[packagers] Fedora Extra's and plans to add support for RHEL, CentOS etc

[Dag Wieers]

On Sun, 24 Sep 2006, Jonathan Underwood wrote:

> Dear Dag, Dries (+Matthias, but I think you're already aware),
> 
> I'm not sure if you're aware, but over on the fedora-extras-list there
> has been discussion about extending Fedora Extras to add support for
> RHEL and CENTOS ("Enterprise Extras").
>
> I started a bit of a discussion, pointing out that what you guys are
> doing at rpmforge already does similar things, and that the FE guys
> should consider involving you in discussions, since you've no doubt
> got a lot of wisdom to add from your own efforts, and that anyway way
> of working together would be beneficial. Anyway, you might want to
> check out the following thread:
> 
> https://www.redhat.com/archives/fedora-extras-list/2006-September/msg00714.html

The irony is that when I proposed at the start nobody was interested. 
The biggest difference in the approach (and you can find discussions 
about that on the fedora lists in the beginning back when I was still 
motivated) is that I think SPEC files should not be forked for different 
distributions as the rule, but only when it's necessary.

When the complexity of maintaining a single SPEC file of a project becomes 
a bigger effort than forking, that's when you'd fork.

Besides that I think Red Hat should have spend more effort in making the 
SPEC syntax more flexible and providing more infrastructure (and 
backporting it to EL) for cross-building packages from a single SPEC file.

The Fedora project never was interested in this because they never needed 
to keep compatibility. Every package they release will be out of 
support in a year. With a turn-around like that you only focus on the 
future and not on the past. You can fork and work on the fork.

Now Extras has much more resources and if they plan to do it, it doesn't 
matter whether the process is efficient or not. At least they have an 
active community that is larger than 3 persons so they will prevail.

Our biggest problem is that we never succeeded in making a community 
effort out of it and that's probably my fault more than anyone elses.

On the other hand I would have felt really bad if I had spend lots of 
effort trying :) At least now I can wish them luck with the effort without 
regretting anything.

But if people are still interested, we need someone to take care of the 
RPMforge website, someone who likes to lead the project and motivate 
people and a better hosting solution so we can allow people access to 
subversion etc...

If one thing is certain, I don't have more time to spend then I already do 
packaging. And I'm pretty sure that I'd be less useful in a Fedora-alike 
structure than what I am able to do outside of it. (email backlog included)
I still have a vivid memory of the email discussions that lead nowhere :)

On the other hand, the Fedora packages that are maintained properly have a 
higher quality than the RPMforge packages. One only wonders if they can 
keep that up once they support also EL2.1, EL3, EL4 and EL5.

*If* they broaden the scope to that. My bet is that it will be the latest 
EL release only (because that fits in their infrastructure) and that will 
be similar to what Karanbir Singh is already doing.

Fact is that in most companies that latest release of EL is always at 
least one year away from general availability. So when companies start 
deploying EL5 (and still need support for EL3 and EL4), FC7 will be 
mainstream and FC8 is there to replace it.

For example, the company I now work for is moving from EL2.1 to EL3. They 
can't go to EL4 because proprietary applications are not supported. They 
have no use to Extras if Extras only focuses on EL5. Once they switch to 
EL5, Fedora Extras will be looking at EL7...

Jonathan, thanks for the reference though. It's nice to know what changes 
we can expect in the RPM market :)

Kind regards,
--   dag wieers,  dag at wieers.com,  http://dag.wieers.com/   --
[all I want is a warm bed and a kind word and unlimited power]