| 1 |
On Thu, 2006-05-18 at 16:37 +0200, Paul de Vrieze wrote: |
| 2 |
> On Thursday 18 May 2006 16:03, Stephen Bennett wrote: |
| 3 |
> > On Thu, 18 May 2006 15:34:28 +0200 |
| 4 |
> > |
| 5 |
> > Paul de Vrieze <pauldv@g.o> wrote: |
| 6 |
> > > Requiring duplication of profiles for every package manager. |
| 7 |
> > |
| 8 |
> > It requires duplicating nothing. This is exactly why we have cascading |
| 9 |
> > profiles. |
| 10 |
> |
| 11 |
> Cascading profiles form a tree with N nodes. Some of these nodes are |
| 12 |
> abstract in the sense that they are not directly usable. Say that leaves |
| 13 |
> M possible profiles. To have paludis be on par with portage, each of |
| 14 |
> these M profiles would have a leaf added for paludis. The same holds for |
| 15 |
> pkgcore and for any other package manager. This would mean that we have |
| 16 |
> N+2M profiles. With a paludis and pkgcore toplevel profile this would |
| 17 |
> even be worse and amount to approximately 3N profiles. |
| 18 |
> |
| 19 |
> In the leaf version, all M paludis specific profiles are equal. |
| 20 |
|
| 21 |
But Paludis supports multiple inheritance. Would it be feasible to have |
| 22 |
Paludis users create /etc/make.profile as a directory, |
| 23 |
with /etc/make.profile/parent inheriting from both their chosen |
| 24 |
gentoo-x86 profile and a profile in the paludis tree? |
| 25 |
|
| 26 |
(I guess this looks like offering a technical solution to a political |
| 27 |
problem... sorry about that.) |
| 28 |
|
| 29 |
Ed |
| 30 |
|
| 31 |
-- |
| 32 |
gentoo-dev@g.o mailing list |
Archives