| |
|
|
|
cURL Mailing List Monthly Index Single Mail
curl-users Mailing List Archives
Re: curl vs. socks5
From: Jochen Hayek <Jochen+curl_at_Hayek.name>
Date: Mon, 12 Feb 2007 15:50:59 +0100
>>>>> "DS" == Daniel Stenberg <daniel_at_haxx.se> writes:
Hi, Daniel,
maybe our communication speed is a little to fast,
I don't want to sound insulting,
DS> On Mon, 12 Feb 2007, Jochen Hayek wrote:
DS> Yes, you tell the server to connect to a private IP address:
DS> PORT 10,0,4,119,180,254
>> Yes, it looks like that -- but when the socks server goes through
DS> The fact that this works with "socksify" is then only because that
Right, "socksify" replaces my *client* *host*'s private IP address by my *socks* *host*'s private IP address.
But (as my previous posting tried to show) socksify also does this for curl -- and still curl does know, what I expect it to do.
And "curl --socks ..." also applies this replacement -- still to no final success.
And the "*magical*" translation from private addresses (i.e. the private address of my socks server) to my static public IP address
DS> curl doesn't magically translate for you,
DS> This isn't strictly a SOCKS issue.
>> Today's NAT routers are wonderful little (Linux) machines :-)
DS> Perhaps, but this problem has nothing to do with the actual NAT or
No, dear friend!
DS> that can't reach you back on that address.
>> So that private IP address should not be a problem.
DS> Yes it should. Unless you introduce a middle-man that translates it for you.
The NAT router.
DS> BTW, protocol-wise, when you use SOCKS for an outgoing
Sure, yes.
Actually my socks server ("dante") only did the right thing
Cheers,
These mail archives are generated by hypermail. |
Page updated November 12, 2010.
web site info