Make _Utf8Encoder create its buffer through an external function.

sdk/lib/convert/utf.dart

 // Copyright (c) 2013, 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.
 
 part of dart.convert;
 
 /** The Unicode Replacement character `U+FFFD` (�). */
 const int UNICODE_REPLACEMENT_CHARACTER_RUNE = 0xFFFD;
 
 /** The Unicode Byte Order Marker (BOM) character `U+FEFF`. */
@@ skipping 115 matching lines @@
 class _Utf8Encoder {
   int _carry = 0;
   int _bufferIndex = 0;
   final List<int> _buffer;
 
   static const _DEFAULT_BYTE_BUFFER_SIZE = 1024;
 
   _Utf8Encoder() : this.withBufferSize(_DEFAULT_BYTE_BUFFER_SIZE);
 
   _Utf8Encoder.withBufferSize(int bufferSize)
       // TODO(11971, floitsch): use Uint8List instead of normal lists.

[floitsch, 2014/01/31 13:03:49]

Remove TODO.

-      : _buffer = new List<int>(bufferSize);
+      : _buffer = _createBuffer(bufferSize);
+
+  /**
+   * Allow an implementation to pick the most efficient way of storing bytes.

[floitsch, 2014/01/31 13:03:49]

Maybe add a TODO(11971): just always use a Uint8List.

+   */
+  external static List<int> _createBuffer(int size);
 
   /**
    * Tries to combine the given [leadingSurrogate] with the [nextCodeUnit] and
    * writes it to [_buffer].
    *
    * Returns true if the [nextCodeUnit] was combined with the
    * [leadingSurrogate]. If it wasn't then nextCodeUnit was not a trailing
    * surrogate and has not been written yet.
    *
    * It is safe to pass 0 for [nextCodeUnit] in which case only the leading
@@ skipping 365 matching lines @@
       }
       break loop;
     }
     if (expectedUnits > 0) {
       _value = value;
       _expectedUnits = expectedUnits;
       _extraUnits = extraUnits;
     }
   }
 }