1 |
On Sun, 23 Sep 2012 13:03:56 +0200 |
2 |
Pacho Ramos <pacho@g.o> wrote: |
3 |
|
4 |
> El dom, 23-09-2012 a las 12:40 +0200, Michał Górny escribió: |
5 |
> > On Sun, 23 Sep 2012 12:33:58 +0200 |
6 |
> > Pacho Ramos <pacho@g.o> wrote: |
7 |
> > |
8 |
> > > El dom, 23-09-2012 a las 11:56 +0200, Michał Górny escribió: |
9 |
> > > > On Sun, 23 Sep 2012 11:07:30 +0200 |
10 |
> > > > Thomas Sachau <tommy@g.o> wrote: |
11 |
> > > > |
12 |
> > > > > Matt Turner schrieb: |
13 |
> > > > > > On Sat, Sep 22, 2012 at 2:24 PM, Michał Górny <mgorny@g.o> wrote: |
14 |
> > > > > >> It is a simple eclass using autotools out-of-source builds to build |
15 |
> > > > > >> packages for multiple ABIs when multilib is supported. |
16 |
> > > > > > |
17 |
> > > > > > Thanks a lot, Michał! This looks good to me. |
18 |
> > > > > > |
19 |
> > > > > >> Use case: xorg packages, ask Matt. |
20 |
> > > > > > |
21 |
> > > > > > So the idea is that users want up-to-date 32-bit drivers for games and |
22 |
> > > > > > WINE. The emul- packages aren't a very good solution for a number of |
23 |
> > > > > > reasons. |
24 |
> > > > > > |
25 |
> > > > > > I'd like to add multilib USE flags to Mesa and thus its dependencies. |
26 |
> > > > > > I realized that almost everything in x11-libs/ could be converted very |
27 |
> > > > > > easily, which would allow us to get rid of emul-linux-x86-xlibs in |
28 |
> > > > > > addition to emul-linux-x86-opengl. |
29 |
> > > > > > |
30 |
> > > > > > |
31 |
> > > > > |
32 |
> > > > > This looks like a shortened duplication of a subset of multilib-portage |
33 |
> > > > > features. While this wont hurt multilib-portage (since it does exclude |
34 |
> > > > > most actions on ebuilds with USE=multilib), it will mean a rewrite for |
35 |
> > > > > many ebuilds, which then again need another rewrite (or more likely |
36 |
> > > > > revert), when multilib-portage is accepted in a future EAPI. |
37 |
> > > > |
38 |
> > > > s/when/if/ |
39 |
> > > > |
40 |
> > > > > So i would prefer some help/support with multilib-portage to get it |
41 |
> > > > > accepted sooner, instead of this additional workaround for a subset of |
42 |
> > > > > packages. |
43 |
> > > > |
44 |
> > > > I prefer the simpler solution. |
45 |
> > > > |
46 |
> > > > > P.S.: I know, that users, who want up-to-date 32bit drivers for games |
47 |
> > > > > and wine do use multilib-portage, so we already have a working solution |
48 |
> > > > > for this issue. |
49 |
> > > > |
50 |
> > > > They will no longer have to do that. |
51 |
> > > > |
52 |
> > > |
53 |
> > > I would prefer if eclass way could be extended to packages not using |
54 |
> > > autotools too, otherwise, we will still need emul packages for, for |
55 |
> > > example, qt libs. If that would be possible via eclass, maybe |
56 |
> > > multilib-portage wouldn't be needed but, if not, we will still need it |
57 |
> > > and, then, would be nice if this inclussion for autotools packages |
58 |
> > > wouldn't cause more problems to get the "strong" solution land in the |
59 |
> > > "near" future :/ |
60 |
> > > |
61 |
> > > The simpler solution (eclass) looks fine to me, but it would need to be |
62 |
> > > extended to more packages than autotools based ones to let it replace |
63 |
> > > portage-multilib/emul packages |
64 |
> > |
65 |
> > Qt uses cmake, doesn't it? |
66 |
> > |
67 |
> > I don't mind having cmake-multilib as well? We could probably move |
68 |
> > the foreach_abi() function to some more common eclass, like multilib |
69 |
> > eclass. |
70 |
> > |
71 |
> |
72 |
> Looks interesting, yes, it uses cmake. I guess we would need to move all |
73 |
> packages needing 32bits libs to this eclasses. Are you sure wouldn't be |
74 |
> better to try to go to an "upper" level like Alexis Ballier suggested |
75 |
> some messages ago?: |
76 |
> "On Sat, 22 Sep 2012 23:24:46 +0200 |
77 |
> Michał Górny <mgorny@g.o> wrote: |
78 |
> |
79 |
> > It is a simple eclass using autotools out-of-source builds to build |
80 |
> > packages for multiple ABIs when multilib is supported. |
81 |
> > |
82 |
> |
83 |
> to some extent, can't you do the same by unpacking twice to different |
84 |
> $S and calling src_prepare/compile/install instead of their |
85 |
> autotools-utils counterpart with tweaked $S so that it works with almost |
86 |
> every ebuild ? |
87 |
> |
88 |
> -- this really starts to resemble multilib portage :)" |
89 |
> |
90 |
> That would be better as there are a ton of ebuilds not inheritting |
91 |
> autotools-utils.eclass even being autotools based (think for example in |
92 |
> gnome packages or many others) |
93 |
|
94 |
You could fix those ebuilds to inherit it too ;). autotools-utils was |
95 |
especially designed to use out-of-source builds for multilib |
96 |
in the future. |
97 |
|
98 |
I'm afraid the 'upper level' is technically impossible without either |
99 |
going into PM itself (which means waiting for EAPI 6 at least |
100 |
and getting some scary logic into it) or reinventing the phases alike |
101 |
ruby-ng/python-distutils-ng. Well, the latter may be useful to some |
102 |
degree; still, it would require each ebuild to redefine all phases. |
103 |
|
104 |
-- |
105 |
Best regards, |
106 |
Michał Górny |