To execute code, use this command:
"java -Dfile.encoding=iso-8859-2 ErrorInEncoding":
----------------------------------------------------------------
import java.io.*;
import java.util.*;

public class ErrorInEncoding {
 static public void main(String[] args)
    throws UnsupportedEncodingException,IOException
 {
  String smallCzechLetters =
  "a\u00e1bc\u010dd\u010fe\u00e9\u011bfghi\u00edjklmn\u0148o\u00f3pqr\u0159s\u0161t\u0165u\u00fa\u016fvwxy\u00fdz\u017e";
 // get 8-bit chars in iso-8859-2
 byte [] ISO2bytes = smallCzechLetters.getBytes("iso-8859-2");
 // convert them
 String explicit = new String(ISO2bytes,"iso-8859-2");
 String implicit = new String(ISO2bytes);
 // compare them
 System.out.println((smallCzechLetters.equals(explicit)?"Match":"Not match"));
 System.out.println((smallCzechLetters.equals(implicit)?"Match":"Not match"));
 }
}
------------------------------------------------------------------
This will produce output "Match, Not match", so default encoding
used in String constructor is not matching the encoding specified
with -Dfile.encoding.

This bugs is in SUN JDK1.1.8 on WinNT,
it is not in  customer  JDK1.1.6 on AIX
and is not in  customer .org JDK1.1.7v1a on Linux
others I haven't tried.
(Review ID: 63197) 
======================================================================

No workaround
======================================================================

file.encoding is intended to reflect the platform encoding; it is not meant to be set from the command line.
  xxxxx@xxxxx   1999-07-26

Evaluation is rather strange for me. What does "platform 
encoding" really mean? IMO impossibility of changing platform 
encoding is very unconvenient and unflexible. I had a lot of 
trouble with it.