Message-ID: <37A75B1E.2F22341E@raleigh.ibm.com>
Date: Tue, 03 Aug 1999 17:11:58 -0400
From: Bill Stoddard <stoddard@raleigh.ibm.com>
MIME-Version: 1.0
To: new-httpd@apache.org
Subject: Re: Where to start the Apache 2.0 code base
References: <9908031318.aa14142@paris.ics.uci.edu>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: new-httpd-owner@apache.org
Precedence: bulk
Reply-To: new-httpd@apache.org
Status: O


> >I'll attempt to restate the essence of what I think you are advocating.
> >Correct me if I miss the point...
> >1. Copy Apache 1.3.8 into the top level Apache 2.0 repository
> >2. Tag the Apache 1.3.8 repository to identify the point the copy was
> >made. Apache 1.3.8 will no doubt be tweaked a few more times, perhaps to
> >1.3.9 or whatever...
> >3. Reversion all the Apache 2.0 files to 1.1
> 
> That would be goofy.  
I knew you didn't intend to save the 1.3.8 history in the 2.0
repository. I wasn't clear on that point.

> This is what I proposed, listed in parallel so
> that you can see that both can occur in parallel:
> 
>         mpm repository            scratch repository
>         =====================     ========================
>    1.   merge in all work      1. cvs export 1.3.8
>    2.   move files to final    2. move files so that the old files
>         locations for dists       correspond to their final loc in mpm
>    3.   continue hacking       3. cvs add all to form rev 1.1 == 1.3.8
>    4.   tag as merge point     4. tag as APACHE_1_3_8_moved
>    5.   continue hacking       5. diff -r APACHE_1_3_8_moved mpm_merge_point
>    6.   identify diffs from scratch as change sets
And this is why I am saying if we are going to base off of Apache 1.3.8,
we need to set up the repository NOW and not spend more time working on
apache-mpm. This is already a huge diff with just HOOKS and MPM. If we
add APR to apache-mpm, the diff will double in size. If we put APR
directly into a clean repository, then we save the effort of sifting
through a huge diff to pull it back out of Apache-mpm. 

>                                7. merge diffs as complete change sets
>    8.   move mpm to archive    8. merge changes after mpm merge point
>                                9. move scratch rcs files to 2.0 directory

I'd like to get group concensus on which path we are going to take to
build the apache-2.0 repository; Roy's plan or base directly off of
apache-mpm. I see the value in adding the change sets to an 1.3.8 base
(to maintain history and for code review purposes). Step 6 will be
painful, but if Roy's plan is the decision, I'd like to get our team
working on it and I don't want to waste time putting APR into
apache-mpm.

-- 
Bill Stoddard
stoddard@raleigh.ibm.com