Charset not present in used JDK

I have a java system communicating that serves as a gateway to different systems (java, mainframe, etc). This java system receives a request using, for example, utf8 and converts it to the encoding of the target

The issue is that there is a mainframe system that is using this obscure encoding that is not available with the JDK that we are using (that charset being IBM-924 which is a "derivative" of IBM-1047 https://en.wikibooks.org/wiki/Character_Encodings/Code_Tables/EBCDIC/EBCDIC_1047). system.

From what I have seen this charset seems to come only with IBM JDK, which is not the one that we use or want to use.

Is there a way to incorporate this specific charset present on the IBM JDK into, for example, an openjdk one? If not, is there any alternative beside creating the charset manually? And if so, is there a way to reuse the IBM-1047 encoding and replace the 11 characters that do not match?

I tried to incorporate the charsets by using the charsets.jar present on the IBM JDK, but it runs into issues because that jar has a dependency of classes only present on the rt.jar of the IBM JDK, and I would certainly not want to replace the rt.jar since it might cause unwanted side effects.

Oof, IBM_924 almost doesn't exist on the entire internet. Looking around, it seems to either be ISO-8859-15 straight up (which is something JDKs ship with, so just go with Charset.forName("ISO-8859-15"), or..

It's this, from ICU Data by the Unicode Consortium.

That is some random mapping file I found from some random project (admittedly, the icu-data project from the unicode organisation; as far as sources go that one should be sufficiently authoritative). You read it as follows:

A line might contain:

<UFF5E> \xA1 |1

And this means:

If you see the byte 0xA1 in the data stream...
Then that represents character 0xFF5E in the unicode table.
That |1 means, more or less: Not roundtrippable; it's an alias. There is a more canonical mapping for the same byte (0xA1) in the file earlier.

And, fortunately, being an ISO-8559 variant, it's just '1 byte = 1 character', therefore, there are only at most 256 different byte values you need to map to a character.

That means it is trivial to write your own charset implementation!

And thus you've solved all problems: House the charset inside your own project, and now you have no dependencies and no need to rely on a JVM that ships with IBM-924 baked in.

It's not too difficult.

Make a source file named Ibm924CharsetProvider.java, extending from Java’s CharsetProvider class. Copy/paste this into it:

// Written by Reinier Zwitserloot
// Public domain - use however you like.

import java.nio.ByteBuffer;
import java.nio.CharBuffer;
import java.nio.charset.Charset;
import java.nio.charset.CharsetDecoder;
import java.nio.charset.CharsetEncoder;
import java.nio.charset.CoderResult;
import java.nio.charset.spi.CharsetProvider;
import java.util.Collections;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;

public class Ibm924CharsetProvider extends CharsetProvider {
    private static final Charset IBM924 = new Ibm924Charset();
    private static final List<Charset> IBM924_LIST = Collections.singletonList(IBM924);
    
    public static final Charset ibm924() {
        return IBM924;
    }
    
    @Override public Iterator<Charset> charsets() {
        return IBM924_LIST.iterator();
    }
    
    @Override public Charset charsetForName(String charsetName) {
        if (charsetName.equals(IBM924.name())) return IBM924;
        if (IBM924.aliases().contains(charsetName)) return IBM924;
        
        return null;
    }
    
    public static void main(String[] args) throws Exception {
        String in = "Hello, World!";
        byte[] b = in.getBytes(ibm924());
        System.out.print(in + " ->");
        for (byte a : b) System.out.printf(" %02X", a);
        System.out.println();
        String in2 = new String(b, ibm924());
        System.out.println("... -> " + in2);
        System.out.println("Equal? " + in.equals(in2));
    }
}

class Ibm924Charset extends Charset {
    public Ibm924Charset() {
        super("IBM-924", new String[] { "IBM924", "Cp924", "Cp-924" });
    }
    
    @Override public boolean contains(Charset cs) {
        return cs.name().equals("IBM-924"); // It isn't even a subset of ASCII!
    }
    
    @Override public CharsetDecoder newDecoder() {
        return new CharsetDecoder(this, 1F, 1F) {
            @Override protected CoderResult decodeLoop(ByteBuffer from, CharBuffer to) {
                while (from.hasRemaining()) {
                    if (!to.hasRemaining()) return CoderResult.OVERFLOW;
                    byte c = from.get();
                    char d = IBM924[c & 0xFF];
                    to.put(d);
                }
                
                return CoderResult.UNDERFLOW;
            }
        };
    }
    
    @Override public CharsetEncoder newEncoder() {
        return new CharsetEncoder(this, 1F, 1F) {
            @Override protected CoderResult encodeLoop(CharBuffer from, ByteBuffer to) {
                while (from.hasRemaining()) {
                    if (!to.hasRemaining()) return CoderResult.OVERFLOW;
                    char d = from.get();
                    Byte v = LOOKUP.get(d);
                    if (v == null) {
                        // 'un'consume the character we consumed
                        from.position(from.position() - 1);
                        return CoderResult.unmappableForLength(1);
                    }
                    to.put(v.byteValue());
                }
                
