Re: [PATCH] Hooks at XactCommand level

From: Gilles Darold <gilles(at)darold(dot)net>
To: Andres Freund <andres(at)anarazel(dot)de>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Nicolas CHAHWEKILIAN <leptitstagiaire(at)gmail(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: [PATCH] Hooks at XactCommand level
Date: 2021-09-04 11:00:58
Message-ID: de32a223-285f-2c9e-9f99-9d3111d49fd4@darold.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I have changed the status of this proposal as rejected.

To resume the final state of this proposal there is no consensus on the
interest to add a hook on start xact commands. Also the only useful case
for this hook was to be able to have a server side automatic rollback at
statement level. It can be regrettable because I don't think that
PostgreSQL will have such feature before a long time (that's probably
better) and a way to external implementation through an extension would
be helpful for migration from other RDBMS like DB2 or Oracle. The only
ways to have this feature is to handle the rollback at client side using
savepoint, which is at least 3 times slower than a server side
implementation, or not use such implementation at all. Outside not being
performant it doesn't scale due to txid wraparound. And the last way is
to use a proprietary forks of PostgreSQL, some are proposing this  feature.

--
Gilles Darold
http://www.darold.net/

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Andrew Dunstan 2021-09-04 13:58:08 Re: Postgres perl module namespace
Previous Message Amit Kapila 2021-09-04 09:51:14 Re: using an end-of-recovery record in all cases