[Yum-devel] Defining other repository types in yum

Bret McMillan bretm at redhat.com
Thu Apr 13 00:13:54 UTC 2006


On Tue, 2006-04-11 at 17:41 -0400, seth vidal wrote:
> On Mon, 2006-03-20 at 15:32 -0500, seth vidal wrote:
> > On Mon, 2006-03-20 at 15:06 -0500, James Bowes wrote:
> > > Hi all:
> > > 
> > > I'm one of the engineers working on Red Hat Network; we've been looking
> > > at yum as a basis for future development work. As a proof-of-concept,
> > > we have a modified yum interacting with RHN through a plugin.
> > > 
> > > The plugin contains RHN specific Repository, PackageSack, and
> > > Package objects. To get these to work, we had to make some changes
> > > to yum proper.
> > > 
> > > Following this email I'm going to send a series of four patches. These
> > > encompass our work on yum so far, and have allowed us to build a plugin
> > > that allows yum to speak to an RHN server.
> > > 
> > 
> > cool.
> > 
> > 
> > > Some things that we might be looking at down the road, and that we'd
> > > like to work with you all on integrating into yum are:
> > > 
> > >     * Token based authentication - Beyond just a username and password,
> > >       we'd like to be able to use arbitrary tickets or credentials to
> > >       authenticate with the server.
> > 
> > if you're thinking kerberos tickets then wouldn't it make more sense to:
> > 
> > - put it into urllib inside python and then have it inherit its way up
> > through urlgrabber and into yum?
> > 
> > 
> > 
> > >     * Broader SSL support - Particularly, we'd like to have yum be able to
> > >       check with a CA (in our case, just a local certificate) to ensure
> > >       that the host is who it says it is.
> > 
> > ditto of above - why not make it possible for urllib to handle this
> > sanely and bubble it up through urlgrabber?
> > 
> > 
> 
> Any responses in general to my comments from before?

Probably makes sense, I don't think Pete or James have had a chance yet
to really take much of a look @ urlgrabber... iirc, it's sortof at the
level that rhnlib is for us, but rhnlib has a lot more business logic
around it.  So the question will be what parts make sense to yank out of
rhnlib and put into urlgrabber, and what bits do we keep separate, and
what's the timeframe, etc.  I'll be working w/ jbowes to sketch out the
plans a little more this/next week.

--Bret




More information about the Yum-devel mailing list