Message-ID: <b8e6905b-bf07-483d-a63d-f43a64606c52@vondra.me>
Date: Fri, 29 Nov 2024 16:32:22 +0100
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: Guidance Needed for Testing PostgreSQL Patch (CF-5044)
To: postgresql_contributors <postgresql_contributors@newtglobalcorp.com>,
 "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Cc: Kiran Suresh Kulkarni <kirans@newtglobalcorp.com>,
 Hari Prasad B <hariprsadb@newtglobalcorp.com>
References: <KL1PR03MB55076EA4B0701791EDD39824E12A2@KL1PR03MB5507.apcprd03.prod.outlook.com>
Content-Language: en-US
From: Tomas Vondra <tomas@vondra.me>
In-Reply-To: <KL1PR03MB55076EA4B0701791EDD39824E12A2@KL1PR03MB5507.apcprd03.prod.outlook.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Archived-At: <https://www.postgresql.org/message-id/b8e6905b-bf07-483d-a63d-f43a64606c52%40vondra.me>
Precedence: bulk

Hi,

On 11/29/24 12:04, postgresql_contributors wrote:
> Hi,
> I recently installed and tested the CF-5044 patch using the query
> provided on the PostgreSQL community page but could not confirm its
> functionality. I followed the standard process:
> 
>  1. Applied the patch using  |git|.
>  2. Rebuilt PostgreSQL from source (|./configure|, |make|, |make install|).
>  3. Tested with the suggested query after restarting PostgreSQL.
> 
> I would appreciate your guidance on verifying patch functionality and
> ensuring compliance with community testing standards. If there are
> additional steps or tools I should use, please let me know.
> Thank you for your support as I represent Newt Global in contributing to
> the PostgreSQL community.

It's not clear to me if you're asking about how to test a particular
patch (I suppose by CF 5044 you mean [1]?), or testing in general.

I don't know what you mean by "suggested query" and why you feel you
could not confirm the functionality of the patch. Can you elaborate?

FWIW the queries mentioned in the e-mail thread are usually included as
a demonstration of what the patch is meant to do, not as a definitive
proof that the patch is working. Likewise, it's not enough to just run
the tests included in the patch - we have automated tools to do that.

Testing a patch usually requires inventing new tests, thinking about
cases the author might not have considered, etc. There's no "standard"
to do that, it's often a surprisingly creative task, and it depends on
what the patch is meant to do.

Maybe start by asking questions about the patch:

- Does the patch address a real problem? Is it useful?

- Does it reliably do what it's supposed to do? Maybe there are false
positives/negatives the author did not consider?

- Could it have some unintended consequences?

- Is is properly testing the change? Maybe it's missing some tests?

- ... and so on ...


Yes, it's going to be very specific to a given patch. I don't think
there's a great "howto" for any patch :-(


Does this answer your question?


[1] https://commitfest.postgresql.org/50/5044/

-- 
Tomas Vondra