Content-Type: text/plain;
	charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3774.200.91.1.1\))
Subject: Re: [PATCH] Add sortsupport for range types and btree_gist
From: Andrey Borodin <x4mmm@yandex-team.ru>
In-Reply-To: <78499920f1e9ed5895423e735272ca2fdb90baf6.camel@oopsware.de>
Date: Sat, 30 Nov 2024 22:18:38 +0500
Cc: Michael Paquier <michael@paquier.xyz>,
 jian he <jian.universality@gmail.com>,
 PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <770DB2DA-0B56-464A-87CE-5620CF12E114@yandex-team.ru>
References: <a380cfd805c37c2aa70b1083d45820e1a7cbe759.camel@oopsware.de>
 <7C9400DF-66FE-4308-BAD8-92FE3D0E55C6@yandex-team.ru>
 <0d9fbd59de15e2fc987a850c3d1f8a79a6ede825.camel@oopsware.de>
 <19c62e62c79b727d2c9a82eaeecfea5017001099.camel@oopsware.de>
 <4BADF8DD-A4D3-43AA-AA08-037F6FFA40BC@yandex-team.ru>
 <61128550659807032dd29c8c8a71674c818cbbb6.camel@oopsware.de>
 <7095B5FA-FADB-4524-B858-ECD14998710B@yandex-team.ru>
 <0176b173adfcff250543fe0aee2ef6137bd71953.camel@oopsware.de>
 <bcdf583f38e7991a374c00ad77bb3adde24d8143.camel@oopsware.de>
 <F6394698-D6BD-469A-BEC4-2D23D348A324@yandex-team.ru>
 <Z0lGG9B1gSJXEwzh@paquier.xyz>
 <78499920f1e9ed5895423e735272ca2fdb90baf6.camel@oopsware.de>
To: Bernd Helmle <mailings@oopsware.de>
Archived-At: <https://www.postgresql.org/message-id/770DB2DA-0B56-464A-87CE-5620CF12E114%40yandex-team.ru>
Precedence: bulk


> On 30 Nov 2024, at 18:14, Bernd Helmle <mailings@oopsware.de> wrote:
>=20
> I don't like this. This smells like we use the wrong tool. Andrey had
> the idea to use them because it looked as a compelling idea to check
> whether sortsupport is used in the code path or not.=20
>=20
> Maybe we should just use a specific DEBUG message and make sure it's
> handled by the tests accordingly.

We tried to go that route, random DEBUG1 lines were breaking tests =
(autovacuum or something like that, there was something =
non-deterministic). I think we can dig exact reason from 2021 thread why =
we abandoned that idea...

I think we have two options:
1. Just do not commit tests. We ran it manually, saw that paths are =
taken, we are happy.
2. Have just one file that builds sorted index on a table with few =
tuples for each data type.

We do not need to test that core sorting (or buffered) build works. =
AFAIR there's plenty of other tests to verify that.

Injection points seemed to me exactly the technogy that could help us to =
have tests for this patch. But at this point It looks like it's =
reasonable to take approach 1, as we did before.


Best regards, Andrey Borodin.=