1 |
Dnia 31 lipca 2017 10:17:56 CEST, Zac Medico <zmedico@g.o> napisał(a): |
2 |
>On Wed, Jul 26, 2017 at 1:44 PM, Michał Górny <mgorny@g.o> |
3 |
>wrote: |
4 |
>> On śro, 2017-07-26 at 11:29 -0700, Zac Medico wrote: |
5 |
>>> You might think that, but I just checked our profiles and the only |
6 |
>one |
7 |
>>> that currently has it enabled is profiles/arch/sparc/make.defaults. |
8 |
>In |
9 |
>>> profiles/targets/developer/make.defaults there's a reference to this |
10 |
>>> bug that's been open for 5 years now: |
11 |
>>> |
12 |
>>> https://bugs.gentoo.org/show_bug.cgi?id=424423 |
13 |
>> |
14 |
>> To be honest, I started rewriting multilib-strict to be sane but lost |
15 |
>> motivation pretty early. Feel free to take it up, it shouldn't be |
16 |
>hard. |
17 |
>> Just horribly boring. |
18 |
>> |
19 |
>> Basically, it should iterate over all ABIs, get libdirs for them |
20 |
>> and look for libraries *directly* in those libdirs. Anything below |
21 |
>first |
22 |
>> level can be ignored as unclear with a good hit rate anyway. |
23 |
> |
24 |
>Assuming that MULTILIB_STRICT_DIRS contains a reasonable list of |
25 |
>libdirs for the profile, can't we get that behavior by adding |
26 |
>-maxdepth 1 to the find call inside |
27 |
>bin/install-qa-check.d/80multilib-strict? If so, then why haven't we |
28 |
>done this already? |
29 |
|
30 |
Probably because nobody touched it. But yeah, killing the exempt first would be a reasonable first step. |
31 |
|
32 |
|
33 |
-- |
34 |
Best regards, |
35 |
Michał Górny (by phone) |