Re: Use correct collation in pg_trgm

Attached is v3 of the patch with the following changes:

On 26.01.2026 14:33, Alexander Korotkov wrote:
>> Updated patch looks good, I only noticed two minor things:
>>
>> * the new test should have a newline at the end of the file

Done.

>> * and probably a conditional skip based on locale availability, the
>> citext_utf8.sql test case does something similar

Done, the same way collate.icu.utf8.sql does it. This test also uses the
Turkish collation tr-x-icu.

> I wonder about existing indexes, which already use default collation
> not column collation. Should we add a release note saying they might
> be obsolete? Alternatively, given we now have opclass options, we may
> add a new opclass option defining whether to use column collation
> (must be false for existing indexes).

I think adding a release note is fine. We discussed this in [1] already
and there's precedence for doing it this way. See [2].

Looking at [3], it seems like we don't include release notes in
bug fix commits but rather collect them retroactively before cutting the
release. The file doc/src/sgml/release-19.sgml is also still completely
empty on master.

--
David Geier

[1]
https://www.postgresql.org/message-id/CAEze2WiUL9idZBbuUN%2BMuWqr6DcPr_-C91E9MTx%3DH62Xx5fHaQ%40mail.gmail.com

[2] https://www.postgresql.org/docs/18/release-18.html#RELEASE-18-MIGRATION

[3]
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=fb1a18810f0