On Sat, 2005-11-26 at 01:01 -0500, Dave Jones wrote: > On Sat, Nov 26, 2005 at 12:54:23AM -0500, Jeremy Katz wrote: > > On Sat, 2005-11-26 at 00:48 -0500, Dave Jones wrote: > > > On Sat, Nov 26, 2005 at 12:39:07AM -0500, Jeremy Katz wrote: > > > > On Fri, 2005-11-25 at 10:54 +0100, Joachim Frieben wrote: > > > > > "up2date-gnome" has a very clear, informative interface. You start with a > > > > > channel windows that allows you review the available channels and to make > > > > > your choice. Very nice, indeed. > > > > > > > > Why would you want to not download updates from all repositories you > > > > have configured? Or at least see what's available. The fact that the > > > > updates come from multiple repositories is a detail that I don't think > > > > users really want to / should need to care about. > > > > > > The last time it happened to me, when yum fails to contact one repo, > > > a 'yum update' fails completely. Ie, it fails for every repo, instead > > > of downloading updates from the repos that it _could_ contact. > > > > > > If current yum has been fixed to work in the face of adversity, > > > then I agree, otherwise, the ability to easily disable a broken repo > > > is useful. > > > > How do you know that updating when a repository is broken is safe, > > though? If you have a local repository which shadows the main repo but > > with packages with config changes, going back to base could completely > > break your system. While punt to the user may be sufficient for you or > > for me (or pretty much everyone on this list), I don't think that it is > > for the "typical" end-user. > > it's arguable that that configuration is also not 'typical end-user' ;-) > Really, I think if someone goes far enough to setup shadow repos, they > should know what their doing enough to figure out what to do when > repo breakage occurs. I also worry about third party repos which change out core packages :-/ There are just things a number of things which can go wrong/strange if you have repositories not available, which is why things do what they do now[1] Jeremy [1] Okay, I'm all but certain that pup needs to give a better error message when this occurs right now, but that's a side point ;-) -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
