Attached test case shows the problem, canConvert() returns "false"
for legal unicode character, but it only fails the first time, the
sequential calling rturns the correct result.

Need either fix the problme in CharToByteConverter.canConvert() or
make CharToByteUnicode have its own canConvert() implementation.


import java.util.*;
import java.nio.charset.*;
import sun.io.*;

public class TestUnicodeConverter {
    public static void main(String[] args) {
        try {
            System.out.println("----------------------");
            CharToByteConverter ctb = CharToByteConverter.getConverter("Unicode");
            System.out.println("CharToByteConverter=" + ctb);
            System.out.println("1: canConvert('A')=" + ctb.canConvert('A'));
            System.out.println("2: canConvert('A')=" + ctb.canConvert('A'));
            System.out.println("----------------------");
            Charset cs = Charset.forName("UTF-16");
            CharsetEncoder cse = cs.newEncoder();
            System.out.println("Encoder=" + cse);
            System.out.println("1: canEncode('A')=" + cse.canEncode('A'));
            System.out.println("2: canEncode('A')=" + cse.canEncode('A'));
        }
        catch (Exception e) {
            System.out.println("e=" + e);
        }
    }
}