From: Tom Lane <tgl@sss.pgh.pa.us>
To: Fabien COELHO <coelho@cri.ensmp.fr>
cc: Thomas Munro <thomas.munro@gmail.com>, didier <did447@gmail.com>,
        Alvaro Herrera <alvherre@2ndquadrant.com>,
        Pg Hackers <pgsql-hackers@lists.postgresql.org>, buildfarm@coelho.net
Subject: Re: LLVM compile failing in seawasp
In-reply-to: <alpine.DEB.2.21.1907280935520.1292@lancre>
References: <20190606173216.GA6306@alvherre.pgsql>
 <CAJRYxuLbtcN+D+KrD6wui-rhXoo7XOungCB5vAech=R90LOTuA@mail.gmail.com>
 <5762.1559866377@sss.pgh.pa.us>
 <CA+hUKGJUGreiJART78O=pd4k9x0fZebL5BfGe8_mHBJe6Qd_WQ@mail.gmail.com>
 <alpine.DEB.2.21.1907270700360.21735@lancre>
 <CA+hUKGKxQLVQ+W6fcvN5jPyPNX+6CAXDy0_n+xvWPnMn73O+wg@mail.gmail.com>
 <CA+hUKG+vW2BHoJXJYhzBpzwC-pH854MguThsLewLcGh4pst0Jg@mail.gmail.com>
 <alpine.DEB.2.21.1907280935520.1292@lancre>
Comments: In-reply-to Fabien COELHO <coelho@cri.ensmp.fr>
	message dated "Sun, 28 Jul 2019 09:47:17 +0200"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <24643.1564322045.1@sss.pgh.pa.us>
Date: Sun, 28 Jul 2019 09:54:05 -0400
Message-ID: <24644.1564322045@sss.pgh.pa.us>
Precedence: bulk

Fabien COELHO <coelho@cri.ensmp.fr> writes:
> Otherwise, why not simply move llvm C++ includes *before* postgres 
> includes?

We've been burnt in the past by putting other headers before postgres.h.
(A typical issue is that the interpretation of <stdio.h> varies depending
on _LARGE_FILES or a similar macro, so you get problems if something
causes that to be included before pg_config.h has set that macro.)
Maybe none of the platforms where that's an issue have C++, but that
doesn't seem like a great assumption.

> They should be fully independent anyway, so the order should 
> not matter?

On what grounds do you claim that's true anywhere, let alone
everywhere?

			regards, tom lane