1 |
"Santiago M. Mola" <coldwind@g.o> said: |
2 |
> On Sat, Apr 4, 2009 at 2:59 PM, Thomas Sachau <tommy@g.o> wrote: |
3 |
> > Hi folks, |
4 |
> > |
5 |
> > |
6 |
> > i would like to hear about other opinions about real multilib support |
7 |
> > within our tree and package managers. From what i know, there are |
8 |
> > mainly 2 different ideas: |
9 |
> |
10 |
> The proposals are not exactly these. |
11 |
> |
12 |
> 1. Make package managers multilib-aware [1][2]. |
13 |
> |
14 |
> Package managers would be able to have a default ABI (say, x86_64) and |
15 |
> optional ones (x86). Everything would be built for the default ABI, |
16 |
> and the package manager could build things for optional ABIs on an as |
17 |
> needed basis. That is, if I install a 32bit binary package, the |
18 |
> package manager will build any 32bit libraries it needs automatically. |
19 |
> |
20 |
> Package managers will have to expose to ebuilds a mechanism to iterate |
21 |
> over enabled ABIs and build anything needed for each one. |
22 |
> |
23 |
> Pros: |
24 |
> - Any package can be made multilib aware, getting rid of the emul-* |
25 |
> packages. - 32bit libraries are built automatically and as needed. |
26 |
> - This system can be extended to support other kind of ABIs. Making it |
27 |
> possible to build packages for various versions of Python/GHC/etc |
28 |
> simultaneously. |
29 |
> |
30 |
> Cons: |
31 |
> - Needs to be implemented on the PM-side and needs a new EAPI. |
32 |
> |
33 |
> 2. Implement multilib on the ebuild level. |
34 |
> |
35 |
> For amd64, this would mean adding a 'lib32' USE flag to every multilib |
36 |
> ebuild, and use it for building 32bit libs as needed. |
37 |
> |
38 |
> Pros: |
39 |
> - Any package can be made multilib aware, getting rid of the emul-* |
40 |
> packages. - Doesn't need PM changes. |
41 |
> |
42 |
> Cons: |
43 |
> - Package manager won't be multilib-aware, so it won't be able to |
44 |
> build 32bit libraries automatically and as needed. |
45 |
> - Users will have to enable 'lib32' USE flag manually for every |
46 |
> library they needed. Enabling 'lib32' by default is not an option |
47 |
> since it would build tons of unneeded 32bit libraries for every user. |
48 |
|
49 |
reading the proposals so far, it sounds to me that only the one that |
50 |
requires pms changes would qualify as 'real multilib support'. |
51 |
|
52 |
> |
53 |
> |
54 |
> [1] http://dev.exherbo.org/~pioto/abi-ideas.html |
55 |
> [2] http://bugs.gentoo.org/show_bug.cgi?id=145737 |
56 |
> |
57 |
> Regards, |