Subject: Re: Consulting MAINTAINER before updating a package
To: None <>
From: Tobias Nygren <>
List: tech-pkg
Date: 06/19/2007 02:00:15
On Tue, 19 Jun 2007 01:11:11 +0200
Roland Illig <> wrote:

> Greg Troxel wrote:
> > "Julio M. Merino Vidal" <> writes:
> > 
> > 
> >>
> >>Why does this need to be a *variable*?  Will it be of any use,
> >>anywhere in the pkgsrc infrastructure?  If not it should be a comment
> >>as Johnny (IIRC) mentioned.
> > 
> > 
> > I don't know; probably not. (How is MAINTAINER used?  If displayed on a
> > web page, this is almost as useful.)  My thought is that if we are going
> > to agree on semantics and a have a limited set of codepoints, then the
> > information should be expressed in machine-parseable form to reduce
> > confusion and unneeded variability.  A variable lets pkglint detect
> > values not in the enum, but I suppose slows down make.  A commented
> > variable assignment seems fine in order to get the rigid semantics
> > without the
> > 
> pkglint can not only check variable assignments, it also knows about 
> comments and much other stuff, so this is a lame excuse. :)
> Just tell me what pkglint should check or add the code yourself. By the 
> way, all my maintained packages have MAINTAINERSTYLE=weak.
> Roland

Now that there seems to be a consensus that it is a good idea
to formalise this, and that exclusive maintainership can be bad
for many noncritical/leaf packages because it results in
unnecessary e-mail overhead and delays, how about this scheme:

1) "weak" maintainership declared the usual way as:



2) exclusive or special maintainership declared as:


# Some comment here. It's a free-form field.

3) "abandoned" maintainership declared as:


The absence of the comment above MAINTAINER would mean that updating
without asking is allowed. Anyone can step up and steal weak
maintainership of an abandoned package. If people care strongly about
what's going on with their packages they should either

1) read pkgsrc-changes@
2) Add a comment field to their packages

This scheme is good because:

 o most packages can be left untouched
 o the comment field is flexible
 o promotes collaboration
 o many packages currently have MAINTAINERs who are not NetBSD
   developers. They would formally be recognised as weak maintainers.

Kind regards,