Attached JISTest.java to reproduce NIO problem.

On my environment Java(TM) 2 Runtime Environment, Standard Edition (build

1.5.0-b64), i got the following result.

 

Original:

0000: 1B 24 42 46 7C 4B 5C 38 6C 1B 28 42

 

Decoded and Encoded by NIO:

0000: 1B 24 42 46 7C 4B 5C 38 6C 00 00 00

 

This result means that NIO iso-2022-jp converter does not add

the escape sequence '1B 28 42' to the end of japanese characters

'46 7C 4B 5C 38 6C', just fills '00 00 00'.

According to iso-2022-jp specification(rfc1468), the escape

sequence '1B 28 42', i.e. 'ESC ( B' is used to switch back to

ASCII mode. and so, '00 00 00' is still handled japanese characters,

then is interpreted broken japanese characters.
###@###.### 2005-1-25 01:51:29 GMT

Here is a test case
=================================JISTest.java=========
import java.nio.*;
import java.nio.charset.*;
import sun.misc.HexDumpEncoder;

public class JISTest {

  public static void main(String args[]) {

    HexDumpEncoder dumper = new HexDumpEncoder();

    String jisx0208 = "\u65e5\u672c\u8a9e";

    try{
      byte[] b = jisx0208.getBytes("ISO-2022-JP");
      System.out.println("Original: ");
      System.out.println(dumper.encode(b));

      System.out.println("");

      System.out.println("Decoded and Encoded by NIO: ");
      Charset iso2022jp = Charset.forName("ISO-2022-JP");
      CharsetDecoder decoder = iso2022jp.newDecoder();
      ByteBuffer jpBytes = ByteBuffer.wrap(b);
      CharBuffer jpChars = decoder.decode(jpBytes);
      CharsetEncoder encoder = iso2022jp.newEncoder();
      ByteBuffer jpBytes2 = encoder.encode(jpChars);
      System.out.println(dumper.encode(jpBytes2.array()));

    } catch (Exception e) {
      e.printStackTrace();
    }
  }
}

###@###.### 2005-1-26 00:47:21 GMT