Re: [HACKERS] Can ICU be used for a database's default sort order?

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com>
Cc: Peter Geoghegan <pg(at)bowt(dot)ie>, Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: [HACKERS] Can ICU be used for a database's default sort order?
Date: 2018-10-02 06:37:06
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

On Wed, Sep 12, 2018 at 01:06:12PM +1200, Thomas Munro wrote:
> The ideal scope would be to track all referenced collation versions on
> every index, and only update them at CREATE INDEX or REINDEX time
> (also, as discussed in some other thread, CHECK constraints and
> partition keys might be invalidated and should in theory also carry
> versions that can only be updated by running a hypothetical RECHECK or
> REPARTITION command). Then a shared pg_collation catalog would make
> perfect sense, and there would be no need for it to have a collversion
> column at all, or an ALTER COLLATION ... REFRESH VERSION command, and
> therefore there would be no way to screw it up by REFRESHing the
> VERSION without having really fixed the problem.

Please note that the latest patch set does not apply, so this has been
switched to commit fest 2018-11, waiting on author for a rebase.

In response to


Browse pgsql-hackers by date

  From Date Subject
Next Message Michael Paquier 2018-10-02 06:40:14 Re: [PATCH] Add support for ON UPDATE/DELETE actions on ALTER CONSTRAINT
Previous Message Masahiko Sawada 2018-10-02 06:25:17 Re: [HACKERS] Transactions involving multiple postgres foreign servers, take 2