1 |
On Wed, 27 Feb 2013 22:08:45 +0100 |
2 |
Thomas Sachau <tommy@g.o> wrote: |
3 |
|
4 |
> Alexis Ballier schrieb: |
5 |
> > On Wed, 27 Feb 2013 18:10:30 +0100 |
6 |
> > hasufell <hasufell@g.o> wrote: |
7 |
> > |
8 |
> >> The other thing is: |
9 |
> >> We still have the conflict with eclass-solution vs PM-solution |
10 |
> >> (multilib-portage) and I propose not to convert ANYTHING else until |
11 |
> >> that conflict is solved, even if it means a council vote (that's |
12 |
> >> what I actually think makes sense here). |
13 |
> >> I understand both sides and somehow find it appealing to have a |
14 |
> >> quicker solution, but since this could damage years of work on a |
15 |
> >> portage fork I think we should slow down here. |
16 |
> > |
17 |
> > except there _has_ been a discussion: |
18 |
> > http://article.gmane.org/gmane.linux.gentoo.devel/80330 |
19 |
> > |
20 |
> > where, at least for me, it appeared that the eclass solution was the |
21 |
> > right way and portage-multilib had its defects that could not be |
22 |
> > solved without such an eclass solution. |
23 |
> > Long story short: portage-multilib does not handle deps needing |
24 |
> > multilib and deps not needing them. Only packages maintainers know |
25 |
> > that, you cannot guess it at the PM level. Doing unpack twice, while |
26 |
> > bearable, is also suboptimal. portage-multilib already disables its |
27 |
> > multilib support for multilib-enabled packages, thus there is not |
28 |
> > even a conflict there. |
29 |
> |
30 |
> So you discussed with mgorny (who does not like multilib-portage) and |
31 |
> not me and then assume that all details have been written in |
32 |
> there? :-) |
33 |
|
34 |
It is probably rather that mgorny's approach is much more transparent: |
35 |
He sends his ideas to the ml, ask for comments, etc. while |
36 |
the multilib-portage approach is not clear to anybody. "All details" |
37 |
are, in fact, what I understood from the list of commands and |
38 |
pseudo-algorithm you sent as a spec some time ago :) |
39 |
|
40 |
[...] |
41 |
> Anyway, in short: the current implementation does add dependencies so |
42 |
> that all dependencies need the same ABIs enabled as the package you |
43 |
> want. If we move the features into a future EAPI, we can of course |
44 |
> drop this and leave the dependency part to the ebuild maintainer. |
45 |
|
46 |
How do you achieve that currently? We _do_ need to leave the dependency |
47 |
part to the ebuild maintainer: How will you achieve that? |
48 |
cat/pkg[$MULTIlIB_USE_DEP] ? :) I don't think we need a future EAPI for |
49 |
this. If you spec it, do we need a new EAPI everytime we want to |
50 |
introduce a new ABI? |
51 |
|
52 |
> > On the other hand, Michal has been doing the work and got things |
53 |
> > done when portage-multilib has never reached mainline after several |
54 |
> > years of development. So, while breaking the tree like the freetype |
55 |
> > case is really bad, please do not use this for killing his efforts, |
56 |
> > esp. when it is now masked. |
57 |
> |
58 |
> If you did not know it: anyone can add an eclass, while adding new |
59 |
> features via package manager requires a new EAPI. |
60 |
> I have written about it on this list for many months, if not years. |
61 |
> And every time i solved a request, a new one was raised. And you want |
62 |
> to blaim me for multilib-portage not reaching the main tree? |
63 |
|
64 |
I don't blame multilib-portage for not reaching mainline. From what |
65 |
I've seen, the only needed change I've understood is changing how the |
66 |
ebuild phases are called. The rest is obscure and not precise enough to |
67 |
me, unfortunately. |
68 |
You always claimed that multilib-portage works on the current tree, so |
69 |
I do not see what eapi part you need, just change your pm to do |
70 |
multilib and be done, no ? |
71 |
|
72 |
> Anyway, if there is a sane and easy to use eclass created, which does |
73 |
> just multilib and does it properly for all multilib arches also |
74 |
> supporting per ABI headers and binaries, i am not opposed against it. |
75 |
|
76 |
Good. |
77 |
|
78 |
> I just see issues the way a work-in-progress is pushed into the main |
79 |
> tree without prior discussion and additional hacks for issues |
80 |
> (freetype headers) forcing other devs to do more work instead of |
81 |
> asking for another solution not needing any additional work for |
82 |
> depending packages. |
83 |
|
84 |
He sent all is changes over the ml for review. Everybody was happy, he |
85 |
committed them. Nothing wrong here. The freetype issue is different, |
86 |
and IMHO the solution is wrong there. Let me check bgo and file a bug |
87 |
if not. |
88 |
|
89 |
Alexis. |