MIME-Version: 1.0
References: <e4ac16908dad3eddd3ed73c4862591375a3f0539.camel@j-davis.com>
 <47e1b4f72fe732c5ae85c6cf2c085b4e99a10120.camel@j-davis.com>
In-Reply-To: <47e1b4f72fe732c5ae85c6cf2c085b4e99a10120.camel@j-davis.com>
From: Robert Haas <robertmhaas@gmail.com>
Date: Tue, 10 Mar 2026 11:12:16 -0400
Message-ID: 
 <CA+TgmobnOn9ipEUcthLmxvv0ZKBWyCcc048qqgqy1X+tszw_Cg@mail.gmail.com>
Subject: Re: Change initdb default to the builtin collation provider
To: Jeff Davis <pgsql@j-davis.com>
Cc: pgsql-hackers@postgresql.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CA%2BTgmobnOn9ipEUcthLmxvv0ZKBWyCcc048qqgqy1X%2Btszw_Cg%40mail.gmail.com>
Precedence: bulk

On Fri, Oct 31, 2025 at 5:30=E2=80=AFPM Jeff Davis <pgsql@j-davis.com> wrot=
e:
> The attached patches implement a more modest proposal which does not
> conflict with Peter's objection about the display order:
>
> 0001: If the encoding is unspecified, and cannot be determined from the
> locale (i.e. the locale is C), then use UTF-8 rather than SQL_ASCII.

I don't know if this is exactly the right proposal, but I think it's
probably appropriate to start gently pushing people towards UTF-8
rather than anything else. Unicode has largely won, AFAICT, and the
use cases for anything else are increasingly narrow. I don't think we
should try to be coercive, but there's a reasonable presumption that
people who haven't said what they want probably want UTF8.

> 0002: If the provider is unspecified, and the locale is C or C.UTF-8,
> then use the builtin provider.

I'm much less convinced about this idea. I think the number of people
who will be unhappy about the less-user-friendly sort order changes is
probably quite high. It's reasonable to want something more stable and
better version-controlled than libc, but switching to a simple
code-point sort seems like a high price to pay for that.

--=20
Robert Haas
EDB: http://www.enterprisedb.com