1.1.0 (2020-07-24)

• Untracked dev changes

1.0.9 (2020-07-17)

• Untracked dev changes

1.0.8 (2020-07-11)

• Keep-alive feedback message sent to the source server is now only sent every 5 seconds

1.0.7 (2020-07-01)

• Lsn position is no longer flushed at the end of the run, but only at the start of the next run

1.0.6 (2020-07-01)

• Only search for DB specific slot
• Improve version detection and control
• Clarify Update Bookmark Period

1.0.5 (2020-06-27)

• Lower default poll_total_seconds
• When no data is received, poll every 5 seconds for 15 seconds total
• Do not flush lsn at the end of a run, but at the start of the next run

1.0.4 (2020-06-25)

• Some postgres instances do not have hstore available. do not make that assumption
• hstore rec must exist && must have an installed version to use hstore
• Send all stream schemas before syncing for binlog
• dsn string was causing encoding issues for a client with backslashes in their password
• Change bookmark property from primary key to replication key
• Only fetch the end_lsn if a log based stream is in play. aurora, for instances, does not support lsns
• Do NOT consume message if its lsn > the global end_lsn for that tap run
• Don't break logical replication just because we haven't seen a message in 10 seconds

1.0.3 (2020-05-28)

• Initial release with updated README.md