[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: Access for the uncomputed



I see the problem as this: not everyone can run Tor on the machines from
which they browse the web, and not everyone can configure their proxy
settings to point at an open Tor proxy.  

I hacked serifos to function as a proxy that sends traffic through Tor.
It is not entirely perfect, but it usually works.  Known limitations
include failure to properly munge image and link tags on SSL-encrypted
pages and an inability to properly handle redirections resulting from
Javascript code.

So, you can visit:

http://serifos.eecs.harvard.edu/proxy/TOR_HTTP_URL

where TOR_HTTP_URL is any HTTP or HTTPS URL, and the Tor client running
on serifos will take you to your specified URL.  For example,

http://serifos.eecs.harvard.edu/proxy/http://www.whatismyip.com/

This is how it works:

First, I configured apache on serifos to use mod_proxy:

<IfModule mod_proxy.c>
    ProxyPass /proxy/ http://localhost:8119/
</IfModule>

Then, I ran my own proxy script that feeds requests to Privoxy and
munges HTML replies to properly translate image and link tags.  The code
is here:

http://afs.eecs.harvard.edu/~goodell/blossom/src/edgeproxy

I invoke this script as follows:

$ edgeproxy -l localhost:8119 -r localhost:8118

Finally, I ran Privoxy on port 8118, in the normal manner specified
on the Tor website.

Please check it out if you are interested and feed me bug reports if you
find anything broken (other than what I described).

Thanks,

Geoff

On Thu, Oct 20, 2005 at 11:08:43PM +0800, Patrick Coleman wrote:
> CGIProxy is quite good in my experience; it seems to proxify URLs
> better than CECID does. It performs exactly the same task as CECID,
> though CECID has a few extra features (banned word filtering).
> 
> I did mention a while ago that I would modify CECID to work as a
> frontend for tor, but its not looking like I'm going to get time to do
> that anytime soon (though I live in eternal hope). The codes all in
> CVS (http://cecid.sf.net), if anyone's interested in picking it up and
> working on it drop me a line.
> 
> -Patrick
> 
> On 20/10/05, Joel Franusic <jfranusic@xxxxxxxxx> wrote:
> > I just ran across: CGIProxy
> > (http://www.jmarshall.com/tools/cgiproxy/cgiproxy-beta.html)
> >
> > A Proxy over CGI of sorts, similar to CECID (?). This looks like a
> > perfect front end for Tor.
> >
> > It supports SSL and it looks like it can be easily configured to use a
> > proxy (Tor).
> >
> > Has anybody tried this out?
> >
> > --Joel
> >
> > On 6/22/05, Patrick Coleman <blinken@xxxxxxxxx> wrote:
> > > Brilliant. I'll see if I cant get something going.
> > > Thanks,
> > > Patrick
> > >
> > > Roger Dingledine wrote:
> > > > On Wed, Jun 22, 2005 at 10:45:17AM +0800, Patrick Coleman wrote:
> > > >
> > > >>shouldn't be too hard. I was actually considering interfacing it
> > > with a proper anonymizer at some
> > > >>point, like Tor, so I'd be happy to do that if thats what you want.
> > > >
> > > >
> > > > That would be wonderful. We really do need something like this, that
> > > > lets people point their browsers somewhere and be able to access .exit
> > > > or .onion addresses.
> > > >
> > > > It should be even easier to find mirrors for you now too, because the
> > > > mirrors don't need to be exiting the traffic themselves.
> > > >
> > > > Thanks,
> > > > --Roger
> > > >
> > > >
> > >
> > >
> > > On 23/06/05, Patrick Coleman <blinken@xxxxxxxxx> wrote:
> > > > [I'll mail this to the list - I am subscribed, but at blinken@xxxxxxxxx]
> > > >
> > > > Hey,
> > > > The client certainly hasn't had any work done on it for ages, so I was
> > > > thinking of ditching that, certainly after I discovered tor. It was
> > > > certainly a bit more complex than I bargained for :)
> > > >
> > > > With the script, it hasn't been developed in quite a while. I have
> > > > been intending to do some work on it, though - I've got some working
> > > > code that should fix a few problems, like SSL, forms and cookies.
> > > > These fixes will also mean a rewrite of the HTTP fetching code, so
> > > > working in HTTP proxying shouldn't be too hard. I was actually
> > > > considering interfacing it with a proper anonymizer at some point,
> > > > like Tor, so I'd be happy to do that if thats what you want.
> > > >
> > > > The script -shouldn't- be breaking stylesheets, so I'll have a look :)
> > > > Thanks,
> > > > Patrick
> > > > +++
> > > > Public Key ID 0x4A6880B2
> > > > Key Fingerprint: 7867 E238 1608 1A20 89C4  BA6C 8FC3 C6EB 4A68 80B2
> > > > http://warhn.org/pcoleman/pubkey.txt
> > > >
> > > > On 22/06/05, Roger Dingledine <arma@xxxxxxx> wrote:
> > > > > On Tue, Jun 21, 2005 at 03:26:33PM -0700, Joel Franusic wrote:
> > > > > > Some quick searches on sf.net and freshmeat.net turn up:
> > > > > > http://cecid.sourceforge.net/
> > > > > >
> > > > > > Links to servers running CECID:
> > > > > > http://cecid.sourceforge.net/mirrors.php
> > > > >
> > > > > Oh hey, and Patrick Coleman runs a Tor server too:
> > > > > http://serifos.eecs.harvard.edu:8000/cgi-bin/desc.pl?q=hal
> > > > >
> > > > > Patrick, how is this going? It looks like Tor can replace the more
> > > > > ambitious part of your project, but step one is still a hard task to
> > > > > get right too. :)
> > > > >
> > > > > It looks like it's GPL, which is good. But it looks like it breaks
> > > > > stylesheets of the pages it downloads (e.g. tor.eff.org), which is
> > > > > bad. What about SSL to the proxy page? Does it have a back-end that can
> > > > > http-proxy to privoxy, and/or socks4a-proxy to Tor?
> > > > >
> > > > > Is this still in development, or should I take the "Copyright 2003"
> > > > > to be a bad sign? :)
> > > > >
> > > > > Thanks,
> > > > > --Roger
> > > > >
> > > > >
> > > >
> > > > --
> > > > Public Key ID 0x4A6880B2
> > > > Key Fingerprint: 7867 E238 1608 1A20 89C4  BA6C 8FC3 C6EB 4A68 80B2
> > > > http://warhn.org/pcoleman/pubkey.txt
> > > >
> > >
> > >
> > > --
> > > Public Key ID 0x4A6880B2
> > > Key Fingerprint: 7867 E238 1608 1A20 89C4  BA6C 8FC3 C6EB 4A68 80B2
> > > http://warhn.org/pcoleman/pubkey.txt
> > >
> >
> 
> 
> --
> Public Key ID 0x4A6880B2
> Key Fingerprint: 7867 E238 1608 1A20 89C4  BA6C 8FC3 C6EB 4A68 80B2
> http://warhn.org/pcoleman/pubkey.txt

Attachment: signature.asc
Description: Digital signature