Re: Order changes in PG16 since ICU introduction

Jeff Davis <pgsql@j-davis.com>

From: Jeff Davis <pgsql@j-davis.com>

To: Andrew Gierth <andrew@tao11.riddles.org.uk>

Cc: Peter Eisentraut <peter.eisentraut@enterprisedb.com>, Sandro Santilli <strk@kbt.io>, Tom Lane <tgl@sss.pgh.pa.us>, Regina Obe <lr@pcorp.us>, pgsql-hackers@lists.postgresql.org

Date: 2023-04-26T01:15:21Z

Lists: pgsql-hackers

Attachments

v2-0001-ICU-do-not-convert-locale-C-to-en-US-u-va-posix.patch (text/x-patch) patch v2-0001
v2-0002-ICU-support-locale-C-with-the-same-behavior-as-li.patch (text/x-patch) patch v2-0002
v2-0003-Make-LOCALE-apply-to-ICU_LOCALE-for-CREATE-DATABA.patch (text/x-patch) patch v2-0003

On Fri, 2023-04-21 at 22:35 +0100, Andrew Gierth wrote:
> > > > > 
> Can lc_collate_is_c() be taught to check whether an ICU locale is
> using
> POSIX collation?

Attached are a few small patches:

  0001: don't convert C to en-US-u-va-posix
  0002: handle locale C the same regardless of the provider, as you
suggest above
  0003: make LOCALE (or --locale) apply to everything including ICU

As far as I can tell, any libc locale has a reasonable match in ICU, so
setting LOCALE to either C or a libc locale name should be fine. Some
locales are only valid in ICU, e.g. '@colStrength=primary', or a
language tag representation, so if you do something like:

  create database foo locale 'en_US@colStrenghth=primary'
    template template0;

You'll get a decent error like:

  ERROR:  invalid LC_COLLATE locale name: "en_US@colStrenghth=primary"
  HINT:  If the locale name is specific to ICU, use ICU_LOCALE.

Overall, I think it works out nicely. Let me know if there are still
some confusing cases. I tried a few variations and this one seemed the
best, but I may have missed something.

Regards,
	Jeff Davis

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

Introduce "builtin" collation provider.
- 2d819a08a1cb 17.0 landed
ICU: do not convert locale 'C' to 'en-US-u-va-posix'.
- f3a01af29b19 16.0 landed
- f7faa9976cc0 16.0 landed
initdb: change default --locale-provider back to libc.
- 2535c74b1a61 16.0 landed
CREATE DATABASE: make LOCALE apply to all collation providers.
- a14e75eb0b6a 16.0 landed
Remove read-only server settings lc_collate and lc_ctype
- b0f6c437160d 16.0 landed
ICU: use uloc_getDefault() for initdb.
- ec1264f01e37 16.0 landed
Doc fixes for commit 1e16af8ab5.
- f4a9422c0c37 16.0 landed
Doc: fix some rendering problems in the PDF docs build.
- b62381d9a23b 16.0 landed
Doc improvements for language tags and custom ICU collations.
- 1e16af8ab5d7 16.0 landed
ICU: check for U_STRING_NOT_TERMINATED_WARNING.
- 1c634f6647c2 16.0 landed
Reduce icu_validation_level default to WARNING.
- 6de31ce446e5 16.0 landed
Revert "ICU: do not convert locale 'C' to 'en-US-u-va-posix'."
- 455f948b0d03 16.0 landed
Fix initdb --no-locale.
- 5cd1a5af4d17 16.0 landed
Avoid character classification in regex escape parsing.
- c04c6c5d6f5f 16.0 cited