Re: Newly created replication slot may be invalidated by checkpoint - Mailing list pgsql-hackers

> On Jan 26, 2026, at 13:35, Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com> wrote:
>
> Dear Hou,
>
> Thanks for updating the patch. I ran tests and reproducer [1] for all versions.
> Confirmed the issue happened before the patch and fixed after applying them.
>
> One difference between master and others is that pg_sync_replication_slots()
> on HEAD is not returned. Because the behavior was changed by 0d2d4a0, not
> related with the fix.
>
> I feel these patches are enough good shape.
>
> [1]:
https://www.postgresql.org/message-id/TY7PR01MB14554DBE84290130EB421DD28F596A%40TY7PR01MB14554.jpnprd01.prod.outlook.com
>
> Best regards,
> Hayato Kuroda
> FUJITSU LIMITED
>

Yeah, I agree v4 is solid. I have a few nitpicks:

1 - commit message
```
WAL reservation and checkpoints, this creates a race conditions where
```

Typo: a race conditions => a race condition

2 - commit message
```
before the the slotsync updates the restart_lsn. This is because in slotsync
```

Typo: the the => the

3 - slotsync.c
```
+     * This can happen because the initial restart_lsn received from the
+     * remote server can precede redo pointer. Therefore, when selecting the
```

I think we need to add “the” before “redo pointer”: the remote server can precede the redo pointer.

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/