FULL PRODUCT VERSION :
java version "1.8.0_92"
Java(TM) SE Runtime Environment (build 1.8.0_92-b14)
Java HotSpot(TM) 64-Bit Server VM (build 25.92-b14, mixed mode)

Tried latest 1.8.0 also.

A DESCRIPTION OF THE PROBLEM :
Decoding a byte array to a string using the charset "MS50220" fails to decode some characters which the real Microsoft code page 50220 can decode.


STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
Run the test program.

EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
Should print:

鄧

If there is any doubt, here is the equivalent program written in C#, and you can run it here:
https://dotnetfiddle.net/zDmDiG

namespace test
{
class MainClass
{
public static void Main(string[] args)
{
byte[] data = { 0x1B, 0x24, 0x42, 0x7B, 0x7D, 0x1B, 0x28, 0x42 };
string str = System.Text.Encoding.GetEncoding(50220).GetString(data);
System.Console.WriteLine(str);
}
}
}

ACTUAL -
Actually prints what looks like nothing, but it seems to be a special character.


REPRODUCIBILITY :
This bug can be reproduced always.

---------- BEGIN SOURCE ----------
class Test
{
    public static void main(String[] args) throws Exception
    {
        byte[] data = { 0x1B, 0x24, 0x42, 0x7B, 0x7D, 0x1B, 0x28, 0x42 };
        String str = new String(data, "MS50220");
        System.out.println(str);
    }
}

---------- END SOURCE ----------

CUSTOMER SUBMITTED WORKAROUND :
Use "x-windows-iso2022jp" instead of "MS50220" . But easier said than done when the code page declaration is inside a file parsed by someone else's code.