Java: BufferedReader Keeps Writing Values 128-159 as 63 When Converting to Char

Question

I am trying to write a hex editor. I'm trying to store values by writing a char to a text file. For some reason every decimal number 128-159 is being written or read (not sure which) as 63. I took measures to isolate the problem. Here's an example of it happening:

import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.io.File;

public class Why {

    public static File file = new File("why.txt");

    public static void main(String[] args) throws IOException {
        if(!file.exists())
            file.createNewFile();

        BufferedWriter bw = new BufferedWriter(new FileWriter(file));
        bw.write((char) 144);
        bw.close();

        BufferedReader br = new BufferedReader(new FileReader(file));
        System.out.println(br.read());
        br.close();
    }
}

Any help is appreciated.

I figured it out using a FileOutputStream and FileInputStream. Thanks all.

Answer 1

When you use FileReader and FileWriter, they will use the default encoding for your platform. That's almost always a bad idea.

In your case, it seems that that encoding doesn't support U+0092, which is fairly reasonable given that it's a private use character - many encodings won't support that. I suspect you don't actually want (char) 144 at all. If you really, really want to use that character, you should use an encoding which can encode all of Unicode - I'd recommend UTF-8.

It's important to differentiate between text and binary, however - if you're really just interested in bytes, then you shouldn't use a reader or writer at all - use an InputStream and an OutputStream. Hex editors are typically byte-oriented rather than text-oriented, although they may provide a text view as well (ideally with configurable encoding). If you want to know the exact bytes in the file, you should definitely be using FileInputStream.