Subject: Re: CVS commit: pkgsrc/textproc/dict-client
To: Thomas Klausner <wiz@NetBSD.org>
From: Jim Wise <email@example.com>
Date: 03/19/2005 11:27:37
-----BEGIN PGP SIGNED MESSAGE-----
On Sat, 19 Mar 2005, Thomas Klausner wrote:
>On Fri, Mar 18, 2005 at 09:34:03PM -0500, Jim Wise wrote:
>> It is visible to a user setting up the package for the first time (as it
>> affects the default behavior, and thus reduces the chance that you'll
>> need to set up a custom config file). It is invisible to users
>> upgrading -- the users who are served by bumping PKGREVISION for changes
>> which do affect them, and badly disserved by bumping PKGREVISION -- as
>> you just did -- for invisible changes, thus spamming their pkg_chk
>I think you're forgetting the users who install a package for the
>first time. If they install the binary package, and there is no
>PKGREVISION bump, how are they supposed to know how much they have
>to configure it?
>It is a new version because it has different features from the
>previous one, and we usually try to keep these apart.
If we provided documentation anywhere saying ``here's how you have to
configure this package, but it changed after PKG_REVISION n'', this
would make sense. But without doing that work, we don't get this
benefit -- and we certainly _do_ get the harm that pkg_chk has become
nearly useless these days, due to constant PKG_REVISION bumps for
invisible or barely-visible changes.
>To get back to your original question from the first mail:
>> I really don't think a change in a packages default installed
>> configuration qualifies.
>> Do you?
>Yes, I do. I guess this is where the main difference lies between
>our points of view. :)
Indeed. I'll now leave it up to others to weigh the pros and cons of
each approach and reach a final policy. :-)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (NetBSD)
-----END PGP SIGNATURE-----