Horton

Re: Mark all GUC variable as PGDLLIMPORT

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Chapman Flack <chap@anastigmatix.net>
Cc: Peter Eisentraut <peter.eisentraut@enterprisedb.com>, Julien Rouhaud <rjuju123@gmail.com>, Bruce Momjian <bruce@momjian.us>, Tom Lane <tgl@sss.pgh.pa.us>, Michael Paquier <michael@paquier.xyz>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2021-08-24T20:31:32Z
Lists: pgsql-hackers 
On Tue, Aug 24, 2021 at 3:36 PM Chapman Flack <chap@anastigmatix.net> wrote:
> Peter may have advocated for that kind of across-the-board adoption;
> my leaning is more to add an API that /can/ be adopted, initially with
> separately-linked extensions as the audience. Nothing would stop it being
> used in core as well, but no reason to change any site where it did not
> offer an advantage.
>
> I generally tend to be an incrementalist.

Sure, me too, but the point for me is that there doesn't seem to be a
shred of a reason to go this way at all. We've turned a discussion
about adding PGDLLIMPORT, which ought to be totally uncontroversial,
into some kind of a discussion about adding an API layer that no one
wants to prevent a hypothetical failure mode not in evidence.

-- 
Robert Haas
EDB: http://www.enterprisedb.com

Commits

  1. Remove PGDLLIMPORT marker from __pg_log_level

  2. Mark a few 'bbsink' related functions / variables static.

  3. Add some missing PGDLLIMPORT markings

  4. Apply PGDLLIMPORT markings broadly.

  5. Helper script to apply PGDLLIMPORT markings.

  6. Simplify declaring variables exported from libpgcommon and libpgport.