MIME-Version: 1.0
References: 
 <CAN55FZ1YsO7e7Wr_49gi=2ne+Ov_YPY6q0EE-OxCJ1kWSJfEwg@mail.gmail.com>
 <D18650AE-65BF-431D-9FE0-6AD118F757A3@yesql.se>
 <CAApHDvq-2Zdqytm_Hf3RmVf0qg5PS9jTFAJ5QTc9xH9pwvwDTA@mail.gmail.com>
 <CAApHDvr3zgeb+2cQTsbyRcV2ZkjeMiHyDXYSWB5z71WwSyxwDA@mail.gmail.com>
 <ae9f2fcb-4b24-5bb0-4240-efbbbd944ca1@gmail.com>
 <CAApHDvq7chgw8=7TU4-hLpQNw8e7qN9_oA4ew2UH05RCcihEDQ@mail.gmail.com>
 <c1d63754-cb85-2d8a-8409-bde2c4d2d04b@gmail.com>
 <CAApHDvoYsiFkZpjWeLGejibAXi=T3B70MzFr7z-1EiJ4pVX2hA@mail.gmail.com>
 <f7e514cf-2446-21f1-a5d2-8c089a6e2168@gmail.com>
 <ZtadkEfXje7zvOs4@paquier.xyz>
 <Ztaj7BkN4658OMxF@paquier.xyz>
 <5812a0b9-b0cf-4151-9a14-d9f00e4f2858@gmail.com>
 <2d7d47ef-98a3-47ff-9564-c4c4f7b0889f@gmail.com>
In-Reply-To: <2d7d47ef-98a3-47ff-9564-c4c4f7b0889f@gmail.com>
From: Daniil Davydov <3danissimo@gmail.com>
Date: Mon, 20 Apr 2026 13:11:48 +0700
Message-ID: 
 <CAJDiXggv2M724eyHkUQqPiC3jjNmB0uOTh0utSbzFqAdiA_U0g@mail.gmail.com>
Subject: Re: Typos in the code and README
To: Alexander Lakhin <exclusion@gmail.com>
Cc: Michael Paquier <michael@paquier.xyz>,
 David Rowley <dgrowleyml@gmail.com>,
	PostgreSQL Developers <pgsql-hackers@lists.postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CAJDiXggv2M724eyHkUQqPiC3jjNmB0uOTh0utSbzFqAdiA_U0g%40mail.gmail.com>
Precedence: bulk

Hi,

On Mon, Apr 20, 2026 at 2:00=E2=80=AFAM Alexander Lakhin <exclusion@gmail.c=
om> wrote:
>
> Please look at another set of typos and inconsistencies introduced since
> 2026-01-01

I've found a few other typos. Although they are quite old, I think they sho=
uld
be mentioned here :

1) Typo in tidstore.c
"Find the shared the shared radix tree" --> "Find the shared radix tree".

2) A bit inconsistent comment in twophase.c
```
/*
 * Recreates a state file. This is used in WAL replay and during
 * checkpoint creation.
 *
 * Note: content and len don't include CRC.
 */
static void
RecreateTwoPhaseFile(FullTransactionId fxid, void *content, int len)
```

This comment was introduced in the implementation of the two-phase commit [=
1].
At that time, this function was actually used during replaying the
XLOG_XACT_PREPARE record. Commit [2] changed that code, so now this functio=
n is
used only during checkpoint and restartpoint creation.

If restartpoint creation can be considered as part of the WAL replay, we ca=
n
leave this comment as is. Otherwise, I suggest changing it to something lik=
e
this :
/*
 * Recreates a state file. This is used during checkpoint/restartpoint crea=
tion.
 *
 * Note: content and len don't include CRC.
 */

Another idea - we can stop listing the places of function's usage so that w=
e
don't need to touch this comment later.

[1] d0a89683a3a4dd8e76ef0a99101355999e519df5
[2] 728bd991c3c4389fb39c45dcb0fe57e4a1dccd71

--
Best regards,
Daniil Davydov