On Fri, Nov 19, 2010 at 10:26:37PM +0100, David Kalnischkies wrote: > On Fri, Nov 19, 2010 at 21:06, Bill Allombert > <Bill.Allombert@math.u-bordeaux1.fr> wrote: > > On Wed, Nov 17, 2010 at 03:44:43PM +0100, Holger Levsen wrote: > >> if you intend to reply to this subthread, please use the 587279 bug. > > > >> On Mittwoch, 17. November 2010, Bill Allombert wrote: > >> > I do not think it is correct to ever upgrade a free package to a non-free > >> > one. Now, apt is not at fault, the problem rather lie in a strange > >> > interpretation of policy 2.1.2 by some developers. But we cannot ignore the > >> 2.2.1 > >> > issue either. > >> > >> No. The "problem" lies in people adding non-free and contrib to their sources. > > > > I disagree. I put non-free in my source.list so that 'apt-cache show' displays the > > non-free packages, not to get any of them installed. This is important for reporting > > bugs against non-free packages, and not breaking them inadvertently. > > You are free to pin the source to -1 by default and only promote packages > you like to a higher pin or vise versa, pin the specific packages you don't > like down to -1 (= APT doesn't allow this version to be a candidate) > > >> So I think apt is actually right in those cases to upgrade to a non-free > >> alternative. It's the users choice. > > > > There are a variety of licenses in non-free and a user (or their lawyers) can be > > fine with some of them but not all. The choice of non-free packages installed > > should remain with the users. > > Now apt is just a tool and I do not ask apt to be aware of non-free. However the > > change in apt make the non-respect of policy 2.2.1 more problematic. > > I still fail to see why. How can it be more problematic to install alternative > B if (and only if) alternative A is not installable? I don't think that a user > expects that APT ignores or-groups and just always only works with the first > package in the or-group and fails if it doesn't work out with it. It does work > for ages if you install the package - so why must the situation be different > if the package is upgraded? Please give an example why - or at least get your > terms straight, as its problematic to follow an "upgrade a free package to a > non-free" argument as it doesn't make sense: A single package is in a specific > version either free or non-free, if it changes his freeness-status between > different versions is a completly different "problem"… > You seem to want to make the point that a free-dependency shouldn't be > replaced by APT with a non-free-dependency and my answer is that it will > not as long as the free-dependency can be used - in case the or-group is > free | non-free, of course. Your turn. Hmm, what about this, admittedly slightly contrived, but still possible case: 1. A package, at installation time, depends on free1 | free2 | non-free 2. With time, free1 either stops to provide the needed functionality or is simply removed; the next version of the package now depends on free2 | non-free 3. At this point, a user tries to upgrade the original package, but on her system there is a package that conflicts with free2... the only choice for APT now is to install the non-free package. G'luck, Peter -- Peter Pentchev roam@space.bg roam@ringlet.net roam@FreeBSD.org PGP key: http://people.FreeBSD.org/~roam/roam.key.asc Key fingerprint FDBA FD79 C26F 3C51 C95E DF9E ED18 B68D 1619 4553 because I didn't think of a good beginning of it.
Attachment:
signature.asc
Description: Digital signature