1 |
On Fri, Jul 20, 2012 at 2:54 AM, Ciaran McCreesh |
2 |
<ciaran.mccreesh@××××××××××.com> wrote: |
3 |
> On Thu, 19 Jul 2012 18:34:41 -0400 |
4 |
> Mike Gilbert <floppym@g.o> wrote: |
5 |
>> On Thu, Jul 19, 2012 at 5:13 PM, Zac Medico <zmedico@g.o> |
6 |
>> wrote: |
7 |
>> > On 07/19/2012 06:14 AM, Ralph Sennhauser wrote: |
8 |
>> >> Could be that Portage re-exports a sanitized |
9 |
>> >> LINGUAS tough, but I doubt it. |
10 |
>> > |
11 |
>> > Portage does sanitize it if there are any linguas_* flags in IUSE, |
12 |
>> > otherwise it lets the variable pass through without sanitizing it. |
13 |
>> |
14 |
>> That's good; we definitely don't want to "sanitize" it if there are no |
15 |
>> linuguas_* flags in IUSE. This would break LINUGUAS support for many |
16 |
>> autotools/gettext based packages, where the autotools code parses |
17 |
>> LINGUAS directly and the ebuild does nothing with it. |
18 |
> |
19 |
> If there aren't any linguas_* flags in IUSE, LINGUAS should be empty, |
20 |
> and will be in future EAPIs. Without that, USE dependencies on |
21 |
> USE_EXPAND variables don't work. |
22 |
|
23 |
How do you figure that? |
24 |
|
25 |
The current portage behavior works well enough; if linugas_* is in |
26 |
IUSE, LINGUAS is treated as a USE_EXPAND, and use-deps should work |
27 |
fine. |
28 |
|
29 |
Otherwise, it is treated just like a normal environment variable, and |
30 |
portage doesn't touch it. |
31 |
|
32 |
For most gettext packages, there is absolutely no reason that the |
33 |
ebuild maintainer should have to keep track of every translation |
34 |
available in a given package across version bumps. If you change this |
35 |
behavior in a future EAPI, you will break this. |