pgsql: contrib/isn: Make weak mode a GUC setting, and fix related funct - Mailing list pgsql-committers

From Tom Lane
Subject pgsql: contrib/isn: Make weak mode a GUC setting, and fix related funct
Date
Msg-id E1tts3k-002nTZ-0u@gemulon.postgresql.org
Whole thread Raw
List pgsql-committers
contrib/isn: Make weak mode a GUC setting, and fix related functions.

isn's weak mode used to be a simple static variable, settable only
via the isn_weak(boolean) function.  This wasn't optimal, as this
means it doesn't respect transactions nor respond to RESET ALL.

This patch makes isn.weak a GUC parameter instead, so that
it acts like any other user-settable parameter.

The isn_weak() functions are retained for backwards compatibility.
But we must fix their volatility markings: they were marked IMMUTABLE
which is surely incorrect, and PARALLEL RESTRICTED which isn't right
for GUC-related functions either.  Mark isn_weak(boolean) as
VOLATILE and PARALLEL UNSAFE, matching set_config().  Mark isn_weak()
as STABLE and PARALLEL SAFE, matching current_setting().

Reported-by: Viktor Holmberg <v@viktorh.net>
Diagnosed-by: Daniel Gustafsson <daniel@yesql.se>
Author: Viktor Holmberg <v@viktorh.net>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/790bc1f9-74dc-4b50-94d2-8147315b1556@Spark

Branch
------
master

Details
-------
https://git.postgresql.org/pg/commitdiff/44890442398cf3a65230d53167e61905d2b0d348

Modified Files
--------------
contrib/isn/Makefile          |  4 +--
contrib/isn/expected/isn.out  | 44 +++++++++++++++++++++++++++
contrib/isn/isn--1.2--1.3.sql |  7 +++++
contrib/isn/isn.c             | 27 ++++++++++++----
contrib/isn/isn.control       |  2 +-
contrib/isn/isn.h             |  1 -
contrib/isn/meson.build       |  3 +-
contrib/isn/sql/isn.sql       | 13 ++++++++
doc/src/sgml/isn.sgml         | 71 ++++++++++++++++++++++++++++---------------
9 files changed, 137 insertions(+), 35 deletions(-)


pgsql-committers by date:

Previous
From: Alexander Korotkov
Date:
Subject: pgsql: reindexdb: Fix the index-level REINDEX with multiple jobs
Next
From: Tom Lane
Date:
Subject: pgsql: pg_dump, pg_dumpall, pg_restore: Add --no-policies option.