Message-ID: <66439a51-0ff9-4897-b4e5-b6308693bb7f@iki.fi>
Date: Tue, 7 Apr 2026 17:38:36 +0300
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: Assertion failure in hash_kill_items()
To: Andres Freund <andres@anarazel.de>
Cc: "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>,
 Alexander Kuzmenkov <akuzmenkov@tigerdata.com>,
 Ashutosh Sharma <ashu.coek88@gmail.com>
References: <b9de8d05-3b02-4a27-9b0b-03972fa4bfd3@iki.fi>
 <p45puiwvjtbeff6cx2suwk3i35zubdot6iaa3z6buiqnoix7v6@hzygqumsgome>
Content-Language: en-US
From: Heikki Linnakangas <hlinnaka@iki.fi>
In-Reply-To: <p45puiwvjtbeff6cx2suwk3i35zubdot6iaa3z6buiqnoix7v6@hzygqumsgome>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: 
 <https://www.postgresql.org/message-id/66439a51-0ff9-4897-b4e5-b6308693bb7f%40iki.fi>
Precedence: bulk

On 17/03/2026 19:40, Andres Freund wrote:
> On 2026-03-17 19:15:10 +0200, Heikki Linnakangas wrote:
>> The second patch simplifies the condition in the 'unlock_page' part. This
>> isn't new, and isn't needed to fix the bug, it just caught my eye while
>> looking at this. I don't understand why the condition was the way it was,
>> checking just 'havePin' seems sufficient and more correct to me. Am I
>> missing something?
> 
> I can't see anything either, quite odd.  Most likely explanation seems to be
> that something changed during the development of 7c75ef571579.
> 
> 
> Indeed, the first version of the patch from
>     https://postgr.es/m/CAE9k0Pm3KTx93K8_5j6VMzG4h5F%2BSyknxUwXrN-zqSZ9X8ZS3w%40mail.gmail.com
> was using "if (so->hashso_bucket_buf == so->currPos.buf)" both at the start
> and end of _hash_kill_items(). So likely it was just an accident during patch
> revisions.

Thanks for archeological excavation; pushed this second patch now.

- Heikki