Horton

Re: Incorrect logic in XLogNeedsFlush()

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Jeff Davis <pgsql@j-davis.com>
Cc: Melanie Plageman <melanieplageman@gmail.com>, Dilip Kumar <dilipbalaut@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Andres Freund <andres@anarazel.de>
Date: 2025-09-10T23:21:21Z
Lists: pgsql-hackers 
On Wed, Sep 10, 2025 at 09:58:08AM -0700, Jeff Davis wrote:
> It seems like XLogFlush() and XLogNeedsFlush() should use the same
> test, otherwise you could always get some confusing inconsistency.
> Right?

Even if the checks are duplicated (dependency could be documented as
well), it would make sense to me to plant a check of XLogNeedsFlush()
inside XLogFlush(), I think.  I have not tried if some parts of the
tests blow up when trying to do that even after switching
XLogNeedsFlush() to check if WAL inserts are allowed rather than if we
are in recovery.
--
Michael

Commits

  1. Reorder XLogNeedsFlush() checks to be more consistent

  2. Make XLogFlush() and XLogNeedsFlush() decision-making more consistent