On 20 April 2012 23:21, Corentin Chary <corentin.ch...@gmail.com> wrote:
>
> Currently it uses SRC_URI and HOMEPAGE, but honestly it wouldn't be
> hard to use any other environment variable and to do some checks on a
> webservice.
> Anyway for tricky cases it can still be done by hand.
>
> --
> Corentin Chary
> http://xf.iksaif.net
>
If you really want to support Perl Modules, ( which theres not much
need for at present, looks like the team have gone through already for
the most part and added remote-id's where possible already ), anything
that inherits 'perl-module.eclass' has a bit of magic, in that neither
SRC_URI or HOMEPAGE is required in the ebuild, and it just gets the
package name from what gentoo is using. We've tried to be as close to
upstream as possible for the ease of maintenance.
However, there are still exception cases, for instance, BioPerl has to
define 'MY_PN' to tell the perl-module eclass to use a different token
( and when this is present, it should be sufficient to say that that
should be the remote-id instead of the package name:
see dev-perl/Moose # an example with neither src_uri or homepage
see sci-biology/bioperl # an example where the package name has
been forced overridden as its changed upstream
But resolving module names is much trickier, its easy-ish to map a
module name to a package using the service, but doing it the other way
round is not so straight forward, as one package can have many
modules, and its common in perl to state dependencies in terms of the
module to require, not the package its in, but there's also often a
defacto "main module".
But I'm myself still working out how to best do that with the service
, so auto-populating a "cpan-module" identifier can be left to later,
its just something I considered "useful" to have metadata wise because
that value is more useful to users.
--
Kent
perl -e "print substr( \"edrgmaM SPA NOcomil.ic\\@tfrken\", \$_ * 3,
3 ) for ( 9,8,0,7,1,6,5,4,3,2 );"
http://kent-fredric.fox.geek.nz
