From: Tom Lane <tgl@sss.pgh.pa.us>
To: ma100@hotmail.com
cc: pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #17336: logtape sort performance and overflow
In-reply-to: <17336-fc4e522d26a750fd@postgresql.org>
References: <17336-fc4e522d26a750fd@postgresql.org>
Comments: In-reply-to PG Bug reporting form <noreply@postgresql.org>
	message dated "Tue, 14 Dec 2021 08:53:13 +0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <262744.1639499287.1@sss.pgh.pa.us>
Date: Tue, 14 Dec 2021 11:28:07 -0500
Message-ID: <262745.1639499287@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/262745.1639499287%40sss.pgh.pa.us>
Precedence: bulk

PG Bug reporting form <noreply@postgresql.org> writes:
> In logtape.c,  

> * Each  `swap_nodes`  will execute assignment statement three times, while
> we only need execute once like in tuplesort.c

Yeah, that could be improved.  I see similar coding in binaryheap.c, too.

> * right_offset(unsigned i) parameter i seemed should be `unsigned long`

Ugh, that's surely wrong.  I don't know that anyone could get to >4G entries
in this heap, but it's still wrong.

Thanks for the report!

> Here's a patch

FYI, patches tend not to survive being passed through email unless
you add 'em as attachments.

			regards, tom lane