MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII;
 format=flowed
Content-Transfer-Encoding: 7bit
Date: Mon, 25 Sep 2023 14:49:50 +0900
From: Ryoga Yoshida <bt23yoshidar@oss.nttdata.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: pgsql-hackers@lists.postgresql.org
Subject: Re: Doesn't pgstat_report_wal() handle the argument "force"
 incorrectly
In-Reply-To: <ZREc8dGx2EBlJx4x@paquier.xyz>
References: <f87a4d7be70530606b864fd1df91718c@oss.nttdata.com>
 <ZRDatnXJqlx46zV9@paquier.xyz>
 <2f6f68343fcd244ef7be5dfae8e55f63@oss.nttdata.com>
 <ZREC3d5Aya0KPbSl@paquier.xyz>
 <5adaf4e34fca5612ff506b47f6e81e47@oss.nttdata.com>
 <ZREc8dGx2EBlJx4x@paquier.xyz>
User-Agent: Roundcube Webmail/1.4.11
Message-ID: <d2b1d40be36e8d7823a0b0ddc85f31ab@oss.nttdata.com>
Archived-At: <https://www.postgresql.org/message-id/d2b1d40be36e8d7823a0b0ddc85f31ab%40oss.nttdata.com>
Precedence: bulk

On 2023-09-25 14:38, Michael Paquier wrote:
> We would not wait on the lock if force=false, which would do
> nowait=true.  And !force reads the same to me as force=false.
> 
> Anyway, I am OK to remove this part.  That seems to confuse you, so
> you may not be the only one who would read this comment.

When I first read it, I didn't read that !force as force=false, so 
removing it might be better.

> Another idea would be to do like in pgstat.c by adding the following
> line, then use "nowait" to call each sub-function:
> nowait = !force;
> pgstat_flush_wal(nowait);
> pgstat_flush_io(nowait);

That's very clear and I think it's good.

Ryoga Yoshida