Date: Thu, 12 Aug 2004 09:06:26 -0700 (PDT) From: Doug Barton <DougB@FreeBSD.org> To: Will Andrews <will@csociety.org> Cc: Oliver Eikemeier <eikemeier@fillmore-labs.com> Subject: Re: cvs commit: src/usr.sbin/pkg_install/add Makefile add.hextract.c Makefilesrc/usr.sbin/pkg_install/delete Makefile delete.h main.c perform.c ... Message-ID: <20040812085554.M773@ync.qbhto.arg> In-Reply-To: <20040812154941.GJ32036@sirius.firepipe.net> References: <20040812012909.GA25768@electra.cse.Buffalo.EDU> <2CD52765-EC03-11D8-887A-00039312D914@fillmore-labs.com> <1092287083.796.29.camel@tomcat.kitchenlab.org> <20040812154941.GJ32036@sirius.firepipe.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 12 Aug 2004, Will Andrews wrote:
> On Wed, Aug 11, 2004 at 10:31:07PM -0700, Alfred Perlstein wrote:
>> But wouldn't it suck to be that user that actually does use '-c'
>> correctly, and to have it suddenly gone in the middle of a "stable"
>> release?
>
> It is a broken mis-feature. Please tell me how you managed to
> make it work properly. Do you have some sort of script that
> sorts the output in order by dependency tree? Or do you actually
> go through it every time and sort it by hand? Do you think that
> it's justified that so many other users who didn't know what they
> were doing shot themselves in the foot?
I find this line of reasoning very interesting in light of the
disagreement I'm currently having with eik about repo copying a port I'm
working on. On the one hand, you and he are arguing that it's perfectly
ok to break POLA in -stable because the new stuff is better, and the old
stuff sucked anyway. On the other hand, eik and one other member of the
portmgr team are arguing that an old port's revision history is so
incredibly valuable that it must be preserved, even though it has
little relevance to the new port. I'd find this funny if it weren't so
sad.
This is exactly the opposite of what it should be. In the past, the very
definition of a -stable branch included that features were NEVER
removed. It doesn't matter how much YOU as an individual developer don't
like a feature, you have absolutely no way of knowing how many users
depend on it, how they are using it, etc.
On the other hand, repo copying a port has always been a judgement call,
based on the merits of the individual case, repo bloat, etc., and now
we're being told that this rule must remain inviolate. I think that this
bodes very poorly for the project.
Doug
--
This .signature sanitized for your protection
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040812085554.M773>