[jdev] SOCKS5 Bytestream File Transfer problem

[Justin Karneges]

On Friday 25 February 2005 02:50 pm, Anthony Ortiz wrote:
> [1] If the Target is willing to accept the bytestream, it MUST attempt
> to open a standard TCP socket on the network address of the StreamHost
> communicated by the Initiator. If the Initiator provides more than one
> StreamHost, the Target SHOULD try to connect to them in the order they
> occur.
>
> ...
>
> [3] If the Target is able to open a TCP socket on a StreamHost, it
> MUST utilize the SOCKS5 protocol specified in RFC 1928 [7] to
> establish the connection with the StreamHost. In accordance with the
> SOCKS5 RFC, the Target MAY have to authenticate in order to use the
> proxy. However, any authentication required is beyond the scope of
> this JEP.
[snip]
> At first, my understanding was that I should open a standard TCP
> connection to the streamhost(s) (192.168.1.100:8010 in this particular
> case) in order to test if the connnection could be made (as per the
> paragraph [1]). I did this at first and that part worked fine & was
> able to connect, but then I stripped that code out because it seemed
> redundant after I read paragraph [3] which states that I will have to
> open a SOCKS5 connection to that streamhost anyway, and in the

Maybe it's not fully clear (and I'm simply reading from the quoted text you 
provided), but what [3] is intended to mean is that you would perform the 
SOCKS5 protocol over the established TCP connection from [1].  There are not 
two connections.  I suppose it is spelled out in steps like this in case 
you're writing your own SOCKS protocol code.  If you're using a SOCKS library 
to perform the connection, which will both create the TCP connection as well 
as negotiate the SOCKS5 protocol over it, then that's all you need to use.

> Proxy.setDefaultProxy(host, port); // sets SOCKS5 proxy to be
> 192.168.1.100:8010
>
[snip]
> So this tells me that the actual SOCKS5 host to connect to is the SHA1
> hash of sid + from + to and that the port is 0. That is why I have the
> following line in my code :

Well, just to ensure our terminology is clear, the SOCKS5 host (or proxy, or 
server, or whatever) is 192.168.1.100.  The destination (or DST) host, ie the 
one that you'll ask the SOCKS5 host to bridge you to, is the hash.

Of course, you're aware that the hash is not a real host, and that the SOCKS5 
host (in this example, Psi) is not actually going to perform an outbound 
connection to the DST as a result.  It is simply used as a handshake.

> Socket ss = new SocksSocket(SocksHost, 0);

My confusion with this line of code is that I don't see how it relates to the 
Proxy object.  In other words, how does the SocksSocket know what SOCKS5 host 
to connect to?  I know nothing about the library you're using though, so 
maybe the Proxy object interfaces with some global settings in the 
background, and SocksSocket draws from that.

It could be that you understand the protocol, and are coding with the right 
intent, but simply coding it wrong. ;-)

> p.s. btw Justin... great work on the PSI client!

Glad you like it. :)

Have fun,
-Justin