1 |
On Thursday, March 10, 2016 5:40:18 PM CET, Ulrich Mueller wrote: |
2 |
>>>>>> On Thu, 10 Mar 2016, Alexis Ballier wrote: |
3 |
> |
4 |
>> On Wednesday, March 9, 2016 5:37:08 PM CET, Andreas K. Hüttel wrote: |
5 |
>>> 1) All pre-PMS, non-PMS-conformant behaviour should be considered |
6 |
>>> deprecated immediately. |
7 |
>>> 2) We encourage creation of trackers to hunt down and kill pre-PMS, |
8 |
>>> non-PMS-conformant behaviour of ebuilds, eclasses, package managers |
9 |
>>> 3) We introduce a hard deadline when all this should be fixed. |
10 |
> |
11 |
>> You seem to be generalizing to all cases from a very specific one: |
12 |
>> multislot is breaking an important assumption (SLOT being constant) |
13 |
>> and dropping it is not breaking anything. |
14 |
> |
15 |
>> Some examples that would fall under the scope of your proposal: |
16 |
> |
17 |
>> https://bugs.gentoo.org/show_bug.cgi?id=202631 |
18 |
>> -> Needed to comply with other PMS rules on some systems ('patch' |
19 |
>> being GNU patch inside ebuilds, etc.) |
20 |
> |
21 |
> No, it isn't. The package manager is required to ensure that "sed", |
22 |
> "patch", "find" etc. are the GNU versions. This is independent of any |
23 |
> profile.bashrc. If Portage relies on aliases set in profiles instead, |
24 |
> then this is a Portage bug which should be fixed. |
25 |
> |
26 |
> PMS reference: |
27 |
> https://projects.gentoo.org/pms/6/pms.html#x1-12600011.3.1.1 |
28 |
|
29 |
|
30 |
Well, you can go into the debate whether perfectly working and needed |
31 |
behavior predating PMS which EAPI0 was supposed to normalize is a PMS bug |
32 |
or a portage bug, but your point would be much better emphasized with a |
33 |
patch providing the behavior you believe to be correct. It might even be a |
34 |
better solution since profiles aliases do not really work when e.g. |
35 |
building freebsd stuff from a gnu userland. |
36 |
Mind the upgrade path though: If you remove the aliases from profiles, you |
37 |
might not be able to install the portage version providing them. Kind of |
38 |
defeats one of the goals of PMS :) |
39 |
|
40 |
|
41 |
>> https://bugs.gentoo.org/show_bug.cgi?id=203891 |
42 |
>> -> Without this, we'd install a half-broken glibc by default. Any |
43 |
>> deadline would have to take in consideration the time needed to have |
44 |
>> a fixed glibc in stable. |
45 |
> |
46 |
>> (some ocaml stuff are also offenders here, but it is really minor in |
47 |
>> comparison, and I've been trying to move away from the "feature" |
48 |
>> causing the need for it as much as I could) |
49 |
> |
50 |
> We have discussed this at length (on the verge of derailing) in |
51 |
> the bug. Someone has to write a spec, then it can go into EAPI 7. |
52 |
> I actually like mgorny's proposal for a "dostrip" in comment 39. |
53 |
|
54 |
|
55 |
I like it too and consider it a much better solution, but that's not the |
56 |
point. |
57 |
|
58 |
|
59 |
> Until then, ebuilds should neither rely on the variable, nor should |
60 |
> they call functions like prepallstrip that are internal to the package |
61 |
> manager. |
62 |
> |
63 |
> PMS reference: |
64 |
> https://projects.gentoo.org/pms/6/pms.html#x1-14500011.3.3.16 |
65 |
|
66 |
|
67 |
You prefer to install half-broken libc by default rather than fixing PMS, |
68 |
noted. |
69 |
|
70 |
|
71 |
|
72 |
>> https://bugs.gentoo.org/show_bug.cgi?id=573306 |
73 |
>> -> Needed to get cross-compilation (or even ROOT!=/) to work |
74 |
>> properly. (Independtly of PM getting cross-compilation deps |
75 |
>> properly). |
76 |
> |
77 |
> I cannot say much about this one, or about cross-compilation in |
78 |
> general. Reading the bug, I am not too optimistic, though. Seems there |
79 |
> aren't even clear answers to the three simple questions that have been |
80 |
> posed. |
81 |
|
82 |
|
83 |
You should probably read again then: A few additive answers without |
84 |
ping-pong usually means all the parties agree. |
85 |
|
86 |
[...] |
87 |
|
88 |
|
89 |
Alexis. |