[cpanrpm] Fwd: PAUSE indexer report SHLOMIF/CPANPLUS-Dist-Fedora-0.0.1.tar.gz

[Chris Weyl]

On Thu, Sep 4, 2008 at 5:33 AM, Shlomi Fish <shlomif at iglu.org.il> wrote:
>
> On Wednesday 27 August 2008, Jerome Quelin wrote:
> > On 08/08/25 23:27 +0300, Shlomi Fish wrote:
> > > On Monday 25 August 2008, Jerome Quelin wrote:
> > > > ==> what do you think of creating a cpanplus::dist::rpm which would
> > > > provide all the duplicated logic, and cpanplus::dist::mdv / fedora
> > > > inheriting from this module instead of cpanplus::dist::base?
> > >
> > > Sounds good.
> >
> > well, in fact, since cpanplus::dist::rpm and subclasses fedora /
> > mandriva / others will have to work together, it might be better to
> > provide only one dist including the subclasses... this way, we would be
> > sure that a given cpanplus::dist::mdv release always works.
> >
> > wdyt?
> >
>
> Sounds good.
>
> However, note that I received a message from Chris Weyl here with a complex
> patch for CPANPLUS::Dist::Fedora that aims to solve some problems. This patch
> uses Template-Toolkit to create a more compliant spec, and so is worth a look
> as the basis for CPANPLUS::Dist::Rpm .

Having a base ::Rpm class would make sense.  To me, it seems to break
out into (and I'm stating the obvious here):

* code to interact with a given dist and figure out "stuff" about it,
e.g. license, status, etc
* code to handle the basics of taking that, generating a spec, and
running rpm against it
* distro-specific spec template(s)
* distro-specific overrides to build behaviour

Using Template::Toolkit to deal with the spec templates will hopefully
help, in terms of keeping the spec generation somewhat distinct from
determining the relevant information from the dist.

What's the next step?  Note I assume you don't mind me helping out :-)

                                    -Chris
--
Chris Weyl
Ex astris, scientia