Subject: Re: CVSup collections for a NetBSD CVS tree
To: NetBSD-current Discussion List <current-users@netbsd.org>
From: Feico Dillema <dillema@acm.org>
List: current-users
Date: 05/04/1999 21:18:44
Your message dated: Tue, 04 May 1999 14:34:39 EDT
>[ On Tuesday, May 4, 1999 at 10:34:51 (-0700), Jonathan Stone wrote: ]

>The pressure is on those who want to use the CVSup client to do the
>porting themselves, or to find someone to do it for them.  If they do
>the latter and put pressure on you, and you don't just let it bounce
>off, then *that* is your problem.  You need to know how to manage
>people's expectations or they'll simply expect everything of you.
>
>Don't worry -- be happy!

This seems to me a nice point in the discussion for a small summary
and a suggestion for a pragmatic approach to the issues raised. I
see two issues remaining:

1. Partial support for CVSup may increase pressure on some portmasters.

I think this has been discussed at length and people seem to agree that
this is not a real issue as long as CVSup doesn't have an `official
supported by TNF' label on it. I think the misunderstanding is on what
`officially supported' exactly means. At least, that's not clear to me
either. So, couple of questions:

-Would cvsup.netbsd.org imply official support? 
-Could a disclaimer message (this service is unsupported by TNF) nuke the
official-support implication of that?
-Would a link on the NetBSD site to a `cvsup.not-netbsd.tld' imply official
support too?
-In other words; what makes a service officially supported by TNF?

2. CVSup is technically a flawed solution for NetBSD (in the long run
as officially supported solution).

This one still seems to stand. We have the following choices for the 
future to get a CVSup style service:

1. Get Modula-3 + CVSup ported to many or most NetBSD-supported archs.
2. Develop C-version of CVSup.
3. Develop something entirely new and better than CVSup.

All three choices take a considerable amount of work to accomplish. This means
that consenses on the approach to take would be valuable in order not to waste
a lot of effort. Getting some experience with CVSup may help in making the
right choice. My suggestion would be to set up a CVSup server either 
unofficially or `officially under probation' and evaluate how useful a service
it really is and  to evaluate whether to go for 1., 2. or 3. above.

Please tell me if I completely missed the point...

Feico.