Send DHT queries to "router.bittorrent.com" response garbled text

Question

I read the DHT Protocol in bep_0005 page.

But when I send a ping query or a find_node query, the server response a garbled text (both of router.bittorrent.com:6881 or dht.transmissionbt.com:6881)

Here is the Java source code bellow

    public String ping(final String id) {
    System.out.println("Start ping:" + id);
    Bencode bencode = new Bencode();
    byte[] encoded = bencode.encode(new HashMap<Object, Object>() {
        private static final long serialVersionUID = 4225164001818744013L;

        {
            put("t", "tr");
            put("y", "q");
            put("q", "ping");
            put("a", new HashMap<Object, Object>() {
                private static final long serialVersionUID = -6092073963971093460L;

                {
                    put("id", id);
                }
            });
        }
    });
    byte[] result = client.send(new String(encoded, bencode.getCharset()));
    Map<String, Object> dict = bencode.decode(result, Type.DICTIONARY);
    System.out.println("Bdecoded Data:" + dict);
    return "";
}

Send Packets

ping Query = {"t":"aa", "y":"q", "q":"ping", "a":{"id":"abcdefghij0123456789"}}

bencoded = d1:ad2:id20:abcdefghij0123456789e1:q4:ping1:t2:aa1:y1:qe

Acrodding to the bep_0005 protocol the response with be like:

Response = {"t":"aa", "y":"r", "r": {"id":"mnopqrstuvwxyz123456"}}

bencoded = d1:rd2:id20:mnopqrstuvwxyz123456e1:t2:aa1:y1:re

But my response is:

Response = {ip=��P��, r={id=2�NisQ�J�)ͺ��F|�g}, t=tr, y=r}

bencoded = d2:ip6:��P��1:rd2:id20:2�NisQ�J�)ͺ��F|�ge1:t2:tr1:y1:re

Send udp part Java code is:

    public byte[] send(String sendData) {
    DatagramSocket client;
    try {
        client = new DatagramSocket();
        client.setSoTimeout(5000);
        byte[] sendBuffer;
        sendBuffer = sendData.getBytes();
        InetAddress addr = InetAddress.getByName("router.bittorrent.com");
        int port = 6881;
        DatagramPacket sendPacket = new DatagramPacket(sendBuffer, sendBuffer.length, addr, port);
        client.send(sendPacket);
        byte[] receiveBuf = new byte[512];
        DatagramPacket receivePacket = new DatagramPacket(receiveBuf, receiveBuf.length);
        client.receive(receivePacket);
        System.out.println("Client Source Data:" + Arrays.toString(receivePacket.getData()));
        String receiveData = new String(receivePacket.getData(), "UTF-8");
        System.out.println("Client String Data:" + receiveData);
        client.close();
        return receivePacket.getData();
    } catch (SocketException e) {
        e.printStackTrace();
    } catch (UnknownHostException e) {
        e.printStackTrace();
    } catch (IOException e) {
        e.printStackTrace();
    }
    return null;
}

read the response in UTF-8 , but the iso-8859-1 is also a garbled text.

who can help me,thanks!

if you want to talk to the mainline dht you could use an [existing library](https://github.com/the8472/mldht). it also contains a prettyprinter for bencoding. — the8472, Jul 21 '17 at 06:06

Encombe · Accepted Answer · 2017-07-26T06:49:54.350

1

the server response a garbled text

No, the response is bencoded and contains raw binary data.
It CAN NOT be treated as text.

In BEP5, to make the raw binary node_id in the examples printable, it has cleverly been chosen to consist of only alphanumeric characters.
See:
Bittorrent KRPC - Why are node ID's half the size of an info_hash and use every character a-z?

The ip key is a extension explained in: BEP42 - DHT Security extension

The received response is fully valid.

TODO: Working Java code

edited Jul 26 '17 at 06:49

answered Jul 19 '17 at 06:23

Encombe

2,003
1
17
26

Thanks for your response! if the response cant not be treated as text,then how can i get response info? to get ip and id value? – fish Jul 20 '17 at 06:41
Sorry! Can't help you with that. I don't java. Try add a [tag:java] tag to your question. – Encombe Jul 20 '17 at 08:25
Also thank you very much. I've added java tag .Can you add your develop language code? is it hex decode is right? – fish Jul 21 '17 at 02:17
Maybe this helps: https://stackoverflow.com/questions/45043257/why-is-hashing-the-info-dict-turning-out-wrong – Encombe Jul 21 '17 at 10:03

score 0 · Answer 2 · answered Jul 26 '17 at 09:59

Map<String, Object> dict = bencode.decode(result, Type.DICTIONARY);

This gives you the decoded root dictionary of the message as Map. Within that you should find the r dictionary as another map and with in that map the id value. What type the id has will depend on the bedecoding library you are using.

If it is ByteBuffer or byte[] then you should have 20 bytes that you can hexencode (to 40 characters) if you need it to be human-readable. The DHT protocol deals in raw hashes, not hex values.

If it is a String then you will have to convert the string back into byte[] before hex-encoding it. That is only possible when the bdecoder used ISO 8859-1 to decode because that charset is roundtrip-safe while utf-8 is not for arbitrary byte sequences.

thank you! I want to know if i want to do something with the id value, is it necessary decode it to human-readable string? or i just use it in byte[]? — fish, Jul 28 '17 at 01:33

Send DHT queries to "router.bittorrent.com" response garbled text

2 Answers2