On Wed, Oct 08, 2003 at 06:40:52PM -0400, Jeremy Katz wrote: > On Tue, 2003-10-07 at 17:28, Christopher C. Weis wrote: > > 2) The ability to download, but not install, RPMs, using Yum. I don't have strong feelings on the matter. I'm just discussing because I like to :) > I've heard a few reasons, and all of them seem ... questionable, at > least to me. Here's the sample of the ones I remember: > 3) "I don't trust the code which installs packages" -- Then stop using > it. The ones I've snipped, I agree with you on, Jeremy. This is the one you've dealt with most crudely, both in the representation and the response. Trust is not binary. It seems perfectly reasonable that someone would want to physically be in front of the box when the upgrade happens. In fact, I think that's pretty common important machines. It's not exactly unprecedented that an upgrade breaks something (usually because some obscure condition wasn't checked by the author/maintainer/packager). However, this only really matters (AFAICT, someone please correct me if I'm wrong) when you ALSO have a slow connection to the repository. It seems like the overlap would be rather small. I guess I can think of a company or two that is cursed (due to location) with dialup that would nonetheless be in rough shape if their internal database/file server went down. /me shrugs... I know _I_ don't want to implement it :) -Michael -- Michael Stenner Office Phone: 919-660-2513 Duke University, Dept. of Physics mstenner@xxxxxxxxxxxx Box 90305, Durham N.C. 27708-0305
