Horton

Re: Mark all GUC variable as PGDLLIMPORT

Julien Rouhaud <rjuju123@gmail.com>

View thread View on postgr.es
From: Julien Rouhaud <rjuju123@gmail.com>
To: Pavel Stehule <pavel.stehule@gmail.com>
Cc: Michael Paquier <michael@paquier.xyz>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2021-08-22T12:25:51Z
Lists: pgsql-hackers 
On Sun, Aug 22, 2021 at 02:17:16PM +0200, Pavel Stehule wrote:
> 
> Is true, so it increases differences between Windows and Unix, and fixing
> these issues is not fun work. On the other hand, maybe direct access to
> these variables from extensions is an antipattern, and we should use a
> direct function call API and functions current_setting and set_config. The
> overhead is usually not too big.

Yes, and that's what I did for one of my extensions.  But that's still a bit of
overhead, and extra burden only seen when trying to have Windows compatiblity,
and I hope I can get rid of that at some point.

If direct variable access shouldn't be possible, then we should explicitly tag
those with __attribute__ ((visibility ("hidden"))) or something like that to
have a more consistent behavior.

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.