On Fri, Nov 18, 2005 at 01:23:42PM +0100, Thomas Bächler wrote: >Magnus Therning schrieb: >>>Our release structure isn't like that at all. We just take a snapshot >>>of the state of current and that's a release. The rolling part means >>>that the difference between releases is all the updates we've done. >>> >>>A release is out of date seconds after it's been made. We don't touch >>>it after that. >> >> >> How does Arch cope with big changes that affect a large set of the >> available packages? > >Look at what happened with gcc4 migration and libtool-slay recently: >The new packages went to [testing] until the migration was (almost) >finished, then the new versions were moved from [testing] to >[current]/[extra] when everything seemed stable. That was a 500MB >Update for many people :)
Wow, and was it relatively painless for users to do the upgrade? >Unfortunately, [current] and [extra] didn't see many updates during the >testing period, so many packages there were out of date. I guess that's to be expected. It's probably also a good thing in many cases. It'd be a bit of a waste of time/energy to modify something that's on top of a base that's about to move signigicantly. As long as [current] and [extra] picks up pace again after the update I don't think it's too much of a problem. /M -- Magnus Therning (OpenPGP: 0xAB4DFBA4) [EMAIL PROTECTED] http://therning.org/magnus Software is not manufactured, it is something you write and publish. Keep Europe free from software patents, we do not want censorship by patent law on written works. The sort of person who uses phrases like `incompletely socialized' usually thinks hackers are. Hackers regard such people with contempt when they notice them at all.
pgpei6CcHTL1L.pgp
Description: PGP signature
_______________________________________________ arch mailing list [email protected] http://www.archlinux.org/mailman/listinfo/arch
