Subject: Re: slight API change
To: new-httpd@hyperreal.com
Date: Sat, 2 Mar 1996 13:32:39 +0000 (GMT)
From: Ben Laurie <ben@gonzo.ben.algroup.co.uk>
In-Reply-To: <199603021241.HAA10546@shado.jaguNET.com> from "Jim Jagielski" at
 Mar 2, 96 07:41:34 am
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 8bit
Content-Length: 3495      
Message-ID: <9603021332.aa29459@gonzo.ben.algroup.co.uk>
Sender: owner-new-httpd@apache.org
Precedence: bulk
Reply-To: new-httpd@hyperreal.com
Status: O

> 
> Alexei Kosut wrote:
> > 
> > But since this is actually a change to the API, I thought I'd pass it by 
> > the list first, see what people thought. I might have missed something, 
> > maybe it will mess things up, maybe it's just a bad idea, etc... so, 
> > before I commit it to CVS, maybe people could just take a look. Thanks. 
> > The patch is:
> 
> Okay... I'm really confused now... Now that we've switched over to CVS
> does that mean that anyone can add a patch to the present distribution with
> absolutely no vote or test to see how it affects anyone else?

It means that anyone who is on the commit list can add a patch to the present
_development_ version with no vote or test.

> 
> I read the above as "I have a patch that I'm gonna fold in to Apache ('commit
> to CVS') and even though I don't need to ask the group if that's OK, I'm
> gonna this time" Or am I reading into things.

That's right. However, we do expect people to consult the group before making
drastic changes. Repeated failure to do this will result in the removal of
commit privilege.

> 
> As cumbersome as the old method was, at least people were informed when
> there were patches available and were given the opportunity to test and
> vote on them before they were committed to the code; Have we stopped this?

We have stopped this for the development branch. As it stands, we intend to
continue to use the patch and vote system for release branches. When the
development branch becomes a release candidate, we will vote on _all_ changes
since the previous release. So, the voting system is not lost, it is merely
deferred.

> Is the only way to obtain news about new patches is by using CVS? Does
> one need CVS on their local machine to be a "member in good standing" (i.e.
> be able to obtain and submit patches)?

Subscribing to the cvs list will inform you of patches. When we get around to
implementing the n-hourly auto update and tar, you will be able to obtain the
patches by downloading the latest version. It had also occurred to me that
extracting the patches and sending them with the log message to the mailing
list might be useful.

Submitting patches can be done by the old method (except the diff should be
done against the most recent possible version), and one of the committers will
put the patch in.

> 
> I just have visions of people playing around with the main code with
> absolutely no controls over that and poor ignorant me, by not grabbing
> the new distribution everyday and recompiling and testing, finding out
> that something made the port on my OS cough and die. At least with the
> old method I could apply the patch one-at-a-time to see if it worked...

Hmmm ... if you are CVS equipped you can still apply one at a time, with
considerably more ease than with plain old patch. I'm not quite sure how we
would resolve this for a non-CVS user.

> 
> Or maybe I am totally confused about how we (and CVS) are doing this...
> -- 
> Jim Jagielski  << jim@jaguNET.com >>   |      "That's a Smith & Wesson,
>   **  jaguNET Access Services  **      |       and you've had your six" 
>       Email: info@jaguNET.com          |             - James Bond
> ++    http://www.jaguNET.com/         +++      Voice/Fax: 410-931-7060       ++

-- 
Ben Laurie                  Phone: +44 (181) 994 6435
Freelance Consultant and    Fax:   +44 (181) 994 6472
Technical Director          Email: ben@algroup.co.uk
A.L. Digital Ltd,           URL: http://www.algroup.co.uk
London, England.