Re: [squid-users] FTP Browsing

The browser operates very differently when using a proxy or not.

When using a proxy HTTP is used between the browser and the proxy, and
it is the responsibility of the proxy to know how to fetch the requested
URL types. So you have Browser -> HTTP -> Squid -> FTP -> FTP server

When not using a proxy the browser much internally implement all
protocols it is capable of retreiving. Browser -> FTP -> FTP Server.

Squid is a HTTP proxy and as such expects all clients to speak HTTP when
talking to the proxy. Squid knows how to retrieve http://, ftp://,
gopher:// and wais:// URL types.

Regards
Henrik

Andrew Bramble wrote:
>
> Modern browsers seem to have no trouble browsing ftp sites when not
> travelling through the proxy. This leads me to presume that browsers
> already perform a similar function to squid when looking at an ftp site.
> But thinking on that , would not the browser be communicating with both
> ftp ports (20,21) ? Which would explain to me why squid is not suited to
> proxy in this way.
>
> Could this be achieved with an acl and a redirector ?
>
> acl ftpsite url_regex ^ftp://
> redirector_access allow ftpsite
> redirect_program /usr/local/squid/bin/ftpproxy
>
> But what if a different redirector need to be used ( ie adzapper) ?
>
> Thankyou everyone for you help.
>
> AB
>
> Henrik Nordstrom wrote:
>
> >
> > How do you want to have them returned?
> >
> > The browser expects a HTTP reply from Squid in a format that makes
> > sense. The only format browsers know to parse into links the user can
> > click on to receive the actual files is HTML.
> >
> > Sure, Squid could return the FTP directory listing as text/plain, but
> > then you would need to manually enter the URL for each file you want to
> > visit as the browser will then just display the directory listing as
> > text/plain as if it was a text file.
> >
> > Regards
> > Henrik
Received on Wed Nov 27 2002 - 04:01:38 MST