Re: [gentoo-dev] [RFC] Policy for migrating library consumers to subslots

[Kent Fredric]

On 27 September 2013 06:00, Kent Fredric <kentfredric@×××××.com> wrote:

>
> ie: every time there is a new iteration of Foo, one must iterate all the
> virtuals that relate to it, and vet their need to be updated, and every
> single update you do is just something thats easy to get wrong, when it
> always seems to make so much more sense to say "X provides Y", especially
> in the case where X provides a very large list of Y ( virtual/perl-* is a
> maintenance nightmare )
>

Also, can I make a suggestion that if we were to bring back PROVIDES=, that
PROVIDES *NOT* be permitted to provide arbitrary $CAT/$P ?

My very rusty memory of when I last saw that saw that as one sticking point.

It seems to me that PROVIDES would be more useful in a different hierarchy,
ie:
PROVIDES="cpan-module/CLASS::MOP/1.6" or something.


-- 
Kent