Changelog

[0.6.6] - [2019-06-20]

Added

• Auto-coercion of clojure numbers if possible (#135)
• Added more explicit information about commit signing Contributing guide (#150)

Fixed

• Merger instance required for session window aggregation (#142)
• Fixed mis-leading parameter names relating to global ktables (#147)
• Select matching record from union during serialization (#149)
• Fixed typo in one of the code examples in the streams guide (#151)

[0.6.5] - [2019-06-14]

Added

• Add Add :do! and :inspect test commands (#141)

Changed

• Fix regression in multi-topic stream constructor (#143)
• Stop swallowing errors in test-machine (#144)
• Include kafka-streams-test-utils so users don't have to (#138)

[0.6.4] - [2019-05-02]

Added

• Add changelog and contributing files (#122)
• Add test-machine example to the word-count example application (#120)
• Add new arities for aggregate and reduce (#132)
• Add sign-off to contributing file (#123)

Changed

• Ensure user-supplied partitions are cast to int before handing them to the underlying kafka producer (#124)
• Make the test for the rest-proxy transport use keywords to identify topics (#120)
• Resolve dependency conflict reported by lein deps :tree (#125)
• Fix a typo in the service-ready? test fixture (#121)
• Make sure mock dirver is closed after use (#128)
• Fix Jackdaw version for Word Count example (#131)
• Update changelog for 0.6.4 release (#124)

Removed

• Delete a couple of overly verbose logging statements (#124)

[0.6.3] - [2019-03-28]

Added

None

Changed

• Upgrade Kafka dependency to 2.2.0 (#123)

Removed

None

[0.6.2] - [2019-03-21]

Added

• Improvement and clarification of documentation (on-going) (#116, #117, #118)
• Support for including a literal avro schema in the topic definitions resolved by the default resolver (#109)
• Support for including a custom :partition-fn in kafka streams operations that write records (#103)

Changed

• Implement kibit recommendations (#118)
• Word-count example (#108)
  • Log to file instead of stdout
  • A few simplifications in the implementation
• In the streams mock driver, get-records now returns a vector of "datafied" producer-records rather than simply the k,v pairs (part of #103)

Removed

None