use of Yy+0/Nn-1/etc... in LC_MESSAGES yesexpr/noexpr

[Mike Frysinger <vapier at gentoo dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=15262
do we have policy/guidance on the use of english chars in the yes/no
regexes ?  of the 202 locales that define yesexpr/noexpr, 195 of them
include [Yy]/[Nn], most of which aren't english.

my take: at the risk of being called anglocentric, we should add
[Yy] & [Nn] to all locales

related, what about locales that are in territories that are frequently
bilingual ?  en_CA for example allows Yes/Oui/No/Non.  CLDR only lists
one option per language.  it doesn't (currently) define things on a
per-locale basis.  this is a semi-moot point depending on the Yy/Nn
question above.

my take: only list the main language (so en_CA would drop Oui).
if we can get CLDR to list more, it would be easy to support.

related, what about langs that have multiple scripts ?  this comes up
with all the locales that have @latin or @devanagari or @cyrillic.
for yesexpr, sr_RS uses [ÐÐDd] and sr_RS@latin uses [Dd].

my take: i can go either way: we could have every lang support all the
alternative scripts (so sr_RS@latin would add ÐÐ), or we could try and
figure out which script is the "main" one and have it import all its
alternatives (so the sr_RS examples would stay the same).

https://sourceware.org/bugzilla/show_bug.cgi?id=15263
what about [+1]/[-0] ?  this is what the i18n definition uses, and what
about 7 others do as well.  should we include those everywhere too ?

my take: we should add [+1]/[-0] to all locales
-mike

Attachment: signature.asc
Description: Digital signature