Re: Req: Need a way to "wash" CGI parameters of URLs contained in cached

Miernik <[email protected]> writes:

> Andrew M. Bishop <amb-Uxr6IM1mbv2TY6FTCsQk+9Bc4/[email protected]> wrote:
> >> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=203882
> > Two things that I sometimes get asked for as new features in WWWOFFLE
> > are the ability to change URLs or to modify the contents of web pages
> > (both using a user supplied program).
> > 
> > In both cases these are not features that I have a very strong feeling
> > for adding to WWWOFFLE.  It would take a lot of CPU effort to pass
> > every URL or every page through an external script.  You have the
> > overhead of starting the scripting engine and people are going to want
> > to use perl or other fully featured script languages not simple shell
> > scripts for this.
> 
> I think this would be a very useful feature, saving a lot of bandwidth
> and disk space for some sites, but not implemented as an external
> script, but implemented internally in WWWOFFLE in C, in such a way that
> one could write in the config file:
> 
> [<URL-SPEC>] wash-cgi = <string>
> 
> which would remove any CGI parameter which has the <string> in
> it, for example:
> 
> http://www.simplus.pl/*   wash-cgi = 'PHPSESSID'
> http://allegro.pl/*       wash-cgi = 'st=gtext'
> http://cgi.ebay.com/*     wash-cgi = 'ssPageName'
> 
> I don't think such implementation would use too much CPU effort.
> Especially the ssPageName is very annoying on eBay, it has an argument
> with a number which changes with each e-mail notification about an item,
> washing it would save a lot of bandwith here.

This is a special case of the 'washing' of URLs that people sometimes
ask for.  You are only removing whole CGI arguments that match and not
operating on anything larger or smaller than a CGI argument.  There
would be some CPU time in taking apart the arguments and putting them
back together if modified, but it shouldn't be too much.


Have you tested these listed pages to see if they work when you remove
the 'washed' parameter?  Some sites will redirect you to a new session
ID if you don't already have one.

-- 
Andrew.
----------------------------------------------------------------------
Andrew M. Bishop                             amb-Uxr6IM1mbv2TY6FTCsQk+9Bc4/[email protected]
                                      http://www.gedanken.demon.co.uk/

WWWOFFLE users page:
        http://www.gedanken.demon.co.uk/wwwoffle/version-2.8/user.html