Re: Include access method in listTables output - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: Include access method in listTables output
Date
Msg-id 20200901044104.GG3511@paquier.xyz
Whole thread Raw
In response to Re: Include access method in listTables output  (Georgios <gkokolatos@protonmail.com>)
Responses Re: Include access method in listTables output  (Georgios <gkokolatos@protonmail.com>)
List pgsql-hackers
On Thu, Aug 20, 2020 at 08:16:19AM +0000, Georgios wrote:
> Please find version 7 attached which hopefully addresses the error along with a proper
> expansion of the test coverage and removal of recently introduced
> whitespace warnings.

+CREATE ROLE    conditional_tableam_display_role;
As a convention, regression tests need to have roles prefixed with
"regress_" or this would cause some buildfarm members to turn red.
Please see -DENFORCE_REGRESSION_TEST_NAME_RESTRICTIONS (you could use
that in your environment for example).

So, as of the tests..  The role gets added to make sure that when
using \d+ on the full schema as well as the various \d*+ variants we
have a consistent owner.  The addition of the relation size for the
sequence and the btree index in the output generated is a problem
though, because that's not really portable when compiling with other
page sizes.  It is true that there are other tests failing in this
case, but I think that we should try to limit that if we can.  In
short, I agree that having some tests is better than nothing, but I
would suggest to reduce their scope, as per the attached.

Adding \dE as there are no foreign tables does not make much sense,
and also I wondered why \dt+ was not added.

Does the attached look correct to you?
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: PATCH: logical_work_mem and logical streaming of large in-progress transactions
Next
From: Andrey Lepikhov
Date:
Subject: Re: Ideas about a better API for postgres_fdw remote estimates