I know the idea of having multiple trees seems complicated, but it isn't
nearly as complicated as it sounds. The time required to keep things in sync
is reduced to a few hours by only allowing syncs from the top down. On the
one day/month that sync's happen, CVS2 & CVS3 would be completely
overwritten by CVS1. The fact that Savannah already has a patch section
makes the implimentation even easier. In truth, I wish there was a way to
make branches of a CVS tree writable only to a selected group, but I don't
know of any such functionality in current CVS. If you've got a better idea
that allows the world to write to the tree without the risk of someone
committing something heinously wrong and causing many hours of work to be
destroyed by a rollback to fix the mistake, I'm game to listen. Otherwise,
try giving the idea more than a cursory inspection. I know from experience
that such a system is no where near as complicated as it sounds.

R. King
----- Original Message -----
From: "Bryce Denney" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, March 01, 2002 1:37 PM
Subject: Re: [plex86] plex86 status


> On Fri, 1 Mar 2002, Ranando King wrote:
> > Drew having to pick up the load of lead in this project. I've got a
> > suggestion about the CVS situation that might just get things moving
faster.
> > If there were 3 CVS code archives and 1 CVS patch archive, then everyone
> > should be able to work without fear of destroying something vital or
worse
> > yet, suffering a severe rollback on rejected work. Here's the idea:
> > [snip]
>
> Ranando,
>
> This is way too complicated to be practical.  It takes a lot of time to
> keep things in sync, and to decide which changes from one place should be
> accepted into some other place.  Having multiple cvs archives with
> basically the same code defeats the purpose.
>
> I agree that there should be a public place where everybody can submit
> their patches.  The Savannah site has a patch section that is intended for
> this sort of thing, and there are already a few patches there.
> (http://savannah.gnu.org/patch/?group_id=869)  If the patch is acceptable,
> a developer with cvs write access commits the patch.
>
> Keep it simple! ;)
>
> -Bryce
>
>
>


Reply via email to