On Sun, Jun 13, 2004 at 22:46:48 +0200, Wouter Verhelst wrote: > Never the less, I think the original question (whether 7.2.4 should be > accepted for woody) is a good one.
True, though I think a better question would be a somewhat more generalised one: are we happy with the current rules for updates to stable, and if not, what improvements to them can we think of and what support (manpower, infrastructure) would be needed if they are improve? > If I understood Martin Pitt's explanation correctly, the postgresql > upstream doesn't lightly perform third-level updates such as the one from > 7.2.1 to 7.2.4. Debian currently has very long release cycles and very conservative rules for updates to stable. This has contributed much to stable being very stable, but the popularity of backports.org and apt-get.org indicates that there is a large audience who appreciate stability but that have requirements which our stable distribution in its current form doesn't meet. Professionally, I am responsible for several server systems. I try to keep these systems based as closely on stable as possible, but over the last year I have switched to backports.org and apt-get.org versions of several packages, for several reasons: - Ease and flexibility of administration. For instance, I prefer exim4's configuration setup and flexibility over exim3's. - For functionality that is not available in stable. In particular, ClamAV. - For functionality for which the stable version is too outdated to be useful. There are some categories of software for which bits rot faster, in particular spam defenses (like spamassassin) and IDSes (like snort) I am quite happy with backports.org / apt-get.org, but it has some limitations: - Support for non-i386 is spotty at best. Stable is fully supported on several architectures and has a buildd infrastructure to keep it that way. - Packages don't neccessarily cooperate well. - There is no active security team like stable has; security updates are done on a best effort basis at best. I think these kind of concerns affect many system administrators, and I would appreciate suggestions on how to cater better for this audience within the Debian project framework (rather than alongside it as backports.org and apt-get.org currently are). Ray -- RUMOUR Believe all you hear. Your world may not be a better one than the one the blocks live in but it'll be a sight more vivid. - The Hipcrime Vocab by Chad C. Mulligan