- Add
KafkaProducer#producePassthrough
for only keeping the passthrough after producing. (#74) - Change
KafkaConsumer#stream
to be an alias forpartitionedStream.parJoinUnbounded
. (#78)- This also removes
ConsumerSettings#fetchTimeout
as it is now unused.
- This also removes
- Change to improve type inference of
ProducerMessage
. (#74, #76)- To support better type inference, a custom
fs2.kafka.ProducerRecord
has been added. - If you were using the Java
ProducerRecord
, change tofs2.kafka.ProducerRecord
.
- To support better type inference, a custom
- Change to replace
Sink
s withPipe
s, and usage ofStream#to
withStream#through
. (#73) - Remove
ProducerMessage#single
,multiple
, andpassthrough
. (#74)- They have been replaced with
ProducerMessage#apply
andProducerMessage#one
. - If you were previously using
single
in isolation, then you can now useone
. - For all other cases, you can now use
ProducerMessage#apply
instead.
- They have been replaced with
- Rename
KafkaProducer#produceBatched
toproduce
. (#74) - Remove the previous
KafkaProducer#produce
.- For previous behavior,
flatten
the result fromproduce
. (#74)
- For previous behavior,
- Change to include current year in license notices. (#72)
Released on 2019-01-18.