Date: Wed, 21 Aug 2019 09:49:54 -0400
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: Andres Freund <andres@anarazel.de>, Tom Lane <tgl@sss.pgh.pa.us>,
	Dilip Kumar <dilipbalaut@gmail.com>, jungleboogie0@gmail.com,
	pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #15964: vacuumdb.c:187:10: error: use of undeclared
 identifier 'FD_SETSIZE'
Message-ID: <20190821134954.GA26174@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20190821073619.GC26424@paquier.xyz>
User-Agent: Mutt/1.9.4 (2018-02-28)
Precedence: bulk

On 2019-Aug-21, Michael Paquier wrote:

> On Tue, Aug 20, 2019 at 10:40:05PM -0400, Alvaro Herrera wrote:
> > 
> > BTW why fprintf?  Since you can get into this by careless use of -j,
> > ISTM that this should be a translatable string.  I'd consider
> > pg_log_fatal().  Maybe something like "Argument of -j too large for this
> > platform -- try %d", numslots - i.
> 
> If giving a recommendation, shouldn't that be linked only to "i"
> instead?  If for example numslots = 1024, but that we get an error
> after allocating 128 slots because of a lack of fds in the range, then
> we would recommend 896, which would still result in an error.

I was too tired last night to think this through properly ... the first
few times I wrote that were even more wrong.  "i" seems right.

> I quite like the suggestion from Andres to keep the message simple
> with "too many jobs for this platform's select()".  You are indeed
> right about the pg_log_fatal() part here, I have changed the patch to
> do that on my local branch.

Well, it's a user-facing error, so I'd rather make it user-friendly.
It doesn't seem difficult, or unreliable enough not to try.

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services