| 1 |
Dnia 2013-12-08, o godz. 20:26:44 |
| 2 |
Ciaran McCreesh <ciaran.mccreesh@××××××××××.com> napisał(a): |
| 3 |
|
| 4 |
> On Sun, 8 Dec 2013 21:21:52 +0100 |
| 5 |
> Ulrich Mueller <ulm@g.o> wrote: |
| 6 |
> > > The PMS describes package manager behavior required to support an |
| 7 |
> > > ebuild repository. If I read the PMS correctly, SLOT-less |
| 8 |
> > > dependencies have undefined behavior; this makes it so that if you |
| 9 |
> > > have a different package manager using the same ebuild repository, |
| 10 |
> > > it could interpret the dependencies completely different. |
| 11 |
> > |
| 12 |
> > Nothing undefined here. A dependency without a slot means that all |
| 13 |
> > slot values are acceptable. And all package managers interpret it in |
| 14 |
> > the same way. |
| 15 |
> |
| 16 |
> Actually, Paludis interprets a lack of slot dependency as a "don't |
| 17 |
> know", and assumes that it might be unsafe to switch slots at runtime |
| 18 |
> in these cases. |
| 19 |
|
| 20 |
So we should basically encourage people to explicitly state ':*' when |
| 21 |
they mean that, correct? Because I can tell you that right now many |
| 22 |
people don't use that because they see it as redundant. |
| 23 |
|
| 24 |
-- |
| 25 |
Best regards, |
| 26 |
Michał Górny |
Archives