Unsupported or invalid address type (2) in SOCKS5 request

[jz]

I'm trying to use Wingate as a SOCKS proxy to use with Fiorano JMS server. But wingate does not accept the client connection and logs:

04/08/05 09:08:33 <client ip> Guest 0000000278 Created:

04/08/05 09:08:33 <client ip> Guest 0000000278 Error: Unsupported or invalid address type (2) in SOCKS5 request - terminating

04/08/05 09:08:33 <client ip> Guest 0000000278 Traffic 12 5 0 0 0s

04/08/05 09:08:33 <client ip> Guest 0000000278 Terminated exit code 1

Anyone can help?
Does a list of Wingate errors exists somewhere?
Thank you in advance,
Jacopo Zamberlan

[adrien]

What is the client software?

SOCKS 5 defines 3 address types, 1, 3, and 4. 2 is not defined in the spec, so it looks like the client is not SOCKS compliant. I can't find any reference to an additional address type being defined.

Adrien

[jz]

>What is the client software?
I use the client of a jms server: Fiorano MQ.

>SOCKS 5 defines 3 address types, 1, 3, and 4. 2 is not defined in the >spec, so it looks like the client is not SOCKS compliant. I can't find any >reference to an additional address type being defined.


I don't think the client is not SOCKS compliant at all, because I found two SOCKS proxy servers who let me connect without problems.
Perhaps it uses SOCKS4 and does not support SOCKS5: coud it be this the case?

Thank you,
Jacopo Zamberlan

[adrien]

you wouldn't get that error then, the protocol version field in the request would be 4 instead of 5, and that error you see only occurs if the client specified v5.

It is possible some other socks servers ignore the address type field if it is unknown, and assume it is IP4.

You aren't running an IPv6 network or something are you? there was an experimental additional address type defined for IP4 to IP6 bridging (ATYP 8), but that was never ratified.

All info on SOCKS is looked after by NEC, at www.socks.nec.com. If you read all the specs like I did, you'll see there is no mention of any address type 2. I would be interested to hear what the developers of the client have to say about it.

Adrien

[jz]

Developers of the client API (very good Fiorano support) found the problem: it is a java bug of jdk1.4, solved in 1.5 (http://bugs.sun.com/bugdatabase/view_bu ... id=5025217)

Thanks,
Jacopo Zamberlan