[Fedora-packaging] Proposal to reduce anti-bundling requirements

[Adam Williamson]

On Fri, 2015-09-11 at 12:06 -0600, Kevin Fenzi wrote:
> On Fri, 11 Sep 2015 10:51:42 -0700
> Adam Williamson <adamwill at fedoraproject.org> wrote:
> 
> > On Fri, 2015-09-11 at 13:35 -0400, Stephen Gallagher wrote:
> > 
> > > As for which components, it's not about specific examples[1].
> > > It's
> > > about solving the question in a generic way. We have quite a lot
> > > of
> > > software that isn't packaged for Fedora (either not started or
> > > aborted
> > > when the package review couldn't be passed) that has genuine
> > > value.
> > 
> > I can certainly confirm that. I dug through quite a lot of review
> > requests yesterday to look at how the rules have been applied in
> > practice, and found several that have been abandoned because of
> > bundling issues. I'll just link one example -
> > https://bugzilla.redhat.com/show_bug.cgi?id=836810 - but it's
> > trivial
> > to find more.
> 
> But by the same token, a great deal of upstream projects don't bundle
> things and are just fine packaged up in Fedora. 

I don't think anyone was advocating a policy that all packages must
have a minimum of 1 (one) bundled library ;)

I agree that the discussion here needs to be more broad-based; see the
other thread fork. I was just providing support for Stephen's
contention that this is not some airy-fairy theoretical problem, there
are multiple examples of real things that people *wanted* to have
packaged that are not packaged because the unbundling process was too
onerous.

-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net