On Fri, Jun 28, 2013 at 10:44:04PM +0200, Stefano Zacchiroli wrote: > On Thu, Jun 27, 2013 at 04:01:21PM +0100, Ian Jackson wrote: > > 8. The Technical Committee resolves that alternative dependencies of > > the form "Depends: package-in-main | package-in-non-free" > > constitute a non-release-critical violation of the policy > > clause cited in point 1. > > > > 9. When it is necessary to provide a reference in a Depends or > > Recommends from main to non-free, this should be done via a > > neutrally named virtual package. > > > > 10. The Technical Committee requests that the policy editors make > > an appropriate clarification to the policy documents. > > About point (9.), and considering the past discussion on the matter > we've had with Colin [1], I suggest to expand it as follows: > > 9. When it is necessary to provide a reference in a Depends or > Recommends from main to non-free, this should be done via a > neutrally named virtual package. When depending on such a virtual > package, other packages should specify a real package in main as > the first alternative, e.g. > "Depends: package-in-main | virtual-interface".
Yes, this is the main comment I have when reconciling Ian's version with my option B. I suspect - without having checked - that what happened is not so much that Ian intentionally dropped this, but that he took a part of my original text without the amendments I made in response to your previous comments [1]. I've taken the liberty of restoring similar text [2]. [1] http://anonscm.debian.org/gitweb/?p=collab-maint/debian-ctte.git;a=commitdiff;h=a4bb169a96c2a25f0b01b550179256e650b8f3fd [2] http://anonscm.debian.org/gitweb/?p=collab-maint/debian-ctte.git;a=commitdiff;h=923daf9aaa689e924f3cbd136fa1090b2db6fef7 > This might be paranoia on my side, but I'm really worried that, by not > following the above best practice, we risk losing some explicit > preferences for main packages that are currently expressed in the > archive and effective during dependency resolution. I do realize that > the most appropriate place where to document this would be the Policy, > but I doubt it would hurt to mention it in the tech-ctte resolution; > quite the contrary, IMHO. I concur with this. -- Colin Watson [cjwat...@debian.org] -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
