<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> Jesse Keating wrote: <blockquote cite="mid1237563335.9564.10.camel@localhost.localdomain" type="cite"> <pre wrap="">On Fri, 2009-03-20 at 08:07 -0500, Adam Miller wrote: </pre> <blockquote type="cite"> <pre wrap="">Right, but my comment was more geared towards the fact that installing from F11-Alpha and trying to run 'yum update' at one point wasn't possible without first running 'yum update rpm' and then 'yum update'. I think that is a flaw in the priority of package installation, if it had first checked for a new yum or new rpm package, installed those and then re-invoked itself then the problem would be solved. </pre> </blockquote> <pre wrap=""> In the rare scenario of alpha -> missed some updates -> rawhide case, it might help some. In the far more common F10 -> rawhide case it wouldn't help one bit. Why? Because trying to go from an out of date F10 directly to rawhide rpm would drag in hundreds and hundreds of packages due to deps, because python changed. The only way around this is to fully update your F10 with F10 updates, and /then/ proceed to rawhide, which you can do in one full swing. The updater updating itself and restarting doesn't really solve enough scenarios to add in the complication and complexity and numerous new failure cases. </pre> </blockquote> <tt>The need to make sure that you update rpm or yum before you do 'yum update' has happened before and I'm sure this isn't the last time that it will happen either. It just makes sense for a packager to see if there is an update to itself first. Update itself and THEN do the updates for everything else. Regards, Gerry </tt> </body> </html>