1 |
Le samedi 23 juin 2012 à 21:37 +0100, Ciaran McCreesh a écrit : |
2 |
> On Sat, 23 Jun 2012 22:36:14 +0200 |
3 |
> Marien Zwart <marienz@g.o> wrote: |
4 |
> > On za, 2012-06-23 at 17:08 +0100, Ciaran McCreesh wrote: |
5 |
> > > > Is it that Paludis installs a newer SLOT even if a reverse |
6 |
> > > dependency |
7 |
> > > > explicitly requests another SLOT? Sounds like a bug to me. |
8 |
> > > |
9 |
> > > No, it's that if a user requests a "complete" resolution, Paludis |
10 |
> > > installs the newest version of things that it can. Extensive |
11 |
> > > consultation with users has shown that this is a good behaviour, |
12 |
> > > except |
13 |
> > > in the small number of situations that have recently arisen where |
14 |
> > > people are doing weird things with versions and slots. |
15 |
> > |
16 |
> > It surprises me that this behavior is normally desirable for packages |
17 |
> > where all dependencies (including any in the world set or the like) |
18 |
> > are slotted. |
19 |
> |
20 |
> Think || ( a:3 a:2 ). |
21 |
> |
22 |
I would say this is not possible with gtk+ |
23 |
|
24 |
To build a gtk+3 app, you need gtk+3 based libs only, same for gtk+2. |
25 |
Mixing will not work because of symbols conflict iirc. |
26 |
|
27 |
Anyway, I think that we got off track on the basics of the problem. The |
28 |
problem is that you cannot have two ebuilds of the same ${CAT}/${PN} |
29 |
with the same version simply because the files would have the same name. |
30 |
Adding a new property or whatever does not solve this problem unless we |
31 |
propose a way of naming such ebuilds to start with, right ? |
32 |
|
33 |
-- |
34 |
Gilles Dartiguelongue <eva@g.o> |
35 |
Gentoo |