FULL PRODUCT VERSION :
Using Java SDK 1.4.2

ADDITIONAL OS VERSION INFORMATION :
Windows 2000 Version 5.00.2195

A DESCRIPTION OF THE PROBLEM :
The ISCII charset implementation seems to incorrectly encode characters. Characters in the \u0900 page (legal for ISCII91) get encoded with a trailing 0xff. Also, some characters seem to get translated into a sequence the decoder does not recognize, so a sequence encoded with the Charset will not decode.

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
See test case below for example. Pretty much, get a sequence of non-ASCII ISCII91 legal characters and encode them. The result looks suspicious. Decoding them fails to reproduce the original string.


/*
 * Created on Feb 9, 2004
 */
package com.pombi.lib.cvsclient.util;

import java.nio.CharBuffer;
import java.nio.charset.CharacterCodingException;
import java.nio.charset.Charset;
import java.nio.charset.CharsetEncoder;


import junit.framework.TestCase;
import sun.io.CharToByteConverter;
import sun.io.CharToByteISCII91;
import sun.io.MalformedInputException;

/**
 * @author cebert
 */
public class ISCII91Test extends TestCase {

        public ISCII91Test(String name) {
                super(name);
        }

        public final void testConvert() throws CharacterCodingException, MalformedInputException {
                CharToByteConverter toCharsOld = new CharToByteISCII91();
                Charset iscii91Charset = Charset.forName("ISCII91");
                String charsToEncode = getCharsForEncoding("ISCII91");
                byte [] oldBytes = toCharsOld.convertAll(charsToEncode.toCharArray());
                byte [] newBytes = iscii91Charset.encode(charsToEncode).array();
                for (int i = 0; i < oldBytes.length; ++i) {
                        assertEquals("At " + i, oldBytes[i], newBytes[i]);
                }
        }


        static final String getCharsForEncoding(String encodingName) throws CharacterCodingException{
                final Charset set = Charset.forName(encodingName);
                final CharBuffer chars = CharBuffer.allocate(300);
                final CharsetEncoder encoder = set.newEncoder();
                for (int c = 0; chars.remaining() > 0 && c < Character.MAX_VALUE; ++c) {
                        if (Character.isDefined((char) c) && !Character.isISOControl((char) c) && encoder.canEncode((char) c)) {
                                chars.put((char) c);
                        }
                }
                chars.limit(chars.position());
                chars.rewind();
                return chars.toString();
        }
}



EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
CharToByteISCII91 and the ISCII91 Charset should agree. A string encoded to bytes and then decoded to chars by the ISCII91 Charset should be unchanged.


REPRODUCIBILITY :
This bug can be reproduced always.

CUSTOMER SUBMITTED WORKAROUND :
Use old encoding (but have to work around new charset code)