If you maintain some PostgreSQL extensions which rely on PGXS, a build infrastructure for PostgreSQL, the following commit added to Postgres 12 will be likely something interesting, because it adds new options to control more types of regression tests:

A couple of months ago a thread has begun on the PostgreSQL community mailing lists about a set of problems where it is possible to lock down PostgreSQL from connections just by running a set of queries with any user, having an open connection to the cluster being enough to do a denial of service.

A new set of system roles leveraging the past existence of superuser-only access for some features has been added to leverage security on an instance of PostgreSQL. The following commit has added them, and is part of 11:

While PostgreSQL 10 has introduced the basic infrastructure for in-core support of partitioned tables, many new features are introduced in 11 to make the use of partitioned tables way more instinctive. One of them is an executor improvement for partition pruning, which has been mainly introduced by commit 499be01:

PostgreSQL TAP test has become over the last couple of years an advanced facility gaining in more and more features which allow for complicated regression test scenarios written in perl, with perl 5.8.0 being a minimum requirement. As of now, TAP tests are divided into different pieces in the code tree:

The following commit adds a new feature which is part of Postgres 11, and matters a lot for a couple of tools:

The following commit, which has introduced a new feature for PostgreSQL 11, introduces the possibility to lower a bit the set of permissions around data folders:

PostgreSQL 11 is releasing a small binary called pg_verify_checksums which is able to check if page-level checksums are in a sane state for a cluster which has been cleanly shutdown (this is necessary so as not not face checksum inconsistencies because of torn pages).

INCLUDE clause for indexes (as known as covering indexes), is a new feature of PostgreSQL 11 which has been committed recently:

As the week ends, here is an explanation behind the following commit which has changed a bit the way the libpq routine called PQhost behaves: