Re: ON CONFLICT (and manual row locks) cause xmax of updated tuple tounnecessarily be set - Mailing list pgsql-hackers

From Peter Geoghegan
Subject Re: ON CONFLICT (and manual row locks) cause xmax of updated tuple tounnecessarily be set
Date
Msg-id CAH2-Wz=39Q-Lq0xBm+hLn=sGq7EHHwTcxunqokHidrYs0bF2vQ@mail.gmail.com
Whole thread Raw
In response to Re: ON CONFLICT (and manual row locks) cause xmax of updated tupleto unnecessarily be set  (Andres Freund <andres@anarazel.de>)
List pgsql-hackers
On Thu, Jul 25, 2019 at 3:10 PM Andres Freund <andres@anarazel.de> wrote:
> > I agree that this is unfortunate. Are you planning on working on it?
>
> Not at the moment, no. Are you planning / hoping to take a stab at it?

The current behavior ought to be fixed, and it seems like it falls to
me to do that. OTOH, anything that's MultiXact adjacent makes my eyes
water. I don't consider myself to be particularly well qualified.

I'm sure that I could quickly find a way of making the behavior you
have pointed out match what is expected without causing any regression
tests to fail, but that's the easy part.

--
Peter Geoghegan



pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: [Proposal] Table-level Transparent Data Encryption (TDE) and KeyManagement Service (KMS)
Next
From: Stephen Frost
Date:
Subject: Re: [Proposal] Table-level Transparent Data Encryption (TDE) and KeyManagement Service (KMS)