1 |
On Wednesday 25 August 2004 21:05, Stephen P. Becker wrote: |
2 |
> I think that precedent has already been set with linux26-headers |
3 |
> unfortunately. |
4 |
> |
5 |
|
6 |
Doesn't mean it shouldn't be reversed... |
7 |
|
8 |
> -Steve |
9 |
> |
10 |
> Ciaran McCreesh wrote: |
11 |
> > On Wed, 25 Aug 2004 12:40:20 -0700 Greg KH <gregkh@g.o> wrote: |
12 |
> > | > gentoo-sources24 |
13 |
> > | > gentoo-sources26 |
14 |
> > | |
15 |
> > | I don't have a problem with this, but the dependancy stuff might not |
16 |
> > | work out properly for some odd kernel-based userspace packages. |
17 |
> > | |
18 |
> > | Anyone object to this? |
19 |
> > |
20 |
> > Well, it's pretty nasty... Part of the idea of SLOTs is that we never |
21 |
> > need to include version numbers in packages... In fact our docs [1] even |
22 |
> > |
23 |
> > say: |
24 |
> >>Most distributions and ports systems tend to have a "freetype" package |
25 |
> >>for freetype 1.x and "freetype2" for 2.x. We consider this approach a |
26 |
> >>sign of a fundamentally broken package management system. |
27 |
> > |
28 |
> > Do we really want to admit that our package manager is broken? |
29 |
> > |
30 |
> > [1]: http://www.gentoo.org/doc/en/portage-manual.xml |
31 |
> |
32 |
|
33 |
-- |
34 |
Tom Wesley <tom@×××××.org> |