Are datcollate/datctype always libc even under --with-icu ?

From: Chapman Flack <chap(at)anastigmatix(dot)net>
To: PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Are datcollate/datctype always libc even under --with-icu ?
Date: 2021-12-22 21:46:40
Message-ID: 61C39CC0.2010208@anastigmatix.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

... ok, I see that the answer is yes, according to the commit comment
for eccfef8:

Currently, ICU-provided collations can only be explicitly named
collations. The global database locales are still always libc-provided.

I got there the long way, by first wondering how to tell whether a
datcollate or datctype string was intended for libc or ICU, and then
reading pg_perm_setlocale, and then combing through the docs at
CREATE DATABASE and createdb and initdb and Collation Support and
pg_database and the release notes for 10, sure that I would find
the answer staring at me in one of those places once I knew I was asking.

Next question: the "currently" in that comment suggests that could change,
but is there any present intention to change it, or is this likely to just
be the way it is for the foreseeable future?

Regards,
-Chap

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Geoghegan 2021-12-22 22:19:16 Re: Unifying VACUUM VERBOSE and log_autovacuum_min_duration output
Previous Message Peter Geoghegan 2021-12-22 21:33:28 Re: do only critical work during single-user vacuum?