Re: [proxyTools-users] Sorted Out CGI Ambles... a bit.....

[wayne <wa...@ny...>]

> From: Michael Foord <Mic...@tb...>
> To: "pro...@li..." <pro...@li...>
> Subject: [proxyTools-users] Sorted Out CGI Ambles... a bit.....
> 
> Ok - now my cgiambles.txt looks like this
> 
> #Format of entries: firewall, preamble, postamble
> #localProxy will request $preamble.escaped($url).$postamble
> #Use the first line that matches our firewall, or use '*' (all
> firewalls).
> #*, https://vip.megaproxy.com/go/,
> #*, http://web.archive.org/,
> #*, http://216.239.35.120/translate_c?&langpair=en%7Cen&u=,
> #XXX - note the hard-coded spanish to english conversion here.
> #XXX - This gets english pages,but some english words look like
> #XXX - spanish and get translated (e.g. 'can' changes to 'dog' :-)
> #*, http://babelfish.altavista.com/urltrurl?url=, &lp=es_en
> #XXX - We should probably check
> #XXX - the user's system to see if he's probably looking for
> #XXX - pages in the chosen lang here, and choose some other conversion!
> #Use Japanese (ja) (or chinese (zn)), low chance of accidental
> collisions
> #Note: translation ends at this: <blink>*** TRANSLATION ENDS HERE
> ***</blink>
> #*, http://babelfish.altavista.com/urltrurl?url=, &lp=ja_en
> *, http://******.com/cgiproxy/nph-proxy.pl/000100A/,
> 
> 
> 
> Which `works` - still no images from the looks of things - could just be
> our system fluffing it.

Hmm ... surprised it works. LP2 had no code to remove the leading 
http:// from the urls it gets from the browser. Now it does - hope 
it still works :-)

> Log :
> 
> For web browser proxy service, please configure your web browser
>   http proxy entry to localhost:10080 and the
>   https proxy entry to localhost:10076 now
> 
> Only accepting connections from 127.0.0.1, 192.168.4.67
> Executed getConfig(), result (truncated): <perldata>
>  <hash>
>   <item key="10076">
>    <hash>
> ...
> Original request: GET http://www.clanbob.net/ HTTP/1.0
> Modified request: GET http://207.44.136.28/ HTTP/1.0
> access denied or failed from 10080.2(a).0(192.168.4.1:8080).-(-):
> HTTP/1.1 403

Looks like the proxy catches commStrat 2(a). You might as well 
disable it in your config (or temporarily in commStrats.xml). 
That will divide the number of options LP tries by a factor of 4.

> 10080.2.0.-: 11:10 (8.7(80.2)) (6.5) 11(=0.544/0.05)
> Original request: GET http://www.clanbob.net/ HTTP/1.0
> Modified request: GET
> http://*******.com/cgiproxy/nph-proxy.pl/000100A/http://www.clanbob.net/
> HTTP/1.0
> Original request: GET
> http://216.239.51.102/search?client=navclient-auto&failedi
> p=216.239.37.102;2018439&ch=53222229727&freshness_check=3K_9hdhQoW2rDYvATf_sP&fe
> 
> atures=Rank&q=info:http%3A%2F%2Fwww%2Eclanbob%2Enet%2F HTTP/1.0

You've got some sort of automatic Google search set up in your browser?
The request to your cgi proxy failed, probably because of the extra 
http://. Edit the file and use the new lp2 as above to fix.
Untested - let me know.

> access denied or failed from 10080.2(7).0(192.168.4.1:8080).-(-):
> HTTP/1.1 403

This looks like your proxy is denying access to Google searches!

> 10080.2.0.-: 22:20 (14(80.2)) (8.2) 1.1e+003(=0.544/0.0005)
> 10080.2.0.-: 5.4:4.9 (3.8(80.2)) (2.3) 5.3(=88.4/16.6)
> 
> 
> 
> Note
> 
> Modified request: GET
> http://*******.com/cgiproxy/nph-proxy.pl/000100A/http://www.clanbob.net/
> HTTP/1.0
> 
> 
> Ought to read
> GET
> http://*******.com/cgiproxy/nph-proxy.pl/000100A/http/www.clanbob.net/
> 
> How do I `escape` the preamble so that the url reads correctly (minus
> the :/ ) for the James Marshall cgi.

See above.

> As it appears to work and the missing images may be down to my system
> perhaps this is unnecessary ??

Mmm ... that version of Junkbuster is newer than mine (I'm certain!).
Mine fails with these urls with the http://.
So, yeah. I can believe yours works with it, but there is that 
failure above ...

> Also the cgi proxy modifies links so following any links from the page
> showsthe proxy address  in the IE window (which you predicted) - but it
> looks from the log like localproxy figures this out and doesn`t chain
> cgi proxy requests...
> 
> Original request: GET http://******.com/cgiproxy/nph-proxy.pl/00010
> 0A/http/www.clanbob.net/modules.php?name=Content&pa=showpage&pid=4
> HTTP/1.0
> Original request: GET
> http://216.239.51.102/search?client=navclient-auto&failedi
> p=216.239.51.102;266253&ch=51821214745&features=Rank&q=info:http%3A%2F%2F******%2E******%2Ecom%2Fcgiproxy%2Fnph%2Dproxy%2Epl%2F000100A%2Fhttp%2Fwww%2Eclan
> 
> bob%2Enet%2Fmodules%2Ephp%3F HTTP/1.0
> access denied or failed from 10080.2(a7).0(192.168.4.1:8080).-(-):
> HTTP/1.1 403
> 10080.2.0.-: 5.4:4.7 (3.1(80.2)) (1.4) 1.1e+003(=0.544/0.0005)
> 10080.2.0.-: 4.6:4 (3(80.2)) (1.5) 3.1(=25.7/8.24)
> 
> I  think ?? - it worked anyway.

It only checks to see if the CGI has rewritten the links in any new 
request it receives from the browser. If it sees that, it doesn't add 
the preamble again. It never removes that preamble as the page comes 
in, so if commStrat 2 uses the proxy once, and you follow links from 
there, you are always using the cgi proxy.

What you see in the log snippet above is something to do with a 
strange 'auto-search' option you've set up in the browser, I think.

> Is the GET http://216.239.51.102/search?client=navclient-auto&failedi my
> IE still automatically launching a search after a  failed request - if
> so its probably the google toolbar and I need to try and kill that if I
> can.

IE has some option like this too, I think. 

> Anyway - many thanks Wayne
> 
> IT man has told us hes just taking superscout down for an upgrade - so
> who knows what the system will look like when that`s finished !!!!
> :-)

Good. 
All this is making localProxy so damn smart I will soon be redundant!

> Mike

--
wa...@ny...
http://proxytools.sourceforge.net/