1 |
Pacho Ramos schrieb: |
2 |
> El dom, 23-09-2012 a las 13:52 +0200, Thomas Sachau escribió: |
3 |
>> Pacho Ramos schrieb: |
4 |
>>> El dom, 23-09-2012 a las 11:56 +0200, Michał Górny escribió: |
5 |
>>>> On Sun, 23 Sep 2012 11:07:30 +0200 |
6 |
>>>> Thomas Sachau <tommy@g.o> wrote: |
7 |
>>>> |
8 |
>>>>> Matt Turner schrieb: |
9 |
>>>>>> On Sat, Sep 22, 2012 at 2:24 PM, Michał Górny <mgorny@g.o> wrote: |
10 |
>>>>>>> It is a simple eclass using autotools out-of-source builds to build |
11 |
>>>>>>> packages for multiple ABIs when multilib is supported. |
12 |
>>>>>> |
13 |
>>>>>> Thanks a lot, Michał! This looks good to me. |
14 |
>>>>>> |
15 |
>>>>>>> Use case: xorg packages, ask Matt. |
16 |
>>>>>> |
17 |
>>>>>> So the idea is that users want up-to-date 32-bit drivers for games and |
18 |
>>>>>> WINE. The emul- packages aren't a very good solution for a number of |
19 |
>>>>>> reasons. |
20 |
>>>>>> |
21 |
>>>>>> I'd like to add multilib USE flags to Mesa and thus its dependencies. |
22 |
>>>>>> I realized that almost everything in x11-libs/ could be converted very |
23 |
>>>>>> easily, which would allow us to get rid of emul-linux-x86-xlibs in |
24 |
>>>>>> addition to emul-linux-x86-opengl. |
25 |
>>>>>> |
26 |
>>>>>> |
27 |
>>>>> |
28 |
>>>>> This looks like a shortened duplication of a subset of multilib-portage |
29 |
>>>>> features. While this wont hurt multilib-portage (since it does exclude |
30 |
>>>>> most actions on ebuilds with USE=multilib), it will mean a rewrite for |
31 |
>>>>> many ebuilds, which then again need another rewrite (or more likely |
32 |
>>>>> revert), when multilib-portage is accepted in a future EAPI. |
33 |
>>>> |
34 |
>>>> s/when/if/ |
35 |
>>>> |
36 |
>>>>> So i would prefer some help/support with multilib-portage to get it |
37 |
>>>>> accepted sooner, instead of this additional workaround for a subset of |
38 |
>>>>> packages. |
39 |
>>>> |
40 |
>>>> I prefer the simpler solution. |
41 |
>>>> |
42 |
>>>>> P.S.: I know, that users, who want up-to-date 32bit drivers for games |
43 |
>>>>> and wine do use multilib-portage, so we already have a working solution |
44 |
>>>>> for this issue. |
45 |
>>>> |
46 |
>>>> They will no longer have to do that. |
47 |
>>>> |
48 |
>>> |
49 |
>>> I would prefer if eclass way could be extended to packages not using |
50 |
>>> autotools too, otherwise, we will still need emul packages for, for |
51 |
>>> example, qt libs. If that would be possible via eclass, maybe |
52 |
>>> multilib-portage wouldn't be needed but, if not, we will still need it |
53 |
>>> and, then, would be nice if this inclussion for autotools packages |
54 |
>>> wouldn't cause more problems to get the "strong" solution land in the |
55 |
>>> "near" future :/ |
56 |
>>> |
57 |
>>> The simpler solution (eclass) looks fine to me, but it would need to be |
58 |
>>> extended to more packages than autotools based ones to let it replace |
59 |
>>> portage-multilib/emul packages |
60 |
>>> |
61 |
>> |
62 |
>> you mean something like this one? |
63 |
>> |
64 |
>> https://github.com/sjnewbury/multilib-overlay/blob/80c9fd20cfd05481ac19edcadd56ad5e578a8930/eclass/multilib-native.eclass |
65 |
>> |
66 |
>> That was the original eclass allowing cross-compile support, but |
67 |
>> required ebuilds to inherit it. multilib-portage is based on this, but |
68 |
>> does not require to modify the ebuilds themselves. |
69 |
>> |
70 |
> |
71 |
> Yes, that is what I meant... but I don't find hard to modify ebuilds to |
72 |
> inherit it :/ |
73 |
> |
74 |
|
75 |
It is not hard by itself to inherit an eclass. There is just the |
76 |
limitation, that occurs with an eclass, e.g.: |
77 |
|
78 |
-the one from mgorny only does it for autotools based ebuilds only and |
79 |
even there only for libraries, headers and binaries are not done. While |
80 |
one may create the same for cmake based ones, those are still only for a |
81 |
subset of packages, since not all do use autotools/cmake or are |
82 |
satisfied with a libs only solution |
83 |
-the multilib-native eclass does extend it way more, but still has its |
84 |
limitations, e.g. what happens with a new target ABI (like x32 on the |
85 |
amd64 profile) or how are dependencies handled? |
86 |
|
87 |
multilib-portage is the answer to those limitations, since it does work |
88 |
for any target with multilib cross-compile support, can handle things |
89 |
like the dependencies internally and can even work on not |
90 |
prepared/modified ebuilds. |
91 |
|
92 |
So before i invest even more time in getting this official, we should |
93 |
better get to some conclusion, if we either go the route with eclass |
94 |
based multilib cross-compile support with its limitations or if we move |
95 |
this up to the package manager level. |
96 |
|
97 |
-- |
98 |
|
99 |
Thomas Sachau |
100 |
Gentoo Linux Developer |