This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Locales: Cyrillic -> ASCII transliteration table [BZ #2872] re-submission for 2.29


Hi,

On 2018-10-10 01:42, Egor Kobylkin wrote:
> Ups, sorry, wrong link to the patch
> correct link https://sourceware.org/bugzilla/attachment.cgi?id=11303

Although I haven't checked every rule this in general looks very good
(but see below). Not sure do we want to add the few missing characters
mentioned at https://en.wikipedia.org/wiki/Cyrillic_script_in_Unicode,
e.g., one instantly notices that U+0400 is missing. (I wouldn't add at
least initially the more exotic characters, like the historic ones,
though.) Perhaps filing a bug or two for these cases for separate
consideration would be ok.

> On 10.10.2018 00:40, Egor Kobylkin wrote:
>> On 10.10.2018 00:17, Rafal Luzynski wrote:
>>> 9.10.2018 20:34 Egor Kobylkin <egor@kobylkin.com> wrote:
>>>>
>>>> The culprits were the "" around the "<U0423><U0301>" (<U00DA>) and
>>>> "<U0443><U0301>" (<U00FA>).
>>>> It works now with
>>>> % CYRILLIC UNDEFINED
>>>> <U0423><U0301> <U00DA>;"<U0055><U0060>"
>>>> % CYRILLIC UNDEFINED
>>>> <U0443><U0301> <U00FA>;"<U0075><U0060>"
>>>>
>>>> [...]
>>>
>>> I wonder why you need Cyrillic U with acute, and why you comment it
>>> as "undefined" at all.  I know that any Cyrillic vowel may appear with
>>> an acute accent but "the diacritic is used only in dictionaries, children's
>>> books, resources for foreign-language learners (...)". [1]  So maybe
>>> all vowels with an acute accent should be handled (which I think is fine)
>>> rather than just U.
>>
>> I have just taken the https://en.wikipedia.org/wiki/ISO_9 table and
>> implemented it on Marko's suggestion. Personally I have no opinion on
>> what letters should be included and under what name. These funny Us just
>> happened to be in the ISO9 table.
>>
>> There is no codepoint and no name for <U0423><U0301> and <U0443><U0301>
>> in Unicode. That’s why its coming through that way from my worksheet as
>> it does a reverse lookup on the names based on the Unicode codepoints.
>>
>> Manually we can change it to whatever you’d suggest in the
>> translit_cyrillic. I just don’t know the right name.

I'm not sure this will work, no existing rule in translit_* files
contain two characters, I'd assume that the rule for U+0423 is applied
first and then the below rule is never used.

% CYRILLIC UNDEFINED
<U0423><U0301> <U00DA>;"<U0055><U0060>"

Perhaps this should be commented out or removed altogether if it's not
working as intended.

Thanks,

-- 
Marko Myllynen


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]