Feedback-ID: id4a34324:Fastmail
Date: Fri, 3 Jun 2022 09:23:21 -0700
From: Andres Freund <andres@anarazel.de>
To: Aleksander Alekseev <aleksander@timescale.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>,
	PostgreSQL Development <pgsql-hackers@postgresql.org>
Subject: Re: [RFC] building postgres with meson
Message-ID: <20220603162321.x2ldcw7vb32oqumj@alap3.anarazel.de>
References: <20220602163930.chnxatxr5pr472t4@alap3.anarazel.de>
 <1064888.1654189729@sss.pgh.pa.us>
 <20220602172609.3wuouzxhflohddc3@alap3.anarazel.de>
 <1069576.1654191231@sss.pgh.pa.us>
 <20220602174845.l6icnqqu63sghnku@alap3.anarazel.de>
 <1089211.1654196710@sss.pgh.pa.us>
 <20220602191755.tiop4clrbwncyg6q@alap3.anarazel.de>
 <1103549.1654199630@sss.pgh.pa.us>
 <20220602211333.op5mz2z2txqhuxsn@alap3.anarazel.de>
 <CAJ7c6TPAkZb12He_JvMqB0n_RrUeZGfQO0eCCZPbU+uFHjOmoA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <CAJ7c6TPAkZb12He_JvMqB0n_RrUeZGfQO0eCCZPbU+uFHjOmoA@mail.gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/20220603162321.x2ldcw7vb32oqumj%40alap3.anarazel.de>
Precedence: bulk

Hi,

On 2022-06-03 12:35:45 +0300, Aleksander Alekseev wrote:
> > Part of the patch series is to use visibility attributes, and your #if
> > TS_EMPTY(PGDLLEXPORT) thing can't handle that.
> 
> Out of curiosity, how come a patchset that adds an alternative build system
> changes the visibility attributes?

It was the simplest path - on windows (and AIx) extension symbols need to be
explicitly exported. We did that by building the objects constituting
extension libraries, collecting the symbols, generating an export file with
all the symbols, which then is passed to the linker.  It was a lot less work
to just add the necessary PGDLLEXPORT annotations than make that export file
generation work for extensions.


> I would guess they should be the same for both Autotools and Meson.

It is, the patch adds it to both.


> Is it necessary in order to make Meson work?

Yes, or at least the simplest path.


> If not, maybe it should be a separate patch.

It is.

https://commitfest.postgresql.org/38/3396/

Greetings,

Andres Freund