On Thu, 2012-10-04 at 22:36 +0000, John Case wrote: > Tom, Andrew, > > On Thu, 4 Oct 2012, Tom Ritter wrote: > > > Of course those are the huge, monolithic cases. Take simpler apps > > like gpg, ssh, putty, pidgin (god help us), git, svn. While tracking > > upstream would certainly be a problem, having a statically linked tor > > and a modified binary that sent everything over Tor I think would go a > > long way towards getting average users using Tor safely... without > > ever having to say the words "Proxy" or "Socks" to them. > > > I'm a unix engineer and a coder, and even *I* sometime just don't feel up > to constructing an entire proxy environment just to make an SSH connection > over Tor. > > And it's not just a case of setup time - there is assurance and testing > and periodic fiddling with just to make sure you did it right, have it > right and will continue to have it right. And then you wake up at night > and wonder if you really, really have the DNS leaking taken care of. > > I know it's a big can of worms, since everyone has their favorite little > binaries, but surely ssh is a good start :) Have you tried 'torsocks <favorite little binary>'? Tor doesn't really make sense as a static library, or a library at all. If you want to get proxied sockets, use any SOCKS library and point it at Tor. Tor (running only as an Onion Proxy or "client") has a large amount of application-neutral state -- would every 'favorite little binary' set up a separate set of circuits pre-emptively? If you can sit down with the Tor source code and come up with a compelling narrative for a Tor library that isn't just a SOCKS library, I'd be interested in reading it, but I just don't see how the current Tor design/implementation is in any way amenable to this. -- Sent from Ubuntu
Attachment:
signature.asc
Description: This is a digitally signed message part
_______________________________________________ tor-talk mailing list tor-talk@xxxxxxxxxxxxxxxxxxxx https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-talk