[Subject Prev][Subject Next][Thread Prev][Thread Next][Subject Index][Thread Index]
Re: HTTP tunnel
Thats what I am up to. But just disagree with the terms. I would not call
it a HTTP server as it itself does not handle any GET, PUT etc requests it just
forwards them. Similarly I would not call it a proxy server as it does not
handle the request but just forwards them...
Anyway, just the terms used are different, and thats not a big deal as long
as the whole thing works fine.
>see it is this way... for the browser you tunnel is the proxy server..
>(only you know otherwise)... so after you've got the request from the
>browser, you can do some preliminary checking based on the url.. it
>nothing is amiss you pass the url to to squid at its port.. but for this
>your tunnel will have to behave as the HTTP client (browser) so youl
>will have to pass the exact HTTP request to squid... after that let
>squid do its work... when it gets the data back, it will give it to your
>http tunnel program (it will always give it to whosoever gave the
>request - here in this case it will always be you tunnel).. you program
>can then take the data, tear it apart and hunt for whatever you want
>to... if nothing suspective is found your tunnel acts a proxy and
>forwards the data to the client.... (basically, an HTTP proxy is also a
>sort of HTTP server... so your tunnel will also have to be one - nothing
>much - just minimal...) ... so simply put... your client browser got
>what it wanted... ok.. ? Now if you did find something wrong in the html
>content, then you should simply log the url and ip of the client and
>send a denied message...