Read byte value string characters

Question

I am trying to implement a communication protocol in JavaScript (language mandatory). The communication is done in TCP/IP thanks to webtcp.

Every packet is divided this way: content size in big endian + content.

So for a packet of a size 28, let's say {"type": "success", "id": 0}, the packet sent will be 0,0,0,28,'{','"', 't', etc....

I have no problem sending packets in pure JavaScript using this syntax.

The problem is, because of WebTCP, the packet I get on my end is always a string, and when the size of a packet is between 128 and 255 (I'm guessing, I just know that it is greater than 128), the size is read wrong. I think I know where the problem is:

Here is my function which extracts the data.

function extractData() {
    // return empty string if not enough bytes to read the size of next packet
    if (!buffer || buffer.length < 4) { return ""; }

    // read size
    var size = [];
    for (var idx = 0 ; idx < 4 ; ++idx) {
        size.push(buffer.charCodeAt(idx)); // pretty sure the problem comes from here.
    }

    // size from big endian to machine endian
    size = ntohl(size);

    // return empty string if the buffer does not have the complete packet yet
    if (buffer.length < 4 + size) { return ""; }

    // copy the packet content into ret
    var ret = "";
    for (var idx = 4 ; idx < size + 4 ; ++idx) {
        ret += buffer[idx];
    }

    // the buffer removes the packet returned
    buffer = buffer.substring(size + 4);

    // return the packet content
    return ret;
}

buffer is a global variable which is filled every time data is received. ntohl is a function I got from http://blog.couchbase.com/starting-membase-nodejs (without the i offset) which takes a 4 bytes array and returns an integer.

So the line at fault would be size.push(buffer.charCodeAt(idx));, I'm guessing the charCodeAt function overflows when the character code given is greater than an ASCII value (0-127). From printing on the server side (which works, I tried in python and C++), the size sent is 130, and on the JavaScript side, the size array contains [0, 0, 0, 65533] (or something like this, I don't remember the right number. With a size of 30 I get [0, 0, 0, 30] so I know that this is supposed to work.

I have several questions :

1. How can I extract the raw integer value of a char in a string ?
2. Is there an easy way to turn the 4 first bytes into something like a bytearray ? Using only JS and jQuery.

Thanks.

Answer 1

Well according to the definition of ntohl it excepts an integer and an array.

exports.ntohl = function(b, i) {
  return ((0xff & b[i + 0]) << 24) |
         ((0xff & b[i + 1]) << 16) |
         ((0xff & b[i + 2]) << 8) |
         ((0xff & b[i + 3]));
} 

where b is an array and i is an integer.

You are only passing a single parameter in your code sample.

If I pass 0, it returns a value if I pass no parameter it returns 0.

Try your code with nthol (size, 0 );

Here is a fiddle demonstrating this: http://jsfiddle.net/3jxrr3k7/