On Wednesday 21 January 2004 06:13, foser wrote: > On the other hand it is just a fact that a lot of ~arch upgrading > happens trough stumbling over it (so a lot of packages stay longer than > needed in ~arch), there have been some efforts to attack this problem, > but not too successful. This is an issue that needs attention and ideas > on how to solve this in a satisfactory manner. Since you seem to Ideas? me has ideas :).
Seriously, I remember a discussion some time ago about starting a "bump to stable" days, a la bug crunching days we have now. The proposal was for doing it monthly, and I remember countering it with the argument about going biweekly, - to match the "official testing duration" we have in policy. This way all the devs will have a clear "bump day" when they devote an hour or so to the bumps and don't care about this for two more week to come afterwards :). The procedure can be very basic as well. Say "newer" bugs (processed within last two weeks) are marked as candidates (may be even virtually, by looking at the dates? Although given enough interest it may be usefull to have another bug resolution marking..) and former candidates are marked stable. All of course provided that there were no problems reported. In the latter case (problems) they are reopened. Sorry about a letter-by-letter spell of obvious, but I thought it would be good to visualize, so there is a clear picture, bringing the hope that this gets addressed at some point :). George -- [EMAIL PROTECTED] mailing list
