1 |
Stephen Bennett wrote: |
2 |
|
3 |
>As those who hang around in the mysterious realms of Portage development |
4 |
>may know, there's some feeling around that the current system of virtual |
5 |
>packages has some serious limitations. The currently-proposed |
6 |
>alternative (as discussed previously, notably in |
7 |
>http://thread.gmane.org/gmane.linux.gentoo.devel/18922), involves a |
8 |
>system of metapackages. These would essentially consist of a |
9 |
>non-installable ebuild that consists entirely of a set of dependency |
10 |
>information. Once the dependencies for the metapackage are satisfied, |
11 |
>it's considered to be installed, and packages depending on it can go |
12 |
>ahead and be built. |
13 |
> |
14 |
>This approach brings several advantages over the current system, |
15 |
>particularly: |
16 |
>- Allowing one version of a package to provide a different version of a |
17 |
>virtual, where these are necessary. |
18 |
>- Fixing the screwup with .51's virtual handling whereby gcc-2.95.x has |
19 |
>PROVIDE="sys-apps/texinfo", a package depends on >=texinfo-4.6, so |
20 |
>portage tries to install >=gcc-4.6. |
21 |
>- Provides, in one easily accessible place, a list of package that could |
22 |
>be used to satisfy the dependency. This has advantages for speed (no |
23 |
>searching the tree for PROVIDEs) and for user-friendliness. |
24 |
> |
25 |
>Anyway, with portage development as it is now, this got brought up |
26 |
>again, and the current state of the GLEP can be found at |
27 |
>http://dev.gentoo.org/~spb/metapkg-glep.txt. Comments/suggestions/flames |
28 |
>welcome. |
29 |
> |
30 |
> |
31 |
> |
32 |
As long as you keep the PROVIDES line in the ebuild. Assuming repoman |
33 |
does full tree check when commiting it can make sure that all packages |
34 |
that PROVIDE virtual also exist in the metapackage. That would be my |
35 |
only concern, people forgetting to add lines to the correct files :) |
36 |
|
37 |
>-- |
38 |
>gentoo-dev@g.o mailing list |
39 |
> |
40 |
> |
41 |
-- |
42 |
gentoo-dev@g.o mailing list |