Date: Fri, 28 Sep 2018 06:14:59 -0400
From: Bruce Momjian <bruce@momjian.us>
To: Daniel Verite <daniel@manitou-mail.org>
Cc: Tom Lane <tgl@sss.pgh.pa.us>,
	Peter Eisentraut <peter.eisentraut@2ndquadrant.com>,
	Michael Paquier <michael@paquier.xyz>,
	Andres Freund <andres@anarazel.de>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
Subject: Re: transction_timestamp() inside of procedures
Message-ID: <20180928101459.GA8989@momjian.us>
References: <23174.1537970823@sss.pgh.pa.us>
 <426715fa-0dd7-4826-b079-1426ca8e6f12@manitou-mail.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <426715fa-0dd7-4826-b079-1426ca8e6f12@manitou-mail.org>
User-Agent: Mutt/1.5.23 (2014-03-12)
Precedence: bulk

On Wed, Sep 26, 2018 at 09:23:58PM +0200, Daniel Verite wrote:
> 	Tom Lane wrote:
> 
> > I agree that it would be surprising for transaction timestamp to be newer
> > than statement timestamp. 
> 
> To me it's more surprising to start a new transaction and having
> transaction_timestamp() still pointing at the start of a previous 
> transaction.
> This feels like a side-effect of being spawned by a procedure,
> and an exception to what transaction_timestamp() normally means
> or meant until now.
> 
> OTOH transaction_timestamp() being possibly newer than
> statement_timestamp() seems like a normal consequence of
> transactions being created intra-statement.

Yes, that is a good point.  My thought has always been that statements
are inside of transactions, but the opposite is now possible.

-- 
  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 +