Date: Sat, 17 Aug 2019 18:03:17 -0700
From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Alvaro Herrera <alvherre@2ndquadrant.com>,
	Dilip Kumar <dilipbalaut@gmail.com>, jungleboogie0@gmail.com,
	pgsql-bugs@lists.postgresql.org,
	Michael Paquier <michael@paquier.xyz>
Subject: Re: BUG #15964: vacuumdb.c:187:10: error: use of undeclared
 identifier 'FD_SETSIZE'
Message-ID: <20190818010317.kh5pxpfgpsskc6ip@alap3.anarazel.de>
References: <15964-c1753bdfed722e04@postgresql.org>
 <24757.1566073861@sss.pgh.pa.us>
 <20190817210025.v4bblh4lhsxajmlj@alap3.anarazel.de>
 <6402.1566079145@sss.pgh.pa.us>
 <20190817224142.gq57z3cwt7c76dtk@alap3.anarazel.de>
 <20190818001858.ho3ev4z57fqhs7a5@alap3.anarazel.de>
 <28480.1566089996@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <28480.1566089996@sss.pgh.pa.us>
Precedence: bulk

Hi,

On 2019-08-17 20:59:56 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > Hm. This made me think: Why is
> >                 if (concurrentCons > FD_SETSIZE - 1)
> > a useful test / error message?
> 
> Good point, it's not.  Subtracting off 10 or so might be reasonable.

I wonder if we shouldn't just do the same as pgbench now does, and just
only error when adding a too large fd. That does reduce the number of
detected cases, true, but it also adds robustness, because larger fds
are properly handled.

> > What is the reason that this doesn't use poll() in the first place?
> 
> We still support platforms without that, no?  Windows, for one.

Ah, right. I forgot that because we do rely on poll() in latch.c - but
we do have an alternative windows implementation there...

Greetings,

Andres Freund