Question

I have a problem with getting parts of the ByteArray data.
There is a binary text in fileData:

var fileData:ByteArray = new ByteArray();
//..........here's code that fills this var with binary data
.....readBytes(fileData,0,1000);
//

Data is like this:

йYЯyeSВ–нkq(г<<<start>>>:xЪмЅdf”cйxЪsdfмЅ”cйdxЪмЅ”cй<<<end>>>В–нkВ

So, I need to find position of <<< start >>> and <<< end >>> and copy data, that is between them.

But searching fileData.toString().indexOf('<<< start >>>') sometimes gets wrong position of this string, and sometimes can’t find it at all.

What can I do to correctly determine the position of part of the data that I need?

Answer 1

You should not use fileData.toString().indexOf() since you are working with binary data. You have to search a sequence of bytes.

The following function retrieve the position of a specified pattern:

public function indexOf(bytes:ByteArray, search:String, startOffset:uint = 0):void
{
    if (bytes == null || bytes.length == 0) {
        throw new ArgumentError("bytes parameter should not be null or empty");
    }

    if (search == null || search.length == 0) {
        throw new ArgumentError("search parameter should not be null or empty");
    }

    // Fast return is the search pattern length is shorter than the bytes one
    if (bytes.length < startOffset + search.length) {
        return -1;
    }

    // Create the pattern
    var pattern:ByteArray = new ByteArray();
    pattern.writeUTFBytes(search);

    // Initialize loop variables
    var end:Boolean;
    var found:Boolean;
    var i:uint = startOffset;
    var j:uint = 0;
    var p:uint = pattern.length;
    var n:uint = bytes.length - p;

    // Repeat util end
    do {
        // Compare the current byte with the first one of the pattern
        if (bytes[i] == pattern[0]) {
            found = true;
            j = p;

            // Loop through every byte of the pattern
            while (--j) {
                if (bytes[i + j] != pattern[j]) {
                    found = false;
                    break;
                }
            }

            // Return the pattern position
            if (found) {
                return i;
            }
        }

        // Check if end is reach
        end = (++i > n);
    } while (!end);

    // Pattern not found
    return -1;
}

Then you can use the function this way:

var extractedBytes = new ByteArray();
var startPos:int = indexOf(fileData, "<<<start>>>");
var endPos:int;

if (startPos == -1) {
    trace("<<<start>>> not found");
} else {
    endPos = indexOf(fileData, "<<<end>>>", startPos + 11); // "<<<start>>>".length = 11
}

if (startPos == -1) {
    trace("<<<end>>> not found");
} else {
    // Extract the bytes between <<<start>>> and <<<end>>>
    fileData.readBytes(extractedBytes, startPos + 11, endPos);
}

Disclaimer: I haven’t tested my code!