Hi, thanks to the developers and reviewers.
It seems the data type of "num_mxids", which is the result of the "pg_get_multixact_stats" function, does not match the
documentation.
In pg_proc.dat, the data type is int8.
[pg_proc.dat]
+ proname => 'pg_get_multixact_stats', provolatile => 'v',
+ proallargtypes => '{int8,int8,int8,xid}', proargmodes => '{o,o,o,o}',
However, the documentation specifies it as an "integer".
[func-info.sgml]
+ ( <parameter>num_mxids</parameter> <type>integer</type>,
+ <parameter>num_members</parameter> <type>bigint</type>,
There may be debate about whether uint32 should be called integer or bigint, but I think it's better if the
implementationand documentation are consistent. The small attached patch changes the data type of the num_mxids column
inthe documentation to "bigint" to match the implementation.
Regards,
Noriyoshi Shinoda
-----Original Message-----
From: Michael Paquier <michael@paquier.xyz>
Sent: Tuesday, December 30, 2025 4:44 PM
To: Naga Appani <nagnrik@gmail.com>
Cc: Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>; Tomas Vondra <tomas@vondra.me>; Xuneng Zhou <xunengzhou@gmail.com>;
torikoshia<torikoshia@oss.nttdata.com>; Kirill Reshke <reshkekirill@gmail.com>; pgsql-hackers@postgresql.org
Subject: Re: [Proposal] Expose internal MultiXact member count function for efficient monitoring
On Mon, Dec 29, 2025 at 08:57:11PM -0600, Naga Appani wrote:
> The oldest_multixact correctly advances to reflect the cleanup.
>
> Thanks for adding the pg_read_all_stats privilege check!
>
> I think this is ready for RFC.
Thanks for looking. I have done an extra round of brush-up, then applied the set. The buildfarm looks OK with it.
--
Michael
