On Thu, 2004-11-11 at 16:45 -0500, Paul Iadonisi wrote:
[snip]
Subsequently upgrading via yum/rpm would not have caught that
problem
even when upgrading to the officially released rpm-4.3.2-21 and the
officially released rpms of the packages in question (the one that
overwrote the file from another package, and the one who's file got
overwritten).
Sorry. Slight logic problem here. If there were new versions of the
overwriting and the overwritten rpms then the problem would have been
fixed by an upgrade. However, let's say only the overwriting package
had an update from FC3RC5->FC3, but the overwritten package didn't
(which is highly probable to fix the overwriting problem that happened
due to file conflicts being turned off). Then you would be left with
the overwritten package whose version didn't change from FC3RC5->FC3,
because an upgrade would not upgrade that rpm (it's version didn't
change, so why should it?).
Also, I should clarify that when I (and others) say there is no
difference between test releases and release candidates, that's probably
a *little* strong. What I mean to say is that it is *only* a matter of
degree. I.e.: the changes from RC->Official are likely smaller and less
likely to cause problems on upgrade. But it still stands that upgrades
from test(or RC) to Official should not be relied upon. That line isn't
crossed until Official release time.
--
-Paul Iadonisi
Senior System Administrator
Red Hat Certified Engineer / Local Linux Lobbyist
Ever see a penguin fly? -- Try Linux.
GPL all the way: Sell services, don't lease secrets