Re: [PATCH v2] localedata: Updates for Taiwanese locales [BZ #24409]

[Wei-Lun Chao <bluebat@member.fsf.org>]

Dear Shoichi,

Thanks for your explanation.
I'll add this part of patch.

TAMUKI Shoichi <tamuki@linet.gr.jp> 於 2021年5月16日 週日 下午6:04寫道：
>
> Hello Wei-Lun-san,
>
> From: Wei-Lun Chao <bluebat@member.fsf.org>
> Subject: Re: [PATCH v2] localedata: Updates for Taiwanese locales [BZ #24409]
> Date: Thu, 13 May 2021 13:10:12 +0800
>
> > > > Changelog:
> > > > * localedata/locales/cmn_TW: Header cleanup; Remove space (abmon); Add
> > > > (week); Change (thousands_sep); Simplify (yesexpr) and (noexpr).
> > > > * localedata/locales/hak_TW: Likewise and add collation.
> > > > * localedata/locales/nan_TW: Likewise and add collation.
> > > > * localedata/locales/lzh_TW: Likewise and add collation.
> > >
> > > If elements from 0 to 59 are defined for alt_digits in lzh_TW, there
> > > are the following problems:
> > >
> > > $ LANG=lzh_TW date -d "1959-04-01 09:00:00" +"%Oy<U5E74>"
> > > <U4E94><U5341><U4E5D><U5E74>
> > > $ LANG=lzh_TW date -d "1960-04-01 09:00:00" +"%Oy<U5E74>"
> > > 60<U5E74>
> > >
> > > Up to 99 elements should be defined for alt_digits.
> >
> > First, alt_digits works like "alt_numbers". I don't know how to just
> > map each digit to another character.
> > Before this patch, alt_digits covers 0..31, and this patch covers
> > 0..59. I am not sure, is it worthy to extend
> > to 0..99, because those presentations are not suitable for year
> > numbers. For example:
> > $ LANG=lzh_TW date -d "1959-04-01 09:00:00" +"%Oy<U5E74>"
> > We will expect <U4E94><U4E5D><U5E74> instead of <U4E94><U5341><U4E5D><U5E74>
>
> I understand what you are concerned about.
>
> In Japan, for example, it is not sutable to use number names to
> express another form of year notation using alternative numeric
> symbols in years such as AD 1984 and AD 645.  It is commonly expressed
> in positional notation.  I think the situation is the same in Taiwan.
>
> On the other hand, for example, it is not suitable to use positional
> notation to express another form of year notation using alternative
> numeric symbols in years such as AD 78.  It is commonly expressed in
> number names.  Please see the attached file alt_digits.png.
>
> I think that these boundaries are empirically determined by whether
> the numerical value is 2 digits or less or 3 digits or more.
> Therefore, it makes sense to define alt_digits up to 99.
>
> Regards,
> TAMUKI Shoichi