On Thu, Dec 06, 2012 at 09:35:11AM +0000, Colin Watson wrote: > Right. I had this in the back of my mind as something I didn't need to > spell out because policy already discusses real alternatives (7.5), but > on reflection the wording there is much weaker than I remember and in > any case you're correct that this is a disparity between the two ballot > options.
Thanks for taking my remark into account! I think we're in agreement already , but to stress my main point: - Policy §7.5 explains *how* to provide a real package preference when depending on a virtual package, but does not say *when* to do that. (IIRC there is some provision to do that with build-time dependencies, but I haven't found it with a quick search.) > How about this, which I've committed to our git branch: > > B 6. Virtual packages are a suitable existing mechanism for packages to > B declare the set of abstract features they provide, and allow > B packages in main to depend on such abstract features without > B needing to name every (free or non-free) alternative. They should > B nevertheless name at least one free preferred alternative, so that > B the package management system has appropriate defaults. How about: s/at least one free preferred alternative/a default free alternative/ that would be more consistent with the wording of Policy §7.5 (not a big deal, though). > [...] > B 8. We recommend that affected packages consider the use of virtual > B packages instead. When doing so, they should specify a real > B package in main as the first alternative, e.g. "Depends: > B package-in-main | virtual-interface". Looks good. Possible s/first/default/ if you apply the suggestion above, for consistency. > I do find it more in line with our general principles for packages in > main to be preferred for dependency resolution, though, so I went for > "should". I'm still convinced that a non-free default alternative would not be appropriate. But before trying to argue this point, in an attempt to save us all some discussion time, let me try to side-step it :-) In the initial bug report that brought this issue before the tech-ctte, Russ wrote: > (I believe that the question of whether "foo-nonfree | foo" should be > allowed is not at issue and that the consensus is that it's not > permitted. However, the Technical Committee can certainly open that > discussion if desired.) So it seems that, at least for non-virtual packages, the Policy Team already has consensus that a non-free default alternative is not acceptable. I would find surprising if consensus would be different for virtual packages. Granted, this is my interpretation only, Russ (with his policy editor hat on) would likely know better. As noted, the tech-ctte is certainly free to dig into this specific sub-matter further. But if you're simply looking for a sane default, I think sticking with Policy Team consensus would be entirely appropriate. HTH, Cheers. -- Stefano Zacchiroli . . . . . . . z...@upsilon.cc . . . . o . . . o . o Maître de conférences . . . . . http://upsilon.cc/zack . . . o . . . o o Debian Project Leader . . . . . . @zack on identi.ca . . o o o . . . o . « the first rule of tautology club is the first rule of tautology club »
signature.asc
Description: Digital signature