On Sun, Jun 30, 2002 at 02:19:17PM -0500, Scott Dier wrote: > Making such statements on such corner applications is so misleading I > don't feel the need to argue about this, however... I'll be happy to > avoid rpms flaws in dependencies. (file deps, versioned deps that might > be asking for something in a different dist, thank god mandrake renames > their versions to make their end of the pool sane) > The problem is that people come into #linux on efnet and want to know why $XYZ feature a) doesn't exist or b) is broken, and it turns out they're running some incredibly old version, mostly the comment was in response to the general 'd3b1an 0wnz j00 b3cau$3 0f apt-get', apt-get is a great utility assuming you have a decent up-to-date repository. And it's no longer debian specific. Debian has a lot of old packages, redhat (so far) has a lot of non- existant packages (which is better, old or nonexistant?) Personally I'd prefer people use the up to date versions of packages. > Some people really dont care that theres a bleeding edge package out > now, but that what they are using is considered non-crack by the > maintainer. If its true that the current revision is non-crack and > should be packaged, file a bug. If the maintainer doesn't give a rats > ass and can't seem to care, find someone else to package it or do it > yourself. Eggdrop (a project I've been working on for years) isn't nearly a 'corner application', yet stable still has 1.3.28 which is horribly bug-ridden (I do mean horribly). I've been told that this is because no further updates can be made to stable, yet these are remotely exploitable bugs that can grant shell access. (and shell access to every bot connected to that bot) > > If none of those work, it shouldn't be packaged in the first place for > lack of interest. If you cant find one DD to care enough about it and > you can't find one user who can't volunteer some time, oh well. Perhaps the people responsible for allowing people to create the packages should make sure the people are going to continue maintaining them, instead of doing it a few times a year to keep their name in it. > > Yes, we know the devlopment cycle sucks right now. This will be a major > topic later this week in Toronto I'm guessing. Many proposals want > targets of 6-12 months without making major infrastructure changes. I'm > personally up for point releases with unchanged base and many > end-user-application updates, with major releases having the big > infrastructure changes. (new dpkg, new init system, etc.) The nasty > problem about this is that the debian development infrastructure does > not lend itself to this sort of development. I'm not exactly sure yet > how to attack that issue yet other than create infrastructure outside of > debian to experiment. As I said, I run debian on my laptop because the only programs I run on the laptop are dhcpcd, kismet, prismstumber, and ssh. For any system that I use as a desktop debian just doesn't cut it. Hopefully the above changes will come about, I'd be willing to take another look at it as a desktop distribution. > Scott Dier <dieman at ringworld.org> http://www.ringworld.org/ -- Matthew S. Hallacy FUBAR, LART, BOFH Certified http://www.poptix.net GPG public key 0x01938203