Feedback-ID: ie3de48e3:Fastmail
Date: Tue, 28 Oct 2025 16:05:34 +0200
From: =?utf-8?Q?=C3=81lvaro?= Herrera <alvherre@kurilemu.de>
To: Michael Paquier <michael@paquier.xyz>
Cc: Heikki Linnakangas <hlinnaka@iki.fi>,
	Quan Zongliang <quanzongliang@yeah.net>,
 Bertrand Drouvot <bertranddrouvot.pg@gmail.com>,
	pgsql-hackers@lists.postgresql.org
Subject: Re: Consistently use the XLogRecPtrIsInvalid() macro
Message-ID: <202510281353.l6bip2btycse@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <aQDAsFbIaj-b7Ag9@paquier.xyz>
Archived-At: 
 <https://www.postgresql.org/message-id/202510281353.l6bip2btycse%40alvherre.pgsql>
Precedence: bulk

On 2025-Oct-28, Michael Paquier wrote:

> The annoying part with eliminating XLogRecPtrIsInvalid() or replacing
> it is that a bunch of external code would be broken, particularly
> backup tools.  I'd rather leave the beast alone.

Well, we don't have to remove it right away.  We can simply not use it.
And maybe if the compiler supports it, make a static inline function in
the header with the [[deprecated]] attribute or
__attribute__((deprecated)) so that the tool developers get a warning
and migrate to using the new one.  Then in a few years we remove it.

BTW we could use Coccinelle to replace all the XLogRecPtrIsInvalid()
calls with !XLogRecPtrIsValid(), as well as all places comparing an LSN
to InvalidXLogRecPtr or literal zero.

-- 
Álvaro Herrera         PostgreSQL Developer  —  https://www.EnterpriseDB.com/