Re: What's the best package manager for single-package upgrades?

[Joe Rhett <jrhett@isite.net>]

> On Tue, 2003-11-04 at 02:35, Joe Rhett wrote:
> > I find it kindof sad that testing really doesn't appear to have any
> > function any longer.  One would like to run from testing and leave unstable
> > for the well, unstable stuff.  But I haven't really found much in testing,
> > which means one must be stale, or bleed on the edge.  Sux.
 
On Tue, Nov 04, 2003 at 03:23:48AM -0600, Alex Malinovich wrote:
> Well, in my experience, testing is most useful immediately following a
> new stable release, and least useful immediately preceding a new stable
> release. If you were to have started using Sarge right after Woody came
> out, I think you would have been rather happy. But now that everyone's
> trying to get Sarge ready to ship out, there's not many current things
> going in.
 
Isn't the point of testing that it should contain what will become stable?
If testing is what is supposed to be the next release, then it seems
pointless to even bother. "Testing" still has Mozilla 1.0.  That's what,
2 years old?

Unless I misunderstand the structure, shouldn't "testing" have lots of
stuff in it just prior to a new release?  There's almost zero updates in
testing ..

> Though Sid is definitely not the bleeding edge of stuff in Debian. Sid
> is, generally speaking, quite stable. There's the occasional hiccup, but
> I can count on one hand the number of major problems I've had with Sid
> in the entire time I've been using Debian. (About 2 years now)
> 
> If you really want bleeding edge, you add experimental to your
> sources.list. That's where you get all the really fun stuff... :)
 
Okay, so "testing" isn't.  Unstable is really "testing" and experimental
(not described in the debian documentation) is really unstable?

> > In a perfect world, people would hammer things and then roll them into
> > testing once they had been in unstable long enough without bug reports.
> > This would allow us to keep high-uptime systems running the same kernels
> > and such as our test/burn/destroy/rebuild laptops ;-) 
> 
> Well, that's basically exactly how it works. There's quite a few extra
> details but that's the "meat and potatoes" of it so to speak. :)

Then why is there really zero updates in testing?

-- 
Joe Rhett                                                      Chief Geek
JRhett@Isite.Net                                      Isite Services, Inc.