On Mon, Aug 24, 2009 at 09:15:59AM +0200, Bjørn Mork wrote: > > I think the use case is users who are being control freaks about the set of > > packages on their systems. If the set of packages being pulled in as > > recommends is *wrong* (they don't fit the Policy definition of Recommends), > > bugs should be filed against those packages and be fixed. If the set of > > packages is *right*, then there's no good reason to give users a big "ignore > > Policy" button at install time. > The number of "yes, we want this" answers shows that the current set is > wrong. No, it doesn't; all it shows is that there are people who are willing to complain about installation of Recommends by default. > One example just out of my head: I've found that it is nearly impossible > to avoid avahi-daemon if APT::Install-Recommends is true. There are > multiple completely unrelated (IMHO) packages recommending it, like > sane-utils and rhythmbox. But are these bugs? I don't think the rhythmbox one is a bug. I'm not sure if the sane-utils one is a bug. > I guess a system without avahi-daemon is pretty unusual, given the number > of recommends pointing in that direction. That's not really a relevant measure of whether the Recommends are correct. > Should I start filing bugs anyway? Discussion on debian-devel first may be a better way to proceed. > I'm hesitating, and have chosen to set APT::Install-Recommends False > instead. I think that's self-evidently the wrong solution. If you don't want avahi-daemon installed, you can uninstall that one package or use equivs or do a variety of other things. It doesn't make sense to ignore all package Recommends just because there's one particular package you don't want on your system. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer http://www.debian.org/ slangasek@ubuntu.com vorlon@debian.org
Attachment:
signature.asc
Description: Digital signature