Date: Sat, 8 Dec 2018 18:24:35 -0800
From: Noah Misch <noah@leadboat.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: pgsql-hackers@postgresql.org
Subject: Re: Too-short timeouts in test code
Message-ID: <20181209022435.GE2973271@rfd.leadboat.com>
References: <20181209001601.GC2973271@rfd.leadboat.com>
 <20181209021412.GC1833@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20181209021412.GC1833@paquier.xyz>
User-Agent: Mutt/1.5.24 (2015-08-30)
Precedence: bulk

On Sun, Dec 09, 2018 at 11:14:12AM +0900, Michael Paquier wrote:
> On Sat, Dec 08, 2018 at 04:16:01PM -0800, Noah Misch wrote:
> > @@ -72,7 +72,7 @@ my $endpos = $node_master->safe_psql('postgres',
> >  print "waiting to replay $endpos\n";
> >  
> >  my $stdout_recv = $node_master->pg_recvlogical_upto(
> > -	'postgres', 'test_slot', $endpos, 10,
> > +	'postgres', 'test_slot', $endpos, 180,
> >  	'include-xids'     => '0',
> >  	'skip-empty-xacts' => '1');
> 
> Instead of allowing callers of pg_recvlogical_upto() to define the
> timeout they want, perhaps it would be better to hardcode the timeout
> limit within the routine?

pg_recvlogical_upto() has the ability to make timeout non-fatal, by calling it
in an array context.  No in-tree test uses that.  Out-of-tree code using that
feature would likely benefit from the ability to set timeout duration.  Hence,
I'm inclined not remove the timeout duration parameter.