Date: Fri, 21 Sep 2018 06:28:22 -0400
From: Bruce Momjian <bruce@momjian.us>
To: Michael Paquier <michael@paquier.xyz>
Cc: Andres Freund <andres@anarazel.de>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
Subject: Re: transction_timestamp() inside of procedures
Message-ID: <20180921102822.GA22699@momjian.us>
References: <20180920234040.GC29981@momjian.us>
 <20180921045536.GJ1338@paquier.xyz>
 <20180921051206.bzfiew4nthmunpbj@alap3.anarazel.de>
 <20180921053425.GM1338@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180921053425.GM1338@paquier.xyz>
User-Agent: Mutt/1.5.23 (2014-03-12)
Precedence: bulk

On Fri, Sep 21, 2018 at 02:34:25PM +0900, Michael Paquier wrote:
> On Thu, Sep 20, 2018 at 10:12:06PM -0700, Andres Freund wrote:
> > Isn't the point that transaction_timestamp() does *not* currently change
> > its value, even though the transaction (although not the outermost
> > statement) has finished?
> 
> Ouch, yes.  I see the point now.  Indeed that's strange to not have a
> new transaction timestamp after commit within the DO block..

So, this puts us in an odd position.  Right now everyone knows that
statement_timestamp() is only changed by the outer statement, i.e., a
SELECT in a function doesn't change statement_timestamp().   So, there
is an argument that transaction_timestamp() should do the same and not
change in a function --- in fact, if it does change, it would mean that
transaction_timestamp() changes in a function, but statement_timestamp()
doesn't --- that seems very odd.  It would mean that new statements in a
function don't change statement_timestamp(), but new transctions in a
function do --- again, very odd.

-- 
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

+ As you are, so once was I.  As I am, so you will be. +
+                      Ancient Roman grave inscription +