Horton

Re: [WIP]Vertical Clustered Index (columnar store extension) - take2

Timur Magomedov <t.magomedov@postgrespro.ru>

View thread
From: Timur Magomedov <t.magomedov@postgrespro.ru>
To: Peter Smith <smithpb2250@gmail.com>, "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>
Cc: Tomas Vondra <tomas@vondra.me>, "Aya Iwata (Fujitsu)" <iwata.aya@fujitsu.com>, Japin Li <japinli@hotmail.com>, shveta malik <shveta.malik@gmail.com>
Date: 2025-08-15T17:45:10Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Mark ItemPointer arguments as const throughout

  2. Reorganize GUC structs

  3. Eliminate XLOG_HEAP2_VISIBLE from vacuum phase III

  4. Update various forward declarations to use typedef

  5. Pathify RHS unique-ification for semijoin planning

  6. Revert "Don't lock partitions pruned by initial pruning"

  7. For inplace update, send nontransactional invalidations.

  8. Remove nearly-unused SizeOfIptrData macro.

Attachments

On Thu, 2025-08-14 at 11:23 +1000, Peter Smith wrote:
> Here are the latest v18* patches.
> 

Hi Peter!
I've reworked my recent patch [1] so it is now based on v18 and is
divided into several simpler patches. Here they are plus one additional
patch.

0001-Fixed-comment-and-guard-name-in-vci_pg_copy.h.patch
Looks like vci_pg_copy.h was renamed from vci_numeric.h but file name
comment and define guard name were not updated. Fixed it.

0002-Removed-vci_set_merge_and_copy_trans_funcs.patch
Found that vci_set_merge_and_copy_trans_funcs() is not used anywhere,
removed it alogn with the code that was only called inside it.
trans_funcs_table[] now only contains single transfn_oid field, others
(unused) are removed.

0003-Replaced-linear-search-by-switch-case.patch
Replaced linear search inside trans_funcs_table array to more optimal
switch-case.

0004-Removed-worker-name-check-in-lock.c.patch
This is one I'm not sure about.
Found that changes in Postgres core lock.c file check for "backend="
substring in background worker name. There is also a comment in
vci_ros_daemon.c mentioning bgw_name checks of LockAquire(). Names
don't match however. So as far as I understand the check for "backend="
in name is always false since no code in VCI sets bgw_name to something
similar.
This is either forgotten feature that can be easily fixed by removing
bgw_name checks, either some bug, either my misunderstanding.
For the first case, here is a patch that removes bgw_name checks in
lock.c. It makes core patch a bit smaller and not touching lock.c at
all (Yay!).

[1]
https://www.postgresql.org/message-id/8beac6e8a01971b22ccf0f2e2a8eb12a78e5a7ac.camel%40postgrespro.ru

-- 
Regards,
Timur Magomedov