Re: proxy use of jigsaw

I've been using Squid, both as a local cache for my browser, and as an "http
accelerator" for Jigsaw. I find Squid to be faster and more accurate than
Netscape's internal caching. To use Squid as an "http accelerator", you run it
on the port your server was running originally, and move your server to an
alternate port which you tell Squid about. Squid then acts as if it were your
original server, caching responses from your server and serving documents from
its cache until they expire and it needs to re-request them. This improves
the performance of Jigsaw for me, since I can reduce the load on it while
retaining all of its dynamism.

Squid emits logs in various formats, including the common log format, which
means I can use it as a record of everywhere I've visited.

Another cool thing about Squid is ICP, the internet cache protocol, which
allows a proxy server to participate in a hierarchy of proxy servers which
can share each other's caches. Jigsaw contains an ICPFilter in
w3c.www.protocol.http.icp, but I don't know how to use it :(, is anyone
doing this?

Squid is at
 http://squid.nlanr.net/Squid

This isn't supposed to be an ad! Squid is freeware under the GPL. Oh yeah,
it's UNIX-only.

> Javier Arellano writes:
>  > 
>  > I'm attempting to "monitor" browser interaction by having
>  > one's browser set to a proxy web server (jigsaw) and setting up a 
>  > filter on this server to "track" requests/replies
>  > 
>  > Now, my question is: given that I utilizing jigsaw as a proxy server,
>  > if one requests a page (i.e., URL) that has been previously visited by the 
>  > browser in the same session, will jigsaw "see" this request even though it
>  > may be cached in the proxied browser...the key point is that the browser
>  > is proxied to jigsaw, i.e., does it make a difference if it is proxied or 
>  > not.
> 
> That's a neat idea. In fact I would really like to have a filter that
> would keep track of *everything* I have visited (very usefull, you
> often loose track of what you've been up to these days ;-). That
> filter should emit all the visited urls (and time) into some HTML file
> which would be serve by the proxy (typically one file per day, zip
> them every week, keep N weeks of log, etc).
> A friend of mine have written that stuff in C (works by checking the X
> properties of Netscape - only under UNIX). It would be much more nice
> to have a personnal proxy, etc
> 
> Anyway back to the question: this depends on the ordering of the
> filter decl;aration in the relevant property
> (w3c.www.protocol.http.filters), so if you say:
> 
> w3c.www.protocol.http.filters=cache|log
> 
> then log might not get called (if request is fullfilled by the cache),
> however:
> 
> w3c.www.protocol.http.filters=log|cache
> 
> Things will work great :-)
> 
> Anselm.
> btw let me know if you write that and if you're willing to share the
> code. 
> 
> 


-- 
Joe Futrelle
Developer, Joule/Jigsaw Java/HTTP
National Center for Supercomputing Applications
futrelle@ncsa.uiuc.edu
(217) 265-0296

Received on Friday, 30 May 1997 12:53:57 UTC