Re: [PATCH] use separate PartitionedRelOptions structure to store partitioned table options

Michael Paquier <michael@paquier.xyz>

From: Michael Paquier <michael@paquier.xyz>
To: Nikolay Shaplov <dhyan@nataraj.su>
Cc: pgsql-hackers@lists.postgresql.org, Amit Langote <amitlangote09@gmail.com>, Tomas Vondra <tomas.vondra@2ndquadrant.com>, Dent John <denty@qqdd.eu>, "Iwata, Aya" <iwata.aya@jp.fujitsu.com>, Alvaro Herrera <alvherre@2ndquadrant.com>, Dmitry Dolgov <9erthalion6@gmail.com>, Thomas Munro <thomas.munro@gmail.com>
Date: 2019-11-12T04:50:03Z
Lists: pgsql-hackers
On Mon, Nov 11, 2019 at 05:22:32PM +0300, Nikolay Shaplov wrote:
> Here goes v3 patch with the typo fixed

Still one here in v3 of the patch:
+   * Since there are no options for patitioned tables for now, we just do
+   * validation to report incorrect option error and leave.
It looks like you are having a hard time with partitioned tables.

+  if (validate)
+       parseRelOptions(reloptions, validate, RELOPT_KIND_PARTITIONED,
+                       &numoptions);
We have been through great length to have build_reloptions, so
wouldn't it be better to also have this code path do so?  Sure you
need to pass NULL for the parsing table..  But there is a point to
reduce the code paths using directly parseRelOptions() and the
follow-up, expected calls to allocate and to fill in the set of
reloptions.
--
Michael

Commits

  1. Split handling of reloptions for partitioned tables