I tried to run attached testcase (Cp943Test), which checks the Cp943 code range.

Cp943Test returns U+0000 and U+FFFD for unconvertable character.

For example,
0x81ad: 0000
0x81ae: 0000
0x81af: 0000
0x81b0: 0000
...
0x8581: fffd
0x8587: fffd
0x85e0: fffd
0x85f0: fffd

What's difference these two unconvertable character?
I think it should returns U+FFFD instead of U+0000.

=====================================================================================
import java.io.*;

class Cp943Test {
    public static void main (String[] args) {
       byte[] test = new byte[2];
       char[] charBuffer = new char[1];
       sun.io.ByteToCharCp943 convertor = new sun.io.ByteToCharCp943();
       for(int i = 0x81; i<= 0xfc; i++) {
           if (0x9f < i && i < 0xe0) continue;
           for(int j = 0x40; j <= 0xfc; j++) {
               if (j == 0x7f) continue;
               test[0] = (byte) i;
               test[1] = (byte) j;
               try {
                  int nc = convertor.convert (test, 0, 2, charBuffer, 0, 1);
                  System.out.printf("0x%02x%02x: %04x\n", i, j, (int)charBuffer[0]);
               }
               catch (Exception e) { }
           }
        }
    }
}