httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ryan Bloom <>
Subject Re: [PATCH] Allow DocumentRoot within Location blocks
Date Mon, 07 Jan 2002 21:00:01 GMT
On Monday 07 January 2002 11:55 am, William A. Rowe, Jr. wrote:
> From: "Rodent of Unusual Size" <Ken.Coar@Golux.Com>
> Sent: Monday, January 07, 2002 1:47 PM
> > Ryan Bloom wrote:
> > > 
> > > Note that I said the module needed to be written correctly.
> > > Mod_status doesn't have a map_to_storage hook, which would
> > > be required to avoid the stats.
> > 
> > I'm in favour of a mechanism that doesn't require every single
> > generated-content module to be rewritten to include an essentially
> > duplicate function.

I'm in favor of doing this in the code and not forcing every administrator to
reconfigure their server to get this functionality.  Understand that introducing
the DocumentRoot directive inside a <Location> container doesn't remove the
stats automatically, it requires that every admin change their mod_status config

<Location /server-status>
    DocumentRoot unset
    SetHandler server-status

> Worse yet - how do we know that the handler is for mod_status,
> when the handler isn't set in translate names?
> SetHandler/Alias is really showing it's age.  JKMount works because
> the JKMount directive is handled in the translate_name hook.
> SetHandler can't accomplish this.

If the module is implemented correctly it can handle this.  The module can
determine if it is configured for the specific directory inside of the requested
directory, and it can act accordingly.

> Perhaps a LocationMountHandler {handlername} would work, as a 'generic'
> method.  But do we want to force each module to do so?
> I honestly believe that Unsetting the DocumentRoot is simpler.  It still
> might be overridden (causing 'extra' stats) because of a mod_rewrite or
> other directory mapping, but this makes life somewhat easier.

Ryan Bloom
Covalent Technologies

View raw message