svn commit: r302016 - in head: share/timedef tools/tools/locale/tools

Andrey Chernov ache at freebsd.org
Mon Jun 20 02:57:17 UTC 2016


On 19.06.2016 22:40, Baptiste Daroussin wrote:
> On Sun, Jun 19, 2016 at 10:02:26PM +0300, Andrey Chernov wrote:
>> On 19.06.2016 20:21, Andrey Chernov wrote:
>>> On 19.06.2016 20:10, Baptiste Daroussin wrote:
>>>> On Sun, Jun 19, 2016 at 08:05:30PM +0300, Andrey Chernov wrote:
>>>>> On 19.06.2016 19:52, Andrey Chernov wrote:
>>>>>> On 19.06.2016 19:47, Andrey Chernov wrote:
>>>>>>> On 19.06.2016 18:49, Baptiste Daroussin wrote:
>>>>>>>> On Sun, Jun 19, 2016 at 04:52:34PM +0200, Baptiste Daroussin wrote:
>>>>>>>>> On Sun, Jun 19, 2016 at 07:57:49AM +0300, Andrey Chernov wrote:
>>>>>>>>>> On 19.06.2016 6:44, Hajimu UMEMOTO wrote:
>>>>>>>>>>> Log:
>>>>>>>>>>>   - Prefer to use %d over %e where the day of the month should be zero
>>>>>>>>>>>     filled.
>>>>>>>>>>>   - Since %e means the day of the month as well, regard %e as same as %d
>>>>>>>>>>>     in md_order.
>>>>>>>>>>
>>>>>>>>>> Nonsense doubled formats in sr_*_RS locales and nonsense md_order there too.
>>>>>>>>>
>>>>>>>>> Crap I'll dig into it, thanks for spotting it.
>>>>>>>>>
>>>>>>>>> Best regards,
>>>>>>>>> Bapt
>>>>>>>>
>>>>>>>> Should be fixed by:
>>>>>>>> https://people.freebsd.org/~bapt/locale-triple-components.diff
>>>>>>>>
>>>>>>>> The perl script generating the locales was badly handling locales with 3
>>>>>>>> components, fixed now.
>>>>>>>>
>>>>>>>> Sorry about that
>>>>>>>>
>>>>>>>> Best regards,
>>>>>>>> Bapt
>>>>>>>>
>>>>>>>
>>>>>>> sr locales are badly named before anything else, proper format is
>>>>>>> language[_COUNTRY][.encoding][@variant]
>>>>>>> I.e.
>>>>>>> sr_RS.UTF-8 at Latn
>>>>>>> sr_RS.ISO8859-2 at Latn
>>>>>>> sr_RS.ISO8859-t at Cyrl
>>>>>>>
>>>>>>> I don't know, if out utilities (f.e. locale(1)) can handle @variant suffix.
>>>>>>>
>>>>>>>
>>>>>> =sr_RS.ISO8859-5 at Cyrl
>>>>>> =our, and not utilities only but libc too
>>>>>
>>>>> Since right now we have only single @variant per each encoding, and no
>>>>> @variant for other locales, we can just drop it and use right names like:
>>>>> sr_RS.UTF-8
>>>>
>>>> We have 2 for the above the Cyrl version and the Latn version
>>>>
>>>>> sr_RS.ISO8859-2
>>>>> sr_RS.ISO8859-5
>>>>
>>>> We can make aliases easily for that is that is what people want, not the we have
>>>> the same for other locales:
>>>>
>>>> zh_Hans_CN.GB18030
>>>> zh_Hans_CN.GB2312
>>>> zh_Hans_CN.GBK
>>>> zh_Hans_CN.UTF-8
>>>> zh_Hans_CN.eucCN
>>>> zh_Hant_HK.UTF-8
>>>> zh_Hant_TW.Big5
>>>> zh_Hant_TW.UTF-8
>>>>
>>>> With some aliases for some of the previously existing ones:
>>>> zh_TW.Big5
>>>> zh_TW.UTF-8
>>>> zh_CN.GB18030
>>>> zh_CN.GB2312
>>>> zh_CN.GBK
>>>> zh_CN.UTF-8
>>>> zh_CN.eucCN
>>>> zh_HK.UTF-8
>>>>
>>>> Some of the new locales (that didn't exist before) I have only imported the new
>>>> name format:
>>>> kk_Cyrl_KZ.UTF-8
>>>> mn_Cyrl_MN.UTF-8
>>>> sr_Cyrl_RS.ISO8859-5
>>>> sr_Cyrl_RS.UTF-8
>>>> sr_Latn_RS.ISO8859-2
>>>> sr_Latn_RS.UTF-8
>>>>
>>>> I would prefer staying on the new syntax given it is allowed by some RFC and it
>>>> is slowly being adopted everywhere else.
>>>>
>>>> Btw another RFC: https://www.ietf.org/rfc/rfc5646.txt
>>>>
>>>> Best regards,
>>>> Bapt
>>>>
>>>
>>> Old POSIX don't use RFC 5646, there are ISO-639 two-letter codes. When
>>> encodings are different, there is no needs to specify variants, they
>>> have meaning only for the same encoding.
>>>
>>>
>>
>> I found POSIX reference about @-syntax:
>> [language[_territory][.codeset][@modifier]]
>> http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap08.html
>> It says nothing about language, territory or codeset standards.
>>
>> They comes from another standard, ISO 15897:
>> "8. Natural language, as specified in ISO 639
>> 9. Territory, as two-letter form of ISO 3166"
>> I can't find there something about codeset/charmap standartization, but
>> perhaps I just overlook some reference.
>>
> 
> Thanks a lot for that reference I wonder how I failed at finding it, I read
> twice the locale definition in opengroup documents.
> 
> I have found yet proof in the code, but from my testings (using nl_langinfo())
> 
> I can see that the @modifier is properly supported.
> 
> I have 2 proposal then:
> 1/ Keep the current version and add aliases with "@" for the one in variants
> 2/ Rename all the variants we have now into the POSIX version with "@"

To minimize confusion among people and better cross-platform
compatibility IMHO we need to see what Linux have here (in parts it not
violates POSIX and ISO standards, of course). We have different charmap
names (ISO-* vs. ISO*), and we have it already too long to change as
many people thinks, but new sr and other locales can be easily set in
line while they are still new.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 455 bytes
Desc: OpenPGP digital signature
URL: <http://lists.freebsd.org/pipermail/svn-src-all/attachments/20160620/fb21a412/attachment.sig>


More information about the svn-src-all mailing list