, 7.8.5,
   // the scanner should pass uninterpreted bodies to the RegExp
   // constructor.
-  StartLiteral();
+  LiteralScope literal(this);
   if (seen_equal)
     AddChar('=');
 
   while (c0_ != '/' || in_character_class) {
-    if (kIsLineTerminator.get(c0_) || c0_ < 0)
-      return false;
+    if (kIsLineTerminator.get(c0_) || c0_ < 0) return false;
     if (c0_ == '\\') {  // escaped character
       AddCharAdvance();
-      if (kIsLineTerminator.get(c0_) || c0_ < 0)
-        return false;
+      if (kIsLineTerminator.get(c0_) || c0_ < 0) return false;
       AddCharAdvance();
     } else {  // unescaped character
-      if (c0_ == '[')
-        in_character_class = true;
-      if (c0_ == ']')
-        in_character_class = false;
+      if (c0_ == '[') in_character_class = true;
+      if (c0_ == ']') in_character_class = false;
       AddCharAdvance();
     }
   }
   Advance();  // consume '/'
 
-  TerminateLiteral();
+  literal.Complete();
 
   return true;
 }
 
 bool Scanner::ScanRegExpFlags() {
   // Scan regular expression flags.
-  StartLiteral();
+  LiteralScope literal(this);
   while (kIsIdentifierPart.get(c0_)) {
     if (c0_ == '\\') {
       uc32 c = ScanIdentifierUnicodeEscape();
       if (c != static_cast<uc32>(unibrow::Utf8::kBadChar)) {
         // We allow any escaped character, unlike the restriction on
         // IdentifierPart when it is used to build an IdentifierName.
         AddChar(c);
         continue;
       }
     }
     AddCharAdvance();
   }
-  TerminateLiteral();
+  literal.Complete();
 
   next_.location.end_pos = source_pos() - 1;
   return true;
 }
 
 } }  // namespace v8::internal