Bug in sl_SI locale file

Bug in sl_SI locale file

[Jure Sah]

Hello,

I would like to report a bug in the sl_SI locale, the version I have is 
labeled like this:
% Language: sl
% Territory: SI
% Revision: 4.4
% Date: 1997-11-05

The problem is in the LC_NUMERIC section. The Thousands Separator 
(thousands_sep)  is indicated as <U0020> which is a space character. 
This creates problems in scripts, because space is sometimes a delimiter 
and with this setting, numbers are sometimes merged inappropriately. 
This is a problem that does not exist if a different locale is used 
(which I would assume is the usual way people solve this issue).

While a thousands separator is not commonly used, the technically 
correct character for this according to Slovenian number styling 
standards, would be <U002E> aka the dot. Likely this was not used, to 
avoid problems in compatibility with locales where the dot is a decimal 
separator (e.g.: English).

LP,
Jure

-- 
<https://www.t-2.net/> 	*Jure Sah*
IT, sistemska integracija• IT, system integrations
www.t-2.net <https://www.t-2.net/> • Opozorilo | Disclaimer 
<https://disclaimer.t-2.net>

Re: Bug in sl_SI locale file

[Florian Weimer]

* Jure Sah:

> I would like to report a bug in the sl_SI locale, the version I have
> is labeled like this:
> % Language: sl
> % Territory: SI
> % Revision: 4.4
> % Date: 1997-11-05
>
> The problem is in the LC_NUMERIC section. The Thousands Separator
> (thousands_sep)  is indicated as <U0020> which is a space
> character. This creates problems in scripts, because space is
> sometimes a delimiter and with this setting, numbers are sometimes
> merged inappropriately. This is a problem that does not exist if a
> different locale is used (which I would assume is the usual way people
> solve this issue).
>
> While a thousands separator is not commonly used, the technically
> correct character for this according to Slovenian number styling
> standards, would be <U002E> aka the dot. Likely this was not used, to
> avoid problems in compatibility with locales where the dot is a
> decimal separator (e.g.: English).

I can confirm that

<https://www.gov.si/assets/ministrstva/MDDSZ/VZD-/Resolucija-o-nacionalnem-programu-VZD.pdf>

linked from

<https://www.gov.si/podrocja/zaposlovanje-delo-in-upokojitev/varnost-in-zdravje-pri-delu/>

uses a dot as the separator:

| Zaradi duševnih motenj je bilo pri izbranih zdravnikih v Sloveniji med
| letoma 2009 in 2013 v povprečju 18.471 primerov bolniškega dopusta na
| leto, kar je 2,5 odstotka vseh bolniških dopustov.

Translated as (also linked available from the page):

| Doctors in Slovenia registered an average of 18,471 cases of sick
| leave per year due to mental disorders between 2009 and 2013, which is
| 2.5 per cent of all cases of sick leave.

I have filed a bug: <https://sourceware.org/bugzilla/show_bug.cgi?id=25233>

Our version has since switched to a Unicode narrow non-breaking space,
and I think we also disable grouping.  CLDR does not seem to have
accurate data (it uses a decimal point AFAICS).

Thanks,
Florian