On Thu, 20 Apr 1995, Rob Hartill wrote:
> > It's alright with me if our beta has the current behavior, and I would
> > also support extending the log file format if we provided an efficient
> > support/apache2common script.
> >
> > How does this sound: an "ExtendedLogging" srm.conf:
> >
> > ExtendedLogging RealObject Referrer UserAgent
> >
> > Which results in a log file that looks like
> >
> > CLF RealObject Referrer UserAgent
> >
> > Where:
> >
> > RealObject = the real object served, if different from the file
> > requested (i.e., "/401.html" or "/cgi-bin/404-handler"), "-"
> > otherwise. In quotes.
>
> Does this handle the case where multiple redirects take place ?
>
> GET A -> B -> C -> D -> E
>
> ...crazy but perfectly valid I think.
You are crazy. Maybe that's why they elected you into the WWW Hall of
Fame. I wonder what it'll take to get me in.
I thought the server was smart enough to only redirect once? Isn't
redirecting more than once really a configuration error, when it has to
do with custom error messages?
> > Referrer = HTTP_REFERRER, in quotes.
> >
> > UserAgent = HTTP_USER_AGENT, in quotes.
>
>
> If people recorded USER_AGENT in the logfile, we'd need another
> conf option,
>
> WhatToDoWhenLogFileExceeds 80Mb rm access_log
>
> :-) does anyone in their right mind want to log USER_AGENT in their
> logfile. It only makes sense for big sites, and they simply won't be
> able to afford all the space it takes up in the log.
> USER AGENT is a big string for most browsers.
Yes, which is why it'd be a configuration option people could turn on or
off whenever they wanted. In fact if those options could be allows in
srm.conf, it could be turned on or off on a per-directory basis.
Brian
--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com brian@hyperreal.com
http://www.[hyperreal,organic].com/