Re: Order changes in PG16 since ICU introduction

Tatsuo Ishii wrote:

> >> Yes it's a special case but when doing initdb --locale=C, a user does
> >> not need or want an ICU locale. They want the same thing than what v15
> >> does with the same arguments: a template0 database with
> >> datlocprovider='c', datcollate='C', datctype='C', dateiculocale=NULL.
>
> So in this case the only way to keep the same behavior in v16 with "initdb
> --locale=C" (--no-locale) in v15 is, bulding PostgreSQL --without-icu?

AFAIK the --no-locale case in v16 is fixed since:

commit 5cd1a5af4d17496a58678c8eb7ab792119c2d723
Author: Jeff Davis <jdavis(at)postgresql(dot)org>
Date: Fri Apr 21 13:11:18 2023 -0700

Fix initdb --no-locale.

Discussion: https://postgr.es/m/878relf7cb.fsf@news-spur.riddles.org.uk
Reported-by: Andrew Gierth

The --locale=C case is still being discussed. To me it should
produce the same result than --no-locale and --locale=C in v15, that is,
"ICU is the default" does not apply to that case, but currently
it initializes the cluster with an ICU locale.

Best regards,
--
Daniel Vérité
https://postgresql.verite.pro/
Twitter: @DanielVerite