pg_dirtyread

• Set SO_TYPE_SEQSCAN in heap_beginscan() to support PG17.
• Exercise toast in tests.

• Add PG16 to github workflow matrix
• Move index test output to separate file and add variant for PG16

• Add functions to check unique indexes for duplicates and FKs for missing rows
• contrib script fixes

• Support PostgreSQL 14

• Support PostgreSQL 13

• regress: Try reading from an index.
• Contrib: Add function to read all good tuples from a table, skipping
  over all tuples that trigger an error, add function to return ctids of
  all tuples in a table that trigger an error.
• README: Fix instructions on setting PG_CONFIG.

• Support PostgreSQL 12.

pg-dirtyread (1.2) unstable; urgency=medium

• Refuse to return the "dead" pseudo column during recovery,
  because GetOldestXmin() asserts !RecoveryInProgress().
  Spotted by Andreas Seltenreich, thanks!

-- Christoph Berg [email protected] Sun, 06 Aug 2017 16:57:41 +0200

pg-dirtyread (1.1) unstable; urgency=medium

• Changes from 1.0:
  • Fix some crashes.
  • Add ability to retrieve system columns such as xmax and ctid.
  • Add "dead" column to allow identification of removed rows

-- Christoph Berg [email protected] Sun, 23 Jul 2017 12:47:01 +0200

This is the original pg_dirtyread code from 2012:
https://github.com/omniti-labs/pgtreats/tree/master/contrib/pg_dirtyread