1 |
On Wed, 2003-07-02 at 20:36, Masatomo Nakano wrote: |
2 |
> I suggest that we stop using 'cjk' USE flag, which is to support |
3 |
> Chinese/Japanese/Korean, and we add new flags(lang_cn,lang_ja,lang_kr,....,m17n). |
4 |
|
5 |
It would probably be lang_zh for chinese, to match the locale definition |
6 |
in glibc. I've got a few questions: |
7 |
|
8 |
1. What is the difference between m17n (multilingualisation) and i18n |
9 |
(internationalisation)? Maybe we should use i18n instead of m17n since |
10 |
it seems more widely used, especially if it means the same thing. |
11 |
|
12 |
2. If it is only chinese, japanese and korean, maybe we should just use |
13 |
zh, ja, kr as the USE flags. We have /usr/portage/use.desc to give it |
14 |
the right description. Is there a reason to have the "lang_" prefix in |
15 |
the USE flag? |
16 |
|
17 |
> There are two reasons for this. |
18 |
> |
19 |
> 1. USE flag related to countries is only 'cjk' at the moment. But we |
20 |
> need other country flag(ex. Russia). |
21 |
> |
22 |
> 2. It is impossible to apply 1USE flag among two or more |
23 |
> countries because most of localization patches are conflict each other. |
24 |
|
25 |
I agree that there are other countries that may have their own |
26 |
localisation patches. However, some of the cjk patches do apply to all |
27 |
double-byte character and kanji compatibility within applications. With |
28 |
splitting the cjk USE flag, this means for those patches we'll likely |
29 |
have to test for all 3 zh,ja,kr USE flags. |
30 |
|
31 |
With the conflicting patches, it should be up to the maintainer of the |
32 |
package to make sure these patches don't conflict with each other with |
33 |
different USE flag combinations. I don't think introducing more USE |
34 |
flags will solve that problem, in fact it could potentially make it |
35 |
worse. |
36 |
|
37 |
With that said, I can see a number of advantages with having separate |
38 |
language flags because there maybe chinese specific or japanese specific |
39 |
patches that need to be installed. Or for instance, a chinese user would |
40 |
not need to install japanese and korean fonts, and vice versa. |
41 |
|
42 |
I agree with Nakano that the cjk/zh/ja/kr use flags should not be |
43 |
confused with the LANGS discussion about selectively installing |
44 |
translations (.po/.mo). This concerns mainly with patches and configure |
45 |
options to specifically support XIM and double-byte characters. |
46 |
|
47 |
My vote would be to _keep_ the "cjk" flag for instances where a patch or |
48 |
particular configure option would benefit all double-byte character |
49 |
users and then add the "zh", "ja" and "kr" USE flags as additional ones |
50 |
where a particular patch/configure option is specific to only one |
51 |
locale. |
52 |
|
53 |
How do people think about that? |
54 |
|
55 |
Cheers, |
56 |
-- |
57 |
Alastair 'liquidx' Tse |
58 |
>> Gentoo Developer |
59 |
>> http://www.liquidx.net/ | http://cvs.gentoo.org/~liquidx/ |
60 |
>> GPG Key : http://cvs.gentoo.org/~liquidx/liquidx_gentoo_org.asc |
61 |
>> FingerPrint : 579A 9B0E 43E8 0E40 EE93 BB1C 38CE 1C7B 3907 14F6 |