Re: ICU for global collation - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: ICU for global collation
Date
Msg-id YyPfOTUT50q88mu8@paquier.xyz
Whole thread Raw
In response to Re: ICU for global collation  (Marina Polyakova <m.polyakova@postgrespro.ru>)
List pgsql-hackers
On Wed, Sep 14, 2022 at 05:19:34PM +0300, Marina Polyakova wrote:
> I was surprised that it is allowed to create clusters/databases where the
> default ICU collations do not actually work due to unsupported encodings:
>
> $ initdb --encoding SQL_ASCII --locale-provider icu --icu-locale en-US -D
> data &&
> pg_ctl -D data -l logfile start &&
> psql -c "SELECT 'a' < 'b'" template1
> ...
> waiting for server to start.... done
> server started
> ERROR:  encoding "SQL_ASCII" not supported by ICU
>
> $ createdb --encoding SQL_ASCII --locale-provider icu --icu-locale en-US
> --template template0 mydb &&
> psql -c "SELECT 'a' < 'b'" mydb
> ERROR:  encoding "SQL_ASCII" not supported by ICU
>
> The patch diff_check_icu_encoding.patch prohibits the creation of such
> objects...

Agreed that it is a bit confusing to get this type of error after the
database has been created when querying it due to a mix of unsupported
options.  Peter?
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: pg_basebackup's --gzip switch misbehaves
Next
From: bt22kawamotok
Date:
Subject: Re: [PATCH]Feature improvement for MERGE tab completion