core/src/main/scala/org/apache/pekko/persistence/cassandra/Extractors.scala (2 lines):
	- line 175: // TODO performance improvement could be to use another query that is not "select *"
	- line 220: // TODO can be removed in 1.0, this is only used during migration from the old version on initial recovery


core/src/main/scala/org/apache/pekko/persistence/cassandra/EventsByTagMigration.scala (2 lines):
	- line 50: // TODO check this is only created once
	- line 133: // TODO might be nice to return a summary of what was done rather than just Done


core/src/main/scala/org/apache/pekko/persistence/cassandra/journal/CassandraJournal.scala (2 lines):
	- line 509: // TODO could "optimize" away deletes that overlap?
	- line 722: // TODO this serialises and re-serialises the messages for fixing tag_views


core/src/main/scala/org/apache/pekko/persistence/cassandra/query/EventsByTagStage.scala (1 line):
	- line 319: // TODO: the looking for missing currently tries to find all the events that are missing. What if there is


cassandra-launcher/src/main/scala/org/apache/pekko/persistence/cassandra/testkit/CassandraLauncher.scala (1 line):
	- line 184: .filterNot(_.endsWith("assembly.jar")) // TODO required?


core/src/main/scala/org/apache/pekko/persistence/cassandra/query/EventsByPersistenceIdStage.scala (1 line):
	- line 384: // TODO if we are far from the partition boundary we could skip this query if refreshInterval.nonEmpty


example/src/main/scala/org/apache/pekko/persistence/cassandra/example/ReadSide.scala (1 line):
	- line 46: settings.nrProcessors - 1, // bug that creates +1 processor FIXME remove in 2.6.5