1 |
On Sun, Apr 20, 2008 at 12:57:28AM +0100, Ciaran McCreesh wrote: |
2 |
> On Sat, 19 Apr 2008 18:38:06 +0200 |
3 |
> "Marijn Schouten (hkBst)" <hkBst@g.o> wrote: |
4 |
> > I don't know what the general use of pkg_preinst is, but in |
5 |
> > pkg_postinst the package itself should be runnable, so its RDEPENDS |
6 |
> > should be installed and usable at this point. So perhaps we should |
7 |
> > define that "usable" means "each of its RDEPENDs is installed and has |
8 |
> > had its pkg_postinst function run". The recursion of that definition |
9 |
> > then comes from the requirement that RDEPENDs should be usable before |
10 |
> > pkg_postinst starts running. |
11 |
> |
12 |
> No good. That prevents RDEPEND <-> RDEPEND cycles from being solved, |
13 |
> and the package manager has to be able to solve that. |
14 |
|
15 |
Not all cycles are solvable w/in the domain of execution which is an |
16 |
unfortunate fact. This loops back into why jstubbs was after use dep |
17 |
cycle breaking (fair bit more fine grained of an issue), a discussion |
18 |
that thus far has been ignored, and the usage of the build flag. |
19 |
|
20 |
|
21 |
> > SRC_UNPACK_DEP="app-arch/unzip" |
22 |
> > SRC_COMPILE_DEP="dev-scheme/bigloo" |
23 |
> > SRC_INSTALL_DEP="" |
24 |
> |
25 |
> Labels are a cleaner solution to this. But again, we're discussing |
26 |
> current EAPIs here. |
27 |
|
28 |
Stop name dropping labels until you tell folk about what labels are. |
29 |
I know, but I'd rather not have the notion "labels solves all" pushed |
30 |
forth w/ out people knowing what it is, please. |
31 |
|
32 |
thanks, |
33 |
~harring |