Summary of the 2008-04-08 Packaging Committee meeting

[Jason Corley]

> On Tue, 2008-05-13 at 11:21 -0400, Jason Corley wrote:
> > Tom "spot" Callaway wrote:
> > > The approved JPackage naming exception says that when a technical
> > > solution is found to make .jpp tagging obsolete for the purposes of
> > > grouping excludes, the exception will vanish. I think we've done that,
> > > insults and flames aside.
> >
> > The portion of that agreement that you are glossing over nicely was
> > that the solution needed to be in rpm, not yum, apt, smart, up2date,
> > or anywhere else.  Plain and simple a yum plugin has zero to do with
> > the exception and conditions as understood on the JPackage side.
> > Jason
>
> Given that I wrote the exception, I can rather confidently say that it
> doesn't say that.

What was agreed to on the JPackage side is what I wrote.  Feel free to
hop on Red Hat internal IRC and ask fnasser if that was his
understanding at the time as well.  What you have written in no way
changes what we were attempting to get at with the original agreement,
regardless of their divergence at this point.

> ~spot, who isn't sure why he's arguing with you, as he know no good will
> come from it.

Productive.  If it makes you feel any better I think any and all
interaction with you is also downright painful.  You're pedantic on
issues you care about, flippant for all others.  Truth be told I
expected exactly this sort of reaction from the person who said:

    [12:33] <spot> jpackage is no more of an upstream than Fedora is.

The above quote comes from the Fedora Packaging minutes for those that
didn't read them
(http://fedoraproject.org/wiki/Packaging/Minutes20080408).  I find
that position curious given that there is zero debate that the
packages in question are derived from JPackage.  You appear to think
of Fedora as upstream and RHEL as downstream -- or at least that's the
impression I've gotten from email, IRC, and other snippets I've read
in passing -- yet for an external repository that you have no control
over (but which you base a deliverable upon) you change your tune.  I
think if RHEL used Fedora packages and didn't submit patches and
improvements back to the Fedora project you'd have a very different
outlook on the problem.  "Upstream! Upstream! Upstream!" is a
selectively applied mantra I guess.

As an illustrative example, let's look at Fedora's tomcat5.  It's
based upon a package which I wrote and introduced into JPackage (which
in turn was based upon the tomcat4 JPackage work done by Henri Gomez,
and others).

    http://koji.fedoraproject.org/koji/buildinfo?buildID=35821

>From 5.5.23-9jpp.4 the following was never submitted back to JPackage
(even as a patch sent by email):
    - Add jasper-eclipse subpackage which is needed for eclipse 3.3.
    - Inject OSGi manifest into servlet api jar and jsp api jar.

>From 5.5.25-2jpp.2 the following was never submitted back to JPackage
(even as a patch sent by email):
    - Fix init script for bz #380921
    - Fix tomcat5.conf and spec file for bz #253605
    - Fix for bz #426850
    - Fix for bz #312561
    - Fix init script, per bz #247077
    - Fix builds on alpha, per bz #253827

I could go on if you'd like, I have many, many more examples.
Jason