                return CoderResult.UNDERFLOW;
            }
        };
    }
    
    // SOURCE: https://github.com/unicode-org/icu-data/blob/main/charset/data/ucm/ibm-924_P100-1998.ucm
    private static final char[] IBM924 = {
        '\u0000', '\u0001', '\u0002', '\u0003', '\u009C', '\u0009', '\u0086', '\u007F',
        '\u0097', '\u008D', '\u008E', '\u000B', '\u000C', '\r', '\u000E', '\u000F',
        '\u0010','\u0011','\u0012','\u0013', '\u009D','\u0085','\u0008','\u0087',
        '\u0018','\u0019','\u0092','\u008F', '\u001C','\u001D','\u001E','\u001F',
        '\u0080','\u0081','\u0082','\u0083', '\u0084','\n','\u0017','\u001B',
        '\u0088','\u0089','\u008A','\u008B', '\u008C','\u0005', '\u0006', '\u0007',
        '\u0090', '\u0091', '\u0016', '\u0093', '\u0094', '\u0095', '\u0096', '\u0004',
        '\u0098', '\u0099', '\u009A', '\u009B', '\u0014', '\u0015', '\u009E', '\u001A',
        '\u0020', '\u00A0', '\u00E2', '\u00E4', '\u00E0', '\u00E1', '\u00E3', '\u00E5',
        '\u00E7', '\u00F1', '\u00DD', '\u002E', '\u003C', '\u0028', '\u002B', '\u007C',
        '\u0026', '\u00E9', '\u00EA', '\u00EB', '\u00E8', '\u00ED', '\u00EE', '\u00EF',
        '\u00EC', '\u00DF', '\u0021', '\u0024', '\u002A', '\u0029', '\u003B', '\u005E',
        '\u002D', '\u002F', '\u00C2', '\u00C4', '\u00C0', '\u00C1', '\u00C3', '\u00C5',
        '\u00C7', '\u00D1', '\u0160', '\u002C', '\u0025', '\u005F', '\u003E', '\u003F',
        '\u00F8', '\u00C9', '\u00CA', '\u00CB', '\u00C8', '\u00CD', '\u00CE', '\u00CF',
        '\u00CC', '\u0060', '\u003A', '\u0023', '\u0040', '\'', '\u003D', '\u0022',
        '\u00D8', '\u0061', '\u0062', '\u0063', '\u0064', '\u0065', '\u0066', '\u0067',
        '\u0068', '\u0069', '\u00AB', '\u00BB', '\u00F0', '\u00FD', '\u00FE', '\u00B1',
        '\u00B0', '\u006A', '\u006B', '\u006C', '\u006D', '\u006E', '\u006F', '\u0070',
        '\u0071', '\u0072', '\u00AA', '\u00BA', '\u00E6', '\u017E', '\u00C6', '\u20AC',
        '\u00B5', '\u007E', '\u0073', '\u0074', '\u0075', '\u0076', '\u0077', '\u0078',
        '\u0079', '\u007A', '\u00A1', '\u00BF', '\u00D0', '\u005B', '\u00DE', '\u00AE',
        '\u00A2', '\u00A3', '\u00A5', '\u00B7', '\u00A9', '\u00A7', '\u00B6', '\u0152',
        '\u0153', '\u0178', '\u00AC', '\u0161', '\u00AF', '\u005D', '\u017D', '\u00D7',
        '\u007B', '\u0041', '\u0042', '\u0043', '\u0044', '\u0045', '\u0046', '\u0047',
        '\u0048', '\u0049', '\u00AD', '\u00F4', '\u00F6', '\u00F2', '\u00F3', '\u00F5',
        '\u007D', '\u004A', '\u004B', '\u004C', '\u004D', '\u004E', '\u004F', '\u0050',
        '\u0051', '\u0052', '\u00B9', '\u00FB', '\u00FC', '\u00F9', '\u00FA', '\u00FF',
        '\\', '\u00F7', '\u0053', '\u0054', '\u0055', '\u0056', '\u0057', '\u0058',
        '\u0059', '\u005A', '\u00B2', '\u00D4', '\u00D6', '\u00D2', '\u00D3', '\u00D5',
        '\u0030', '\u0031', '\u0032', '\u0033', '\u0034', '\u0035', '\u0036', '\u0037',
        '\u0038', '\u0039', '\u00B3', '\u00DB', '\u00DC', '\u00D9', '\u00DA', '\u009F',
    };
    
    private static final Map<Character, Byte> LOOKUP;
    static {
        Map<Character, Byte> map = new HashMap<>();
        for (int i = 0; i < IBM924.length; i++) map.put(IBM924[i], (byte) i);
        LOOKUP = Collections.unmodifiableMap(map);
    }
}

To use as provider

If you want e.g. new String(bytes, "IBM-924") to work, as well as having it listed in the list of all available charsets, this class needs to be registered. To do this, you stick the fully qualified class name (com.foo.yourapp.util.Ibm924CharsetProvider, that string, for example) in an otherwise empty text file. You name this textfile java.nio.charset.spi.CharsetProvider. This file needs to be in your jar, at the path META-INF/services/java.nio.charset.spi.CharsetProvider. Then, if that jar is on your classpath, things should just work.

Oof that sounds complicated

You don't really need to do that - that's just to make sure "IBM-924" as a string works out. You can use it as is without registering the provider. See the main method in the snippet where I don't bother registering it but use the Charset object directly.

Run that, and it prints:

Hello, World! -> C8 85 93 93 96 6B 40 E6 96 99 93 84 5A
... -> Hello, World!
Equal? true

NB: This simplistic take doesn't deal with the aliases. They are only relevant for encoding (if you need to output IBM-924 formatted text); they aren't useful for decoding (reading IBM-924 formatted text). And only if you use these aliases in your strings.

To use as provider

Oof that sounds complicated

Recommended topics

Hot tags