Polipo web proxy (was Re: Tor and Google Image search)

[Roger Dingledine]

On Fri, Aug 18, 2006 at 06:42:28PM -0500, Mike Perry wrote:
> Actually, I've started noticing this even though my privoxy config
> hasn't changed in a long while. I think its something new that
> images.google.com is doing that privoxy doesn't like.

Exciting. Thanks for tracking this down.

> Perhaps the images.google.com declairation should be added to the
> Privoxy that is shipped with vidalia/tor. It is likely to be pretty
> frustrating to new users.

What I'd like to do actually is move to some other http proxy one day.

(I once dreamed of taking the http proxy out of the loop entirely now that
Firefox supports safe socks, but it turns out that the entire networking
component of Firefox blocks during socks handshakes, so that is not an
option until somebody does a major overhaul of Firefox.)

I am thinking Polipo is a nice next option:
http://www.pps.jussieu.fr/~jch/software/polipo/
I've been using it the past month or two with good success. Can other
people here give it a try and see if we can clean it up? (You will need
the latest development version.)

Known issues when using Polipo with Tor:

1) In the config.sample, it suggests
socksParentProxy = "localhost:9050"
You should either change this to 127.0.0.1:9050, or enable the
dnsUseGethostbyname config option -- otherwise polipo asks your name
servers where "localhost" is, with possibly disastrous implications.

2) Polipo doesn't do as much application-level scrubbing as Privoxy tries
to do. But Privoxy isn't very good at it anymore anyway, and Firefox
is getting better. See previous threads about all the Firefox plugins
you need so you can discard Privoxy -- I recommend Noscript, Adblock,
and Adblock Filterset.G.

3) I've seen some funny behavior from its caching. But Privoxy also
gives funny behavior. And Polipo breaks fewer sites than Privoxy does. :)

4) It crashes (albeit rarely). The developer knows and is looking for
more clues.

5) I've never tried it on Windows. I don't think its developer has either.

6) Polipo writes your hostname in every request. Either define proxyName
to something else, or set DisableVia = true in your config file.

See also http://article.gmane.org/gmane.comp.web.polipo.user/1016

Thanks,
--Roger