Horton

Re: Missing [NO] INDENT flag in XMLSerialize backward parsing

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Jim Jones <jim.jones@uni-muenster.de>
Cc: Andrew Dunstan <andrew@dunslane.net>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Pavel Stehule <pavel.stehule@gmail.com>
Date: 2025-02-21T11:46:03Z
Lists: pgsql-hackers 
On Fri, Feb 21, 2025 at 12:29:12PM +0100, Jim Jones wrote:
> For future reference, what’s the best way to verify this myself? The
> buildfarm was all green.

Don't worry, we are all trapped by that at some point.

One way would be to generate by yourself dumps from an older version
by yourself, as documented by src/bin/pg_upgrade/TESTING, part DETAILS
(the part about USE_MODULE_DB=1 is very important).  Then you can
reuse the SQL files by specifying an old dump files in the TAP tests
of pg_upgrade.  The buildfarm does exactly that.  

For every commit, doing this much testing is a bit annoying, and I
face this issue less than once a year for anything I apply.  So let's
say that I trust the buildfarm to report back at this stage ;)
--
Michael

Commits

  1. Fix cross-version upgrades with XMLSERIALIZE(NO INDENT)

  2. Add missing deparsing of [NO] IDENT to XMLSERIALIZE()