Date: Mon, 11 Oct 1999 14:39:31 -0400 (EDT)
From: Ryan Bloom <rbb@raleigh.ibm.com>
To: new-httpd@apache.org
Subject: Re: Compiling multiple MPM at the same time.
In-Reply-To: <199910111803.OAA09986@devsys.jaguNET.com>
Message-ID: 
 <Pine.LNX.4.04.9910111432460.17868-100000@chosondo.raleigh.ibm.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: new-httpd-owner@apache.org
Precedence: bulk
Reply-To: new-httpd@apache.org

> If runtime then does that imply that each MPM method must compile
> for each platform? If so, then we'll need to wrap each method
> in some sort of #ifdef MPM_PTHREAD_DOES_NOT_WORK ON ME to allow
> a full compile to succeed and a bail out method when a non-compatible
> MPM method is selected at run time.

I don't think so.  I think we just need to expect to have smart users.
This would be an advanced option, that most people wouldn't need.  Most
MPM's won't even compile on platforms they weren't designed for.  For
example, the winNT MPM isn't ever going to compile on Unix, and non of the
Unix MPM's will compile on NT, because they all rely on fork().

We could define certain MPM's that won't compile on certain certain
platforms, so if I user tried to compile winNT MPM on AIX or Linux, it
would just die saying "No way, buddy".

Then, if they tried to load a non-built MPM at run time, everything would
close with an error condition anyway.

Ryan

_______________________________________________________________________
Ryan Bloom		rbb@raleigh.ibm.com
4205 S Miami Blvd	
RTP, NC 27709		It's a beautiful sight to see good dancers 
			doing simple steps.  It's a painful sight to
			see beginners doing complicated patterns.