1 |
Am Sonntag, den 05.04.2009, 10:18 +0200 schrieb Thomas Sachau: |
2 |
> Mike Frysinger schrieb: |
3 |
> > On Saturday 04 April 2009 08:59:22 Thomas Sachau wrote: |
4 |
> >> i would like to hear about other opinions about real multilib support |
5 |
> >> within our tree and package managers. From what i know, there are mainly 2 |
6 |
> >> different ideas: |
7 |
> >> |
8 |
> >> 1. Do the main stuff in the package manager (e.g. if the ARCH is amd64 and |
9 |
> >> the package has x86 keyword, the package manager adds a lib32 useflag, |
10 |
> >> which would additionally install the 32bit variant of that package together |
11 |
> >> with the normal 64bit install). |
12 |
> >> |
13 |
> >> pro: -much lesser work for package maintainers |
14 |
> >> |
15 |
> >> contra: -needs addition in PMS and support in the pms, which will need some |
16 |
> >> work on their side |
17 |
> > |
18 |
> > get a *working* implementation first and *then* worry about specing it. once |
19 |
> > you have something running with portage, the spec should fall naturally out. |
20 |
> > previous multilib methods attempted to spec things out without any real code |
21 |
> > and they've all just died. |
22 |
> > |
23 |
> >> 2. Do the main stuff in the ebuilds themselves (e.g. an additional eclass |
24 |
> >> multilib-native.eclass, any ebuild with 32bit support would then need |
25 |
> >> adaption and of course inheriting that eclass) |
26 |
> > |
27 |
> > this is dead end and useless overhead, and i would reject it from any core |
28 |
> > package someone would try to merge. |
29 |
> > -mike |
30 |
> |
31 |
> |
32 |
> From what i got until now, it seems that all answers prefer this option, so i would like to move |
33 |
> forward and create some aggreement on how this should look like or some implementation that at least |
34 |
> the majority can accept. With this, i would also like to see any changes that need an EAPI to get |
35 |
> into EAPI-3. |
36 |
No. Won't happen. |