Question

I have to encode in base64 an array of bytes (that is a byte stream of a Zipfile) but the array is too big. So I split the array in two parts, encode the first half, then the second one, and in the end I append the second part to the first part. The thing is that sometimes when i decode the result the zipfile seems to be corrupted.

    FileInputStream fis = new FileInputStream(new File(filePath));
    ByteArrayOutputStream baos = new ByteArrayOutputStream();

    while((len = fis.read(buffer)) >= 0) {
        baos.write(buffer);
    }

    byte[] bArray = baos.toByteArray();

    int bArrayLength = bArray.length;

    byte[] part1 = new byte[bArray.length / 2];
    byte[] part2 = new byte[ (bArrayLength%2 == 0) ? (bArray.length / 2) : (bArray.length / 2 + 1)];

    System.arraycopy(bArray, 0, part1, 0, part1.length);
    System.arraycopy(bArray, part1.length, part2, 0, part2.length);

    resultArray[0] = Base64.encodeBytes(part1);
    resultArray[1] = Base64.encodeBytes(part2);

What is the problem? Is that split that causes the issue?

Answer 1

This probably doesn’t work because of padding. The Base64 algorithm pads its output with a == if its encoding results in a trailing group (of 24 bits) that contains only 1 byte, or = if it contains only 2 bytes. That means you cannot just add two encoded strings together and expect to get the correct result.

But how do you know the array is too big — do you get an OutOfMemoryException?

You may try to split not exactly at the halfway point, but choose a different point until you no longer get a trailing = for that part.

See also http://en.wikipedia.org/wiki/Base64.