Handle odd data lengths in UTF-16 decoder

Source/modules/encoding/TextDecoder.cpp

 /*
  * Copyright (C) 2013 Google Inc. 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
@@ skipping 69 matching lines @@
 String TextDecoder::decode(ArrayBufferView* input, const Dictionary& options, ExceptionState& es)
 {
     bool stream = false;
     options.get("stream", stream);
 
     const char* start = input ? static_cast<const char*>(input->baseAddress()) : 0;
     size_t length = input ? input->byteLength() : 0;
 
     bool flush = !stream;
 
-    // FIXME: Not all TextCodec implementations handle |flush| - notably TextCodecUTF16
-    // ignores it and never flushes!
-
     bool sawError = false;
     String s = m_codec->decode(start, length, flush, m_fatal, sawError);
 
     if (m_fatal && sawError) {
         es.throwDOMException(EncodingError, "The encoded data was not valid.");
         return String();
     }
 
     if (!m_bomSeen && !s.isEmpty()) {
         m_bomSeen = true;
         String name(m_encoding.name());
         if ((name == "UTF-8" || name == "UTF-16LE" || name == "UTF-16BE") && s[0] == 0xFEFF)
             s.remove(0);
     }
 
     if (flush)
         m_bomSeen = false;
 
     return s;
 }
 
 } // namespace WebCore