1 |
On 02/27/2013 07:27 PM, Alexis Ballier wrote: |
2 |
> On Wed, 27 Feb 2013 19:14:38 +0100 |
3 |
> hasufell <hasufell@g.o> wrote: |
4 |
> |
5 |
>> On 02/27/2013 06:58 PM, Alexis Ballier wrote: |
6 |
>>> |
7 |
>>>> The other thing is: |
8 |
>>>> We still have the conflict with eclass-solution vs PM-solution |
9 |
>>>> (multilib-portage) and I propose not to convert ANYTHING else until |
10 |
>>>> that conflict is solved, even if it means a council vote (that's |
11 |
>>>> what I actually think makes sense here). |
12 |
>>>> I understand both sides and somehow find it appealing to have a |
13 |
>>>> quicker solution, but since this could damage years of work on a |
14 |
>>>> portage fork I think we should slow down here. |
15 |
>>> |
16 |
>>> except there _has_ been a discussion: |
17 |
>>> http://article.gmane.org/gmane.linux.gentoo.devel/80330 |
18 |
>>> |
19 |
>>> where, at least for me, it appeared that the eclass solution was the |
20 |
>>> right way and portage-multilib had its defects that could not be |
21 |
>>> solved without such an eclass solution. |
22 |
>> |
23 |
>> I don't even know multilib-portage (think is this way around) that |
24 |
>> detailed myself, but Tommy[D] claims that some of those problems will |
25 |
>> be solved in EAPI=6 and that he is willing to work on the spec. |
26 |
> |
27 |
> What are the problems exactly? I'm likely misinformed, but to me it |
28 |
> seemed there is nothing EAPI-related there. What kind of spec (read: |
29 |
> pms diff) do we need? |
30 |
|
31 |
E.g. that you cannot enabled/disable ABIs on a per-package basis. |
32 |
|
33 |
Feb 26 22:04:07 <hasufell> Tommy[D]: per-package setting is possible as |
34 |
well? |
35 |
Feb 26 22:05:19 <Tommy[D]> hasufell: adding the features to EAPI-6, you |
36 |
can do it per package too, yes |
37 |
|
38 |
Feb 26 22:04:40 <mgorny> Tommy[D]: when is it going to be available in |
39 |
mainstream Gentoo? |
40 |
Feb 26 22:07:45 <Tommy[D]> mgorny: you want it? i could switch my free |
41 |
time into it after being done with another recruitment |
42 |
|
43 |
Feb 26 22:12:02 <hasufell> what is the portage maintainers opinion on |
44 |
this fork? |
45 |
Feb 26 22:14:49 <Tommy[D]> afaik zmedico has no issues with it, also i |
46 |
dont know how close his look on the code itself has been |
47 |
|
48 |
|
49 |
Afaiu this seems to be mainly a PMS thing. And changing PMS is slow and |
50 |
painful, so no wonder people rather want to go for eclass based solutions. |
51 |
|
52 |
> |
53 |
>> The reason I bring this up again is that there was a strong argument |
54 |
>> yesterday in #gentoo-dev, so it seems the situation is NOT clear. |
55 |
> |
56 |
> What are these arguments ? The IRC conspiracy is hard to follow :) |
57 |
> |
58 |
>>> Long story short: portage-multilib does not handle deps needing |
59 |
>>> multilib and deps not needing them. Only packages maintainers know |
60 |
>>> that, you cannot guess it at the PM level. Doing unpack twice, while |
61 |
>>> bearable, is also suboptimal. portage-multilib already disables its |
62 |
>>> multilib support for multilib-enabled packages, thus there is not |
63 |
>>> even a conflict there. |
64 |
>>> |
65 |
>> |
66 |
>> It still does not make sense to work in two different directions, |
67 |
>> imo. I was supporting the eclass idea myself by proposing |
68 |
>> autotools-multilib-minimal.eclass, but I think this should be voted |
69 |
>> on, so we don't duplicate work. |
70 |
> |
71 |
> Again, without a summary of pros and cons and an old discussion that |
72 |
> died leaving the eclass solution as a winner, it is a bit harsh to ask |
73 |
> for a vote. |
74 |
> |
75 |
|
76 |
This is just my own view on this and NOT complete, Tommy[D] will |
77 |
probably have a more complete list what the eclasses currently lack and |
78 |
where they will fail. |
79 |
Mgorny will have a more complete list why multilib-portage is a bad hack. |
80 |
|
81 |
|
82 |
PM level: |
83 |
pro: |
84 |
- one-blow solution, tree-wide |
85 |
- _much_ less modification on ebuilds needed |
86 |
- will be properly documented in the spec, something people can |
87 |
rely on |
88 |
- multilib-portage has years of experience on ABI issues |
89 |
con: |
90 |
- difficult to maintain (please count the number of people who |
91 |
understand portage code) |
92 |
- slower fixes |
93 |
- still no merge into mainline in sight, could very well take another |
94 |
few months, if at all |
95 |
|
96 |
|
97 |
eclass level: |
98 |
pro: |
99 |
- easier to maintain (eclasses are generally easy understandable) |
100 |
- quicker to fix and to extend |
101 |
- solution is NOW available |
102 |
con: |
103 |
- more likely to break stuff as all eclass based solutions, because |
104 |
there are no eclass-APIs and no spec |
105 |
- needs much more modification on ebuilds, especially for weird build |
106 |
systems |
107 |
- not tree-wide, slow per-package migration |