I struggle to understand why the Halloween Problem is listed as a Postgres issue. The concept that a where condition would run until all rows are condition false is just categorically false for Postgresql.
> This could even cause an infinite loop in some cases where updates continually place the updated record ahead of the scan performing the update operation.
This just is not true for Postgresql and they don't specify anywhere that they switched from talking about Postgresql to a different database.
So either they believe it's an concern with Postgresql or they aren't being clear when they are talking about different databases.
To be honest, just like you I also thought that it was related to PostgreSQL or other popular SQL databases somehow, until I re-read the section multiple times.