Re: ye olde upgrade vs. dist-upgrade
scripsit Marc Wilson:
> On Thu, Dec 04, 2003 at 02:57:16PM -0700, Thanasis Kinias wrote:
> > I wonder the same thing as Marc.
>
> You're not wondering the same thing as me... I know perfectly well
> what the two targets do. It's Bill Moseley who's doing the wondering.
Sorry, brain-finger connection problem there. I do not doubt your
expertise.
> > I always do dist-upgrade also. Since I also always use -u, I'm not
> > worried about its removing or installing things I don't want...
>
> Uh, no, all that does is show you what it's going to do without actually
> *doing* it. It has nothing to do with what you're *allowing* it to do.
> Assuming it shows you that it intends to remove a package, or install a new
> one... what are you going to do then? Are you going to still turn it
> loose, or are you going to investigate why?
If I discover that it's going to remove something I need (for whatever
reason), I will certainly investigate why, and use pins as necessary to
prevent it. I'm not going to empower apt potentially to remove packages
without checking with me first!
> There should never be a reason to need 'dist-upgrade' if you're running
> stable.
That certainly makes sense. I should have mentioned, I suppose, that I
run mostly testing -- so there is fairly often the need to do
`dist-upgrade'.
> Certainly. See above. If you don't want to give apt the power to
> change the installation state of a package, you don't use
> 'dist-upgrade'. Why would you give it that power, if it weren't
> necessary?
Let me rephrase that: Given that it is (for a system tracking testing)
at times necessary to do `dist-upgrade', is there any reason not to do
it always?
The alternative is to do `upgrade' routinely, and then redo it with
`dist-upgrade' when it fails occasionally, which (unless there's a good
reason to do it that way) seems like adding a needless extra step.
(Analogy: If a script will only ever be run by bash, why do `FOO=bar;
export FOO' when `export FOO=bar' will do?)
--
Pax vobiscum; pax cum omnibus.
Thanasis Kinias
tkinias at asu.edu
Doctoral Student, Department of History
Arizona State University
Tempe, Arizona, U.S.A.
Reply to: