Date: Fri, 12 Jan 2024 09:46:50 -0600
From: Nathan Bossart <nathandbossart@gmail.com>
To: Aleksander Alekseev <aleksander@timescale.com>
Cc: pgsql-hackers@postgresql.org
Subject: Re: reorganize "Shared Memory and LWLocks" section of docs
Message-ID: <20240112154650.GA3565306@nathanxps13>
References: <20240112041430.GA3557928@nathanxps13>
 <CAJ7c6TNC-tC9JS_X_FSShDEeGujrUdunVdtBBm+YGg8VW+AkFg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAJ7c6TNC-tC9JS_X_FSShDEeGujrUdunVdtBBm+YGg8VW+AkFg@mail.gmail.com>
Archived-At: <https://www.postgresql.org/message-id/20240112154650.GA3565306%40nathanxps13>
Precedence: bulk

Thanks for reviewing.

On Fri, Jan 12, 2024 at 05:12:28PM +0300, Aleksander Alekseev wrote:
> """
> Any registered shmem_startup_hook will be executed shortly after each
> backend attaches to shared memory.
> """
> 
> IMO the word "each" here can give the wrong impression as if there are
> certain guarantees about synchronization between backends. Maybe we
> should change this to simply "... will be executed shortly after
> [the?] backend attaches..."

I see what you mean, but I don't think the problem is the word "each."  I
think the problem is the use of passive voice.  What do you think about
something like

	Each backend will execute the registered shmem_startup_hook shortly
	after it attaches to shared memory.

> """
> should ensure that only one process allocates a new tranche_id
> (LWLockNewTrancheId) and initializes each new LWLock
> (LWLockInitialize).
> """
> 
> Personally I think that reminding the corresponding function name here
> is redundant and complicates reading just a bit. But maybe it's just
> me.

Yeah, I waffled on this one.  I don't mind removing it.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com