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

From Michael Paquier
Subject Re: ICU for global collation
Date
Msg-id YrgIwulT/ZIK5hqy@paquier.xyz
Whole thread Raw
In response to Re: ICU for global collation  (Julien Rouhaud <rjuju123@gmail.com>)
List pgsql-hackers
On Sun, Jun 26, 2022 at 11:51:24AM +0800, Julien Rouhaud wrote:
> On Sat, Jun 25, 2022 at 10:19:30AM -0500, Justin Pryzby wrote:
>>> +    if (GET_MAJOR_VERSION(old_cluster.major_version) <= 1500)
>>> +        snprintf(query + strlen(query), sizeof(query) - strlen(query),
>>> +                 "'c' AS datcollprovider, NULL AS daticucoll, ");
>>> +    else
>>> +        snprintf(query + strlen(query), sizeof(query) - strlen(query),
>>> +                 "datcollprovider, daticucoll, ");
>>> +    snprintf(query + strlen(query), sizeof(query) - strlen(query),
>>>               "pg_catalog.pg_tablespace_location(t.oid) AS spclocation "
>>>               "FROM pg_catalog.pg_database d "
>>>               " LEFT OUTER JOIN pg_catalog.pg_tablespace t "
>
> Indeed!

Oops.  Beta2 tagging is very close by, so I think that it would be
better to not take a risk on that now, and this is an issue only when
upgrading from v15 where datcollprovider is ICU for a database.
As things stand, someone using beta1 with this new feature, running
pg_upgrade to beta2 would lose any non-libc locale provider set.
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Julien Rouhaud
Date:
Subject: Re: ICU for global collation
Next
From: Tom Lane
Date:
Subject: Re: Core dump in range_table_mutator()