On Mon, 19 Oct 1998, George Bonser wrote: > On Tue, 20 Oct 1998, Nikolai Andreyevich Luzan wrote: > > I just wonder .... when a freeze is announced, a lot of people load up > their systems from unstable to give it a good workout. You might consider > this a wide but not ubiquitous beta. At this point, now that you have > several more times more systems in sync with the unstable tree, would it > make sense for someone ( other than the package maintainer ) to install > the package and check it out before addig it to the archive? I mean, not > an extensive feature bugcheck, just a basic integrity check to make sure > that it does not crash a system when installed.
A freeze is the responsibilty of the committee that oversees the entire distribution, not that of individual package maintainers. The fact that a freeze is announce does not mean inherant stability, all it means is that no new things will be added for that release of the distribution, the release could be buggy as all hell. The announce of freeze is analogous to the close acceptance of papers for a conference. > I think everyone would agree that before a freeze is released, you should > expect anything to happen if you are in the unstable tree. There is an > implied level of security when it changes to "frozen" that all the basic > system crashing bugs have been worked out .... that is why its status was > changed. A freeze is not released, it is simply another step towards the release of a distribution, by the very nature of the unstable tree you should expect nothing to be perfect in there. And s I have already said freeze does not imply stability, it simplie states that the only submissions that will be accepted for inclusion into the tree are those that fix pre-existing bugs, it also by no means implies that _any_ bugs have been fixed. The statuse merely changes to frozen to tell people that release is approaching and that all the bugs are being fixed. Nikolai
