[Mageia-sysadm] [sysadmin-commits] [2736] do not explictely set the same value as default

[Michael Scherer]

Le dimanche 15 avril 2012 à 15:01 +0200, nicolas vigier a écrit :
> On Sun, 15 Apr 2012, Michael Scherer wrote:
> 
> > 
> > And for the sake of coherence, the whole repository was cleaned a few
> > months ago to have manifests without the same boilerplate, and no one
> > commented, and there is nothing special in repoctl to be different. 
> 
> I didn't see any mail on this mailing list about "cleaning". It would
> have been nice to discuss the puppet style to use before changing
> the whole repository.

I think we all would like to have our code being reused outside of our
infrastructure. And you are the one who try to push for a more direct
approach of reusability, as seen by GSOC Proposal, and various mail.
As seen in this thread, despites agreeing that sharing knowledge is a
good thing, the exact details on how to do it are still unspecified and
we ( as well as the whole puppet community ) didn't agree on how to do
it, ie there is problem to solve like "how to take in account different
packages name, how to take in account different puppet version", etc.

But something rather clear is that the puppet community seems to agree
on a style guide, being recommended and followed by most people. 

For example, puppetlabs use this as a baseline to evaluate modules ( like on
http://puppetlabs.com/blog/module-of-the-week-benoitcattienginx-basic-nginx-configuration/ ), 
and is part of their recommendation ( http://puppetlabs.com/blog/writing-great-modules-an-introduction/ ).

( and in fact, you can often see reference to it on the blog ).

So this looked to me the good thing to do, and that anyone would see the
work ( we have commit diff for that reason ) and would ask questions if
any.

Puppet-lint also return problems that would bring issues when puppet 2.8
will be out ( variable scoping issues ), and like any checking system,
would be ignored if there is a endless pile of false positive, or "stuff
that we do not care" ( I think that we have experimented enough with rpmlint on that
matter, with people no longer using it until we decided to make it mandatory ). 
I also started to add a check for module name with
a -, as this should/would be forbidden for various reasons ( ie, parser
support it for now, but you cannot reference variable in it, and this
would likely break with puppet 2.8 ).

So cleaning seemed to be a good thing to do for all thoses reasons, and still seem.
-- 
Michael Scherer