Commitfest 2021-11 Patch Triage - Part 2 - Mailing list pgsql-hackers

Below is the second pass over the patches in the current commitfest, this time
the patches which have been around for 5 or 6 commitfests.  The usual
disclaimers that I very well might have missed something still apply.


2773: libpq compression
=======================
This patch intended to provide libpq connection compression to "replace SSL
compression" which was doomed when the patch was written, and have since been
removed altogether.  The initial approach didn't get much traction but there
was significant discussion and work, which has since fizzled out.  The patch
has been updated but there hasn't been meaningful review the past months, the
last comments seem to imply there being a fair amount of questionmarks left in
here.  Robert, having been very involved in this do you have any thoughts on
where we are and where to go (if at all IYO)?


2765: Add extra statistics to explain for Nested Loop
=====================================================
There has been almost exclusively positive feedback on this idea, and the patch
has matured over several rounds of fixes.  Any committer keen on picking this
up and getting it across the finishline?


2780: Allow COPY "text" to output a header and add header matching mode to COPY
FROM
===============================================================================
The original patch was rejected for lacking matching, but it has since been
addressed in an updated patch which has seen a lot of review.  Reading the
thread it seems this should really be in Ready for Committer.


2815: CREATE INDEX CONCURRENTLY on partitioned table
====================================================
This spawned off from the work on REINDEX CONCURRENTLY but seems to have
stalled and is awaiting more review.


2800: CLUSTER on partitioned table
==================================
The the patch above, it too spawned off from the REINDEX CONCURRENTLY thread
but seems to have stalled awaiting more review.  There was scepticism raised
over the usefulness vs code weight tradeoff, but the response to this haven't
really been followed up on.  Alvaro and Michael, do you have any changed
thoughts on Justins patch reorg or do you stil feel this isn't worth going
ahead with?


2897: Faster pglz compression
=============================
Tomas did benchmarking recently which shows promising speedups and is actively
working on getting it ready for commit.


2906: Row filtering for logical replication
===========================================
This patch is undergoing lots of active review.


2831: Extended statistics / estimate Var op Var clauses
=======================================================
The patch and approach has been approved by reviewers, it's down to the final
push over the finishline from what I can gather.  Seems to be awaiting
benchmarks by Tomas on the relative costs of the new operations in planning.


2918: New default role allowing to change per-role/database settings
====================================================================
Stephen, you had comments on the patch which have since been addressed.  Any
thoughts on the new revision?


2911: Dynamic result sets from procedures
=========================================
This patch is gated on the SHOW_ALL_RESULTS patch, and is thus not moving until
there is further progress on that one.


2901: SQL/JSON: functions
2901: SQL/JSON: JSON_TABLE
==========================
Grouping these two together as they are in very similar states.  The patches
seems to mostly move forward by rebases fixing smaller things.  Andrew, do you
have a sense of how far these are from a committable state?


2838: brin: avoid errors processing work item following concurrent reindex
==========================================================================
Alvaro proposed an alternative fix, and the thread has since stalled.  Should
this fix be applied and the CF entry closed?


2877: Fix firing of RI triggers during cross-partition updates of partitioned
tables
=============================================================================
This is fixing a live issue in the backbranches, and would be good to get
closure on.  Amit, you mentioned a problem with the patch involving deferrable
foreign constraints, is this all fixed in the latest posted revision?  (sorry
if that's spelled out and I missed it)


2900: On client connection event trigger
========================================
This implements a client login event trigger commonly used for auditing and
connection setup tasks.  I've reviewed this a fair bit and I think it's pretty
much committable, but there is an ongoing discussion on whether it needs an
escape-hatch GUC to disable such triggers in case they brick the database.  I
am against such a GUC but there has been a lot of support voiced in the thread
for it. If we can put that to rest I think we can close this patch.


2903: Parallel Hash Full Join
=============================
This patch is blocked on a reproducible crash which is being worked on.


2864: Fix pg_rewind race condition just after promotion
2869: Mitigate pg_rewind race condition, if config file is enlarged
concurrently.
===================================================================
Grouping similar entries.  Both fix live bugs (one reported in 2015), and both
patches are reviewed and approved since months ago.  Seems like we should be
able to close these ones in the CF.


2824: Functions for sorting GiST build of gist_btree types
==========================================================
This seems to get close to be committable after bug investigation by Heikki.


2837: pg_stat_statements and "IN" conditions
============================================
This patch aims to consolidate (based on a configurable threshold) SELECT ..
WHERE c IN (a, b, ...) queries to variations of the same query from cluttering
up pg_stat_statements.  Seems like a good idea, and the patch has been tested
and reviewed to the point where it seems ready for committer to me.


2839: possibility to rename root namespace in plpgsql
=====================================================
This patch addresses a feature request for making plpgsql easier to work with,
but the approach has been questioned with the thread stalling since March in
one of these discussion.  Is there interest in picking this up or should we
close it as Returned with Feedback?


2863: Printing backtrace of postgres processes
==============================================
This patch is being actively worked on and reviewed.


2851: Consider parallel for LATERAL subqueries having LIMIT/OFFSET
==================================================================
A nice little fix for supporting parallel query execution for LATERAL ..
LIMIT/OFFSET queries.  Recently marked ready for committer, so let's try and
get this in.


2861: pg_upgrade test for binary compatibility of core data types
=================================================================
This patch is also actively worked on and reviewed, and bits from this thread
have already landed.  Do you think this can land in its entirety in this CF
Michael?


--
Daniel Gustafsson        https://vmware.com/

--
Daniel Gustafsson        https://vmware.com/




pgsql-hackers by date:

Previous
From: vignesh C
Date:
Subject: Re: Printing backtrace of postgres processes
Next
From: Daniel Gustafsson
Date:
Subject: Re: pg_dump new feature: exporting functions only. Bad or good idea ?