El lun, 11-06-2012 a las 18:08 +0100, Ciaran McCreesh escribió:
> On Mon, 11 Jun 2012 13:15:40 +0100
> Nirbheek Chauhan <email@example.com> wrote:
> > On Sun, Jun 10, 2012 at 9:49 PM, Ciaran McCreesh
> > <ciaran.mccreesh@...> wrote:
> > > On Sun, 10 Jun 2012 21:45:27 +0100
> > > Nirbheek Chauhan <firstname.lastname@example.org> wrote:
> > >> It's a simple workaround for the lack of proper ebuild namespacing
> > >> on the basis of slots.
> > >>
> > >> So, till we have that, this works pretty well. :)
> > >
> > > Until you have that, or something else designed to do what you want,
> > > don't come up with some disgusting hack.
> > So the PMS process should be a bottleneck to getting software out to
> > users? I think that's counter-productive.
> There is no PMS bottleneck. There is a Portage bottleneck, and there is
> a "figuring out how to ensure new features don't interact badly with
> either old features or stupid hacks people have done". Abuse of the
> kind under discussion is a large contributor to both of those
> > Our goal here is not to facilitate package manager development but to
> > package and distribute software to users.
> No, your goal is to provide a distribution. Gentoo has repeatedly shot
> itself in the foot, leg, groin etc by favouring short-term hacks over a
> well thought out, validated, self-enforcing design. Right now nearly
> all of the package manager work is on paying off previously incurred
> technical debt, and in the mean time you're busy adding to it.
The problem here is that we (or, at least, I) are a bit unsure about how
this could be handled better and, then, try to use that better way in
the future. If you (or any) have some suggestion, it would be nice :)