Allow surrogates in string literals.

pkg/analyzer/lib/src/generated/parser.dart

 // Copyright (c) 2014, 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.
 
 library analyzer.src.generated.parser;
 
 import 'dart:collection';
 import "dart:math" as math;
 
 import 'package:analyzer/dart/ast/ast.dart';
@@ skipping 78 matching lines @@
       0, (Parser target) => target.parseTypeArgumentList()),
   'parseTypeName_0':
       new MethodTrampoline(0, (Parser target) => target.parseTypeName()),
   'parseTypeParameter_0':
       new MethodTrampoline(0, (Parser target) => target.parseTypeParameter()),
   'parseTypeParameterList_0': new MethodTrampoline(
       0, (Parser target) => target.parseTypeParameterList()),
   'parseWithClause_0':
       new MethodTrampoline(0, (Parser target) => target.parseWithClause()),
   'advance_0': new MethodTrampoline(0, (Parser target) => target._advance()),
-  'appendScalarValue_5': new MethodTrampoline(
+  'appendCodePoint_5': new MethodTrampoline(
       5,
       (Parser target, arg0, arg1, arg2, arg3, arg4) =>
-          target._appendScalarValue(arg0, arg1, arg2, arg3, arg4)),
+          target._appendCodePoint(arg0, arg1, arg2, arg3, arg4)),
   'computeStringValue_3': new MethodTrampoline(
       3,
       (Parser target, arg0, arg1, arg2) =>
           target._computeStringValue(arg0, arg1, arg2)),
   'convertToFunctionDeclaration_1': new MethodTrampoline(
       1, (Parser target, arg0) => target._convertToFunctionDeclaration(arg0)),
   'couldBeStartOfCompilationUnitMember_0': new MethodTrampoline(
       0, (Parser target) => target._couldBeStartOfCompilationUnitMember()),
   'createSyntheticIdentifier_0': new MethodTrampoline(
       0, (Parser target) => target._createSyntheticIdentifier()),
@@ skipping 1956 matching lines @@
   }
 
   /**
    * Advance to the next token in the token stream.
    */
   void _advance() {
     _currentToken = _currentToken.next;
   }
 
   /**
-   * Append the character equivalent of the given [scalarValue] to the given
+   * Append the character equivalent of the given [codePoint] to the given
    * [builder]. Use the [startIndex] and [endIndex] to report an error, and
-   * don't append anything to the builder, if the scalar value is invalid. The
+   * don't append anything to the builder, if the code point is invalid. The
    * [escapeSequence] is the escape sequence that was parsed to produce the
-   * scalar value (used for error reporting).
+   * code point (used for error reporting).
    */
-  void _appendScalarValue(StringBuffer buffer, String escapeSequence,
-      int scalarValue, int startIndex, int endIndex) {
-    if (scalarValue < 0 ||
-        scalarValue > Character.MAX_CODE_POINT ||
-        (scalarValue >= 0xD800 && scalarValue <= 0xDFFF)) {
+  void _appendCodePoint(StringBuffer buffer, String source,
+      int codePoint, int startIndex, int endIndex) {
+    if (codePoint < 0 || codePoint > Character.MAX_CODE_POINT) {
+      String escapeSequence = source.substring(startIndex, endIndex + 1);
       _reportErrorForCurrentToken(
           ParserErrorCode.INVALID_CODE_POINT, [escapeSequence]);
       return;
     }
-    if (scalarValue < Character.MAX_VALUE) {
-      buffer.writeCharCode(scalarValue);
+    if (codePoint < Character.MAX_VALUE) {
+      buffer.writeCharCode(codePoint);
     } else {
-      buffer.write(Character.toChars(scalarValue));
+      buffer.write(Character.toChars(codePoint));
     }
   }
 
   /**
    * Return the content of a string with the given literal representation. The
    * [lexeme] is the literal representation of the string. The flag [isFirst] is
    * `true` if this is the first token in a string literal. The flag [isLast] is
    * `true` if this is the last token in a string literal.
    */
   String _computeStringValue(String lexeme, bool isFirst, bool isLast) {
@@ skipping 5978 matching lines @@
             // Illegal escape sequence: incomplete escape
             _reportErrorForCurrentToken(ParserErrorCode.INVALID_UNICODE_ESCAPE);
             return length;
           }
           currentChar = lexeme.codeUnitAt(currentIndex);
         }
         if (digitCount < 1 || digitCount > 6) {
           // Illegal escape sequence: not enough or too many hex digits
           _reportErrorForCurrentToken(ParserErrorCode.INVALID_UNICODE_ESCAPE);
         }
-        _appendScalarValue(buffer, lexeme.substring(index, currentIndex + 1),
-            value, index, currentIndex);
+        _appendCodePoint(buffer, lexeme, value, index, currentIndex);

[Lasse Reichstein Nielsen, 2016/09/15 11:18:44]

There was no need to create a new substring when you are also passing the index
and currentIndex anyway. The substring is only used for reporting errors, so now
it's only created if there is an error.

[floitsch, 2016/09/15 15:09:19]

Acknowledged.

         return currentIndex + 1;
       } else {
         if (currentIndex + 3 >= length) {
           // Illegal escape sequence: not enough hex digits
           _reportErrorForCurrentToken(ParserErrorCode.INVALID_UNICODE_ESCAPE);
           return length;
         }
         int firstDigit = currentChar;
         int secondDigit = lexeme.codeUnitAt(currentIndex + 1);
         int thirdDigit = lexeme.codeUnitAt(currentIndex + 2);
         int fourthDigit = lexeme.codeUnitAt(currentIndex + 3);
         if (!_isHexDigit(firstDigit) ||
             !_isHexDigit(secondDigit) ||
             !_isHexDigit(thirdDigit) ||
             !_isHexDigit(fourthDigit)) {
           // Illegal escape sequence: invalid hex digits
           _reportErrorForCurrentToken(ParserErrorCode.INVALID_UNICODE_ESCAPE);
         } else {
-          _appendScalarValue(
+          _appendCodePoint(

[Lasse Reichstein Nielsen, 2016/09/15 11:18:44]

Retained the call to _appendCodePoint here, but it will always do
buffer.write(value).

[floitsch, 2016/09/15 15:09:19]

Acknowledged.

               buffer,
-              lexeme.substring(index, currentIndex + 1),

[Lasse Reichstein Nielsen, 2016/09/15 11:18:43]

+1 here should have been +4.

+              lexeme,
               (((((Character.digit(firstDigit, 16) << 4) +
                                   Character.digit(secondDigit, 16)) <<
                               4) +
                           Character.digit(thirdDigit, 16)) <<
                       4) +
                   Character.digit(fourthDigit, 16),
               index,
               currentIndex + 3);
         }
         return currentIndex + 4;
@@ skipping 953 matching lines @@
    */
   const ParserErrorCode(String name, String message, [String correction])
       : super(name, message, correction);
 
   @override
   ErrorSeverity get errorSeverity => ErrorSeverity.ERROR;
 
   @override
   ErrorType get type => ErrorType.SYNTACTIC_ERROR;
 }