Inserted string validation as separate task in compiler.

frog/leg/string_validator.dart

+// Copyright (c) 2012, the Dart project authors.  Please see the AUTHORS file
+// for details. All rights reserved. Use of this source code is governed by a
+// BSD-style license that can be found in the LICENSE file.
+
+// Checks a tree structure for whether the string literals are valid.
+
+class StringValidatorTask extends CompilerTask {
+  // Validator has no state except the compiler, so we just create one
+  // per task.
+  final StringValidator validator;
+
+  StringValidatorTask(Compiler compiler)
+      : super(compiler),
+        validator = new StringValidator(compiler);
+
+  void validate(Node tree) {
+    tree.accept(validator);
+  }
+}
+
+class StringValidator extends AbstractVisitor {
+  final Compiler compiler;
+
+  StringValidator(this.compiler);
+
+  void visitNode(Node node) {
+    node.visitChildren(this);
+  }
+
+  visitLiteralString(LiteralString node) {
+    // This is a string literal where the source contains both start and
+    // end quotes.
+    // String literals that are part of string interpolations are
+    // handled there to account for their varying degree of quotedness.
+    SourceString source = node.value;
+    StringQuoting quoting = quotingFromString(source);
+    int leftQuote = quoting.leftQuoteLength;
+    int rightQuote = quoting.rightQuoteLength;
+    SourceString content = source.copyWithoutQuotes(leftQuote, rightQuote);
+    validateString(node, node.token.charOffset + leftQuote, content, quoting);
+  }
+
+  // String interpolation validates the immediate strings and traverses
+  // the sub-expressions.
+  visitStringInterpolation(StringInterpolation node) {
+    LiteralString literalString = node.string;
+    SourceString source = literalString.value;
+    StringQuoting quoting = quotingFromString(source);
+    int offset = literalString.token.charOffset + quoting.leftQuoteLength;
+    SourceString string = source.copyWithoutQuotes(quoting.leftQuoteLength, 0);
+    for (StringInterpolationPart part in node.parts) {
+      validateString(literalString, offset, string, quoting);
+      part.expression.accept(this);
+      literalString = part.string;
+      offset = literalString.token.charOffset;
+      string = literalString.value;
+    }
+    string = string.copyWithoutQuotes(0, quoting.rightQuoteLength);
+    validateString(literalString, offset, string, quoting);
+  }
+
+  static StringQuoting quotingFromString(SourceString sourceString) {
+    Iterator<int> source = sourceString.iterator();
+    bool raw = false;
+    int quoteChar = source.next();
+    if (quoteChar == $AT) {
+      raw = true;
+      quoteChar = source.next();
+    }
+    assert(quoteChar === $SQ || quoteChar === $DQ);
+    // String has one quote. Check it if has three.

[karlklose, 2012/01/25 08:47:26]

one -> at least one

[Lasse Reichstein Nielsen, 2012/01/26 10:14:20]

Done.

+    // If it only have two, the string must be an empty string literal,
+    // and end after the second quote.
+    bool multiline = false;
+    if (source.hasNext() && source.next() == quoteChar && source.hasNext()) {
+      assert(source.next() == quoteChar);
+      multiline = true;
+    }
+    return StringQuoting.get(quoteChar, raw, multiline);
+  }
+
+  void stringParseError(String message, LiteralString node, int offset) {
+    assert(node.quotedString === null);
+    compiler.cancel("$message @ $offset", node);
+  }
+
+  /**
+   * Validates the escape sequences and special characters of a string literal.
+   * Returns the number of actual scalar values in the corresponding Dart

[karlklose, 2012/01/25 08:47:26]

It returns whether the string is valid or not, doesn't it?

[Lasse Reichstein Nielsen, 2012/01/26 10:14:20]

True. Old comment is old.

+   * string, or a negative value if the string is invalid.
+   */
+  bool validateString(LiteralString node,
+                      int startOffset,
+                      SourceString string,
+                      StringQuoting quoting) {
+    // We only need to check for invalid x and u escapes, for line
+    // terminators in non-multiline strings, and for invalid Unicode
+    // scalar values (either directly or as u-escape values).
+    int length = 0;
+    int index = startOffset;
+    for(Iterator<int> iter = string.iterator(); iter.hasNext(); length++) {
+      index++;
+      int code = iter.next();
+      if (code === $BACKSLASH) {
+        if (quoting.raw) continue;
+        if (!iter.hasNext()) {
+          stringParseError("Incomplete escape sequence",  node, index);
+          return false;
+        }
+        index++;
+        code = iter.next();
+        if (code === $x) {
+          for (int i = 0; i < 2; i++) {
+            if (!iter.hasNext()) {
+              stringParseError("Incomplete escape sequence",  node, index);
+              return false;
+            }
+            index++;
+            code = iter.next();
+            if (!isHexDigit(code)) {
+              stringParseError("Invalid character in escape sequence",
+                               node, index);
+              return false;
+            }
+          }
+          continue;
+        } else if (code === $u) {
+          int escapeStart = index - 1;
+          index++;
+          code = iter.next();
+          int value = 0;
+          if (code == $OPEN_CURLY_BRACKET) {
+            // expect 1-7 hex digits.
+            int count = 0;
+            index++;
+            code = iter.next();
+            do {
+              if (!isHexDigit(code)) {
+                stringParseError("Invalid character in escape sequence",
+                                 node, index);
+                return false;
+              }
+              count++;
+              value = value * 16 + hexDigitValue(code);
+              index++;
+              code = iter.next();
+            } while (code != $CLOSE_CURLY_BRACKET);
+            if (count > 7) {
+              stringParseError("Invalid character in escape sequence",
+                               node, index - (count - 7));
+              return false;
+            }
+          } else {
+            // Expect four hex digits, including the one just tread.
+            for (int i = 0; i < 4; i++) {
+              if (i > 0) {
+                index++;
+                code = iter.next();
+              }
+              if (!isHexDigit(code)) {
+                stringParseError("Invalid character in escape sequence",
+                                 node, index);
+                return false;
+              }
+              value = value * 16 + hexDigitValue(code);
+            }
+          }
+          if (0xd800 <= value && ( value <= 0xdfff || value > 0x10ffff)) {
+            stringParseError(
+                "Invalid unicode scalar value U+${value.toRadixString(16)}",
+                node, index);
+            return false;
+          }
+          continue;
+        }
+      }
+      // This handles borth unescaped characters as well as those
+      // characters after a backslash that doesn't have a special
+      // meaning.
+      if (code >= 0xd800 && (code <= 0xdfff || code > 0x10ffff)) {
+        stringParseError(
+            "Invalid unicode scalar value U+${code.toRadixString(16)}",
+            node, index);
+        return false;
+      }
+      if (!quoting.multiline && (code === $LF || code === $CR)) {
+        stringParseError("Line  terminator in single-line string",
+                         node, index);
+        return false;
+      }
+    }
+    // String literal successfully validated.
+    node.quotedString = new QuotedString(string, quoting, length);
+    return true;
+  }
+}