Gary Thomas wrote:

Chris Lumens wrote:
Just thinking out loud. Could there be a requirement path from
kernel to selinux? It could be whats happening here?


Yes, that's exactly what's going on here. Something in @base has a
requirement on an selinux package, and those requirements get pulled in
regardless of what is removed in the kickstart %packages section since


Sorry, I don't buy it. From a fully loaded Fedora 8 system:
# rpm -q --whatrequires selinux-policy
no package requires selinux-policy
# rpm -q --whatrequires selinux-policy-targeted
no package requires selinux-policy-targeted

Also, I used pungi to rebuild my base repository (the one I was
testing with was created by hand) and there's no selinux-*
packages anywhere. Installing from this repository works just
fine.



use `rpmquery` for those types of queries
Additionally the requires does not have to be at the first level. To tell you the truth don't know in what level it might be in. Havent done the math. If you by any chance discover it, post it here. I would be interested in seeing exactly where that happens.

Regards.

--
Joel Andres Granados
Red Hat / Brno, Czech Republic

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

I was doing some work in identifying a truly minimal package set
required to boot a system, and additionally minimal sets that get you
rpm and yum. That would then give us a base we could use to determine
dependencies which need to be made optional and to find packages which
need to be split.

With those lists in hand, we could work on rationalizing @core and
@base (because they're not really well-defined) and perhaps collapse
them into one group.

If anyone else is interested in this, please let me know.

- J<

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

Jason L Tibbitts III wrote:

I was doing some work in identifying a truly minimal package set
required to boot a system, and additionally minimal sets that get you
rpm and yum. That would then give us a base we could use to determine
dependencies which need to be made optional and to find packages which
need to be split.

With those lists in hand, we could work on rationalizing @core and
@base (because they're not really well-defined) and perhaps collapse
them into one group.

If anyone else is interested in this, please let me know.

- J<

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

post them on the list. unless they are huge files.
--
Joel Andres Granados
Red Hat / Brno, Czech Republic

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

Joel Andres Granados wrote:

Jason L Tibbitts III wrote:

I was doing some work in identifying a truly minimal package set
required to boot a system, and additionally minimal sets that get you
rpm and yum. That would then give us a base we could use to determine
dependencies which need to be made optional and to find packages which
need to be split.

With those lists in hand, we could work on rationalizing @core and
@base (because they're not really well-defined) and perhaps collapse
them into one group.

If anyone else is interested in this, please let me know.

- J<

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

post them on the list. unless they are huge files.


Im especially interested in the one with yum and rpm included
Does it have selinux?


--
Joel Andres Granados
Red Hat / Brno, Czech Republic

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

Jason L Tibbitts III wrote:

I was doing some work in identifying a truly minimal package set
required to boot a system, and additionally minimal sets that get you
rpm and yum. That would then give us a base we could use to determine
dependencies which need to be made optional and to find packages which
need to be split.

With those lists in hand, we could work on rationalizing @core and
@base (because they're not really well-defined) and perhaps collapse
them into one group.

If anyone else is interested in this, please let me know.


yes please. Rationalizing and well-defining @core and @base sounds
great to me I mean, if there is a rational reason why selinux policy
should be explicitly listed in those, then fine, but I haven't heard the
reason yet, just the fact.


-dmc

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

On Fri, 01 Feb 2008 15:17:55 -0600
Douglas McClendon <dmc.fedora@filteredperception.org> wrote:

> yes please. Rationalizing and well-defining @core and @base sounds
> great to me I mean, if there is a rational reason why selinux
> policy should be explicitly listed in those, then fine, but I haven't
> heard the reason yet, just the fact.

Fedora == SELinux in use. If you don't want SELinux, you're going to
have to make something that isn't Fedora.

--
Jesse Keating
Fedora -- All my bits are free, are yours?
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

On Fri, Feb 01, 2008 at 03:17:55PM -0600, Douglas McClendon wrote:
> >With those lists in hand, we could work on rationalizing @core and
> >@base (because they're not really well-defined) and perhaps collapse
> >them into one group.
> yes please. Rationalizing and well-defining @core and @base sounds
> great to me I mean, if there is a rational reason why selinux policy
> should be explicitly listed in those, then fine, but I haven't heard the
> reason yet, just the fact.

I think a rational split is:

1) Base is the minimum required to bootstrap -- this should contain no
more than necessary to install a minimal system from which one can
install everything else. (This should include rpm, and arguably yum.)

2) Core is everything that is "expected" to be installed in a minimal
environment as a matter of policy. selinux-policy probably belongs here.
man, vi, traceroute, openssh -- a minimal functional environment.

In other words, Base is a technical minimum, and Core is the
social one. That makes Core much harder to nail down -- but easy to bump
things from Base to Core.

--
Matthew Miller mattdm@mattdm.org <http://mattdm.org/>
Boston University Linux ------> <http://linux.bu.edu/>

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

On Fri, 1 Feb 2008 16:27:47 -0500
Matthew Miller <mattdm@mattdm.org> wrote:

> I think a rational split is:
>
> 1) Base is the minimum required to bootstrap -- this should contain no
> more than necessary to install a minimal system from which one can
> install everything else. (This should include rpm, and arguably
> yum.)
>
> 2) Core is everything that is "expected" to be installed in a minimal
> environment as a matter of policy. selinux-policy probably belongs
> here. man, vi, traceroute, openssh -- a minimal functional
> environment.
>
> In other words, Base is a technical minimum, and Core is the
> social one. That makes Core much harder to nail down -- but easy to
> bump things from Base to Core.

You've just about got it nailed, only switch your uses of Core with
Base, and vice versa (:

Base is the one you can disable during install. You can't get away
without Core though.

--
Jesse Keating
Fedora -- All my bits are free, are yours?
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

On Fri, 2008-02-01 at 16:27 -0500, Matthew Miller wrote:
> On Fri, Feb 01, 2008 at 03:17:55PM -0600, Douglas McClendon wrote:
> > >With those lists in hand, we could work on rationalizing @core and
> > >@base (because they're not really well-defined) and perhaps collapse
> > >them into one group.
> > yes please. Rationalizing and well-defining @core and @base sounds
> > great to me I mean, if there is a rational reason why selinux policy
> > should be explicitly listed in those, then fine, but I haven't heard the
> > reason yet, just the fact.
>
> I think a rational split is:

You have the right idea, but Base and Core backwards. And really,
that's where they stand today[1]. An install with just @core had next
to nothing which could be removed after the install completed. And most
of the things which could were things which are used at install-time
(eg, lokkit, etc). If you shrink the locale archive by cutting it down
to just what you want, remove some translations you get to under 200
megs iirc

Jeremy

[1] Well, I haven't looked at base as closely at core in a little while.


_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list

On Fri, Feb 01, 2008 at 04:36:20PM -0500, Jeremy Katz wrote:
> You have the right idea, but Base and Core backwards. And really,

Err, yes, thanks. But yeah.


--
Matthew Miller mattdm@mattdm.org <http://mattdm.org/>
Boston University Linux ------> <http://linux.bu.edu/>

_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/anaconda-devel-list