On Monday, May 02, 2011 07:31:31 AM Lucas Nussbaum wrote: ... > How we deal with freezes is the hard point in this discussion. I'm > personnally in favor of the "freeze rolling for 3 months, then fork > frozen and unfreeze rolling" plan, though it has some problems too > (it is not clear whether the required manpower really decreases at > the end of freezes). ...
There is a ton of complexity hidden under these simple words. There is also (that I can immediately think of): - How do we provide a reliable path for fixes to Testing once Unstable/Rolling have moved on? - How do we stitch Testing/Rolling back together after a release into the new Rolling? - How do we allow for more parallel transitions so that rolling can actually roll. The first two points have gotten a lot of discussion. The third one, not so much. The Debian archive has gotten large enough with enough non-trivial intersections between groups of packages that transitions of almost any size need coordination and analysis to find an appropriate time to land in order to minimize deadlocks in Unstable -> Testing (or whatever you call it) transitions. If you view this exercise as primarily a PR move to make Testing seem more attractive to users that want a rolling distribution, then I suggest we arrange things so it can actually roll. As an example, my desktop environment of choice (KDE) is still a year (and two major releases) out of date in Debian Unstable/Testing. Current packages exist in Experimental, but can't get to Unstable let alone some theoretical Rolling because there's no transition window. I don't think that someone who is attracted to the idea of a Rolling release to get the latest and greatest would find this met their expectations. Without solving the problem of the need to serialize transitions, I doubt Rolling will match the expectations such a change would engender and while there would no doubt be publicity, I'm skeptical it would be the good kind. Scott K -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201105021148.28250.deb...@kitterman.com
