Re: UTF-8 Sorting

Gerhard Schmidt skrev 2014-03-12 15:58:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 12.03.2014 15:50, Rolf Nielsen wrote:
>>
>>
>> Gerhard Schmidt skrev 2014-03-12 10:31:
>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
>>>
>>> Hi,
>>>
>>> I've a problem with FreeBSD, UTF-8 and Sorting.
>>>
>>> e.g. there is a file with the following content
>>>
>>> Meier M=FCller =D6ger Ofner Schmidt
>>>
>>> I have set my Terminal to ISO-8859-1 Encoding and call sort on
>>> this file I get the following output.
>>>
>>> Meier M=FCller Ofner =D6ger Schmidt
>>>
>>> Which is correctly sorted.
>>>
>>> When i change my Terminal to UTF-8 encoding and convert the file
>>> to UTF-8 and call sort again I get the following output.
>>>
>>> Meier M=FCller Ofner Schmidt =D6ger
>>>
>>> which is wrong.
>>>
>>> The problem seams to be that the LC_COLLATE file in the
>>> de_DE.UTF-8 locale is linked to ../la_LN.US-ASCII/LC_COLLATE (as
>>> are all LC_COLLATE Files in any UTF-8 locale).
>>>
>>> After some Research i found a Mail from Kuba Lida in December
>>> 2008 (yeah that's 5 Years ago) stating the same Problem and got
>>> no response.
>>>
>>> Why isn't there a UTF-8 LC_COLLATE file for any language. Kuba
>>> Lida believed there was a Problem with multibyte collate files in
>>> FreeBSD. Is this true and are there plans to fix this problem.
>>>
>>> The same test under Linux works without problem.
>>>
>>> Regards Estartu
>>>
>>> - -- -
>>
>> Hi,
>>
>> Hmm, to me the result that you claim is wrong looks perfectly
>> correct, however, it may of course differ between languages. In
>> Swedish =D6 is a separate letter, located last in the alphabet (from
>> A to Z we have the exact same alphabet as English, and then come =C5,
>> =C4 and =D6, in that order).
>
> Yeah, Sweedisch sorts these characters after Z but in German =D6 equals
> Oe in Names and O in all other cases. There have to be collation
> tables for different languages as there are different one for dieffent
> languages in ISO encoding. I know that the direfrence in Name and Not
> name will not be implementable but the default whould be much of an
> improvement.
>
> The same difference is between German German (de_DE) and Austrian
> German (de_AT).
>
> Regards
>     Estartu

I see. Well, different countries, different customs. :)

(I should have included the list in my previous reply, but I hit the =

wrong button. I apologise for that).

Regards,
Rolf
_______________________________________________
[e-mail] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "[e-mail]"