[Yum] Feature request(s)

[Robert G. Brown]

On Wed, 8 Oct 2003, Garrick Staples wrote:

> I'd like yum to report full URLs.  Then downloading manually would be
> trivial.  (if you ignore the possiblity of requiring authentication,
> configuring proxies, etc).

Or, as originally suggested, permit/modify yum to just download the
rpm's without installing them.  Jeeze, this is pretty simple and you can
do it now by ^C'ing right after it finishes putting an rpm in the cache
but before it installs.  This is just asking for a more elegant exit
than a program break (and maybe a syntax that moves the rpm to a given
destination path instead of "only" to cache).

Agreed, there are tools like wget you can use instead, but you can't do
the equivalent of "yum list \*frog\*" to see all rpm's with frog in them
in a repository known to match your distribution, then "yum get frog ."
to get frog-1.2-1.i386.rpm and put it in the cwd.

Another thing that this might be useful for is getting src rpm's.  I
predict that with Red Hat going heavy commercial, there is going to be a
lot more interest in repositories that serve src rpm's, not binary.
Maintain a working base system, but fully automate the process of going
from src rpm's to installed binary rpm's.  I could easily see a GPL
future where instead of getting binary distributions at all one builds a
binary distribution in real time from GPL sources.

The issue to me is less one of whether there are other tools one can use
to download files; of course there are.  However, yum is by its nature
an INTEGRATION tool -- it exists as a fundamental one-stop-shop
interface between (hu)Man (of both sexes), Machine, and Repository.  So
any function that can sensibly exploit the layer of abstraction that it
creates between "repository" (collection of rpm's believed to be
consistent for some distribution level) and the rest is worth at least
considering.  This one doesn't seem that crazy to me.

I'd even go it one better.  "yum clone" to perfectly mirror/clone any
repository (set) to which yum has access.  Useful for creating backup
servers/repositories.  Useful for creating/mirroring new ones.  Sure,
rsync will do it.  So will wget.  Both have their OWN flaws -- the need
to use multiple tools, idiosyncracies of their own, maybe access
restrictions.  Also, yum now supports any URI, right?  rsync and wget
don't necessarily work on all URI types.  This would simplify the howto
considerably, as a fairly major chunk of using yum as an admin in a LAN
environment is creating a repository.

At some point you stop making things simpler (or you start making the
originally simple tool too hard to maintain and a bit gaudy).  However,
the real tradeoff is that using MULTIPLE tools to achieve the same work
is often a lot more complex.  Everything yum does could be done by the
use of a few other tools (e.g. wget, rpm) and some sweat.  It is just
that there is a LOT of sweat because yum does a lot for you and hides
complexity.

   rgb

> 
> 
> _______________________________________________
> Yum mailing list
> Yum at lists.dulug.duke.edu
> https://lists.dulug.duke.edu/mailman/listinfo/yum
> 

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu