null directives

["Roy T. Fielding" <fi...@kiwi.ICS.UCI.EDU>]

Is it worth defining the core directives that have been removed
from 2.0 (like ResourceConfig and AccessConfig) and assigning
them to a null routine, just so that 1.3 config files work?
Or is it better to have the error message, so that the user
will know that we now only allow one config file?

....Roy

Re: null directives

[Rodent of Unusual Size <Ke...@Golux.Com>]

Manoj Kasichainula wrote:
> 
> If we were to make those two directives backward compatible, I'd alias
> them to "Include" instead (with a warning in the logs), so they'd at
> least work.

That wouldn't have the same effect; aliasing them to Include would
suck in the contents at that point in the file, rather than at
the end as in 1.3.  The original directives were declarative only,
and just identified which files to process *after* httpd.conf.

If we're not going to look at the files, it should be at least
an error-level message and keep the server from starting -- because
any security stuff in the canonical access.conf won't be getting
processed.  This should be equivalent to what we do if we can't
read an .htacces file.
-- 
#ken    P-)}

Ken Coar                    <http://Golux.Com/coar/>
Apache Software Foundation  <http://www.apache.org/>
"Apache Server for Dummies" <http://Apache-Server.Com/>
"Apache Server Unleashed"   <http://ApacheUnleashed.Com/>

Re: null directives

[Manoj Kasichainula <ma...@io.com>]

On Fri, Mar 24, 2000 at 11:54:10PM -0800, Roy T. Fielding wrote:
> Is it worth defining the core directives that have been removed
> from 2.0 (like ResourceConfig and AccessConfig) and assigning
> them to a null routine, just so that 1.3 config files work?
> Or is it better to have the error message, so that the user
> will know that we now only allow one config file?

Well, that partly depends on whether the 2.0 config file ends up
completely different or not. :) Assuming not...

If we were to make those two directives backward compatible, I'd alias
them to "Include" instead (with a warning in the logs), so they'd at
least work.

But, unless the server is also running in the same process model
(prefork), we'd probably want to force the admin to hack up the config
files anyway.