diff options
author | Michael Paquier <michael@paquier.xyz> | 2022-03-08 10:12:22 +0900 |
---|---|---|
committer | Michael Paquier <michael@paquier.xyz> | 2022-03-08 10:12:22 +0900 |
commit | 5b81703787bfc1e6072c8e37125eba0c5598b807 (patch) | |
tree | 64fa53f8604641fc8ba17da54cf2b127b5016dac /contrib/jsonb_plpython/jsonb_plpython2u.control | |
parent | d5ed9da41d96988d905b49bebb273a9b2d6e2915 (diff) | |
download | postgresql-5b81703787bfc1e6072c8e37125eba0c5598b807.tar.gz postgresql-5b81703787bfc1e6072c8e37125eba0c5598b807.zip |
Simplify SRFs using materialize mode in contrib/ modules
9e98583 introduced a helper to centralize building their needed state
(tuplestore, tuple descriptors, etc.), checking for any errors. This
commit updates all places of contrib/ that can be switched to use
SetSingleFuncCall() as a drop-in replacement, resulting in the removal
of a lot of boilerplate code in all the modules updated by this commit.
Per analysis, some places remain as they are:
- pg_logdir_ls() in adminpack/ uses historically TYPEFUNC_RECORD as
return type, and I suspect that changing it may cause issues at run-time
with some of its past versions, down to 1.0.
- dblink/ uses a wrapper function doing exactly the work of
SetSingleFuncCall(). Here the switch should be possible, but rather
invasive so it does not seem the extra backpatch maintenance cost.
- tablefunc/, similarly, uses multiple helper functions with portions of
SetSingleFuncCall() spread across the code paths of this module.
Author: Melanie Plageman
Discussion: https://postgr.es/m/CAAKRu_bvDPJoL9mH6eYwvBpPtTGQwbDzfJbCM-OjkSZDu5yTPg@mail.gmail.com
Diffstat (limited to 'contrib/jsonb_plpython/jsonb_plpython2u.control')
0 files changed, 0 insertions, 0 deletions