Horton

Re: Mark all GUC variable as PGDLLIMPORT

Chapman Flack <chap@anastigmatix.net>

View thread View on postgr.es
From: Chapman Flack <chap@anastigmatix.net>
To: Bruce Momjian <bruce@momjian.us>, Tom Lane <tgl@sss.pgh.pa.us>
Cc: Julien Rouhaud <rjuju123@gmail.com>, Michael Paquier <michael@paquier.xyz>, pgsql-hackers@lists.postgresql.org
Date: 2021-08-23T14:57:16Z
Lists: pgsql-hackers 
On 08/23/21 10:36, Bruce Momjian wrote:

> So the problem is that extensions only _need_ to use that API on
> Windows, so many initially don't, or that the API is too limited?

I think there can be cases where it's too limited, such as when significant
computation or validation is needed between the form of the setting known
to the GUC machinery and the form that would otherwise be available in
the global.

I'm thinking, for instance, of the old example before session_timezone
was PGDLLIMPORTed, and you'd have to GETCONFIGOPTION("timezone") and
repeat the work done by pg_tzset to validate and map the timezone name
through the timezone database, to reconstruct the value that was
otherwise already available in session_timezone.

Maybe those cases aren't very numerous ... and maybe they're distinctive
enough to recognize when creating one ("hmm, I am creating a check or
assign hook that does significant work here, will it be worth exposing
a getter API for the product of the work?").

Regards,
-Chap

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.