On Aug 6, 2004, Jeff Spaleta <jspaleta@xxxxxxxxx> wrote: > On Fri, 06 Aug 2004 11:16:29 +0200, Farkas Levente <lfarkas@xxxxxxxxx> wrote: >> is there any reason why outdated rpms are in updates/testing/ directory? > Why is it confusing? Only in that one can't assume the packages in testing are always >= packages in released updates. Personally, I'd much rather have packages in testing never *moved* to released updates, but rather have testing be a superset of the released updates. With hard- or soft-links, this would be no additional space, and then those that want to track the testing repo wouldn't have to have another repository in his tree. Why does it matter? My reason has to do with web caching. I do track testing updates on all of my boxes at home. It annoys me immensely when, after downloading a large update such as say kernel, xorg-x11 or openoffice.org on one of the boxes, I start rolling the update onto the other boxes, and, because the rpms were moved from testing to official updates, I either get a failure to download the rpms because (outdated in web cache) header.info for testing updates says they should still be there but they were (re)moved, or I end up having to download the rpms again, now from the released updates. -- Alexandre Oliva http://www.ic.unicamp.br/~oliva/ Red Hat Compiler Engineer aoliva@{redhat.com, gcc.gnu.org} Free Software Evangelist oliva@{lsd.ic.unicamp.br, gnu.org}
