Skip to content

0.4

Compare
Choose a tag to compare
@SaitTalhaNisanci SaitTalhaNisanci released this 27 Nov 08:51
· 329 commits to master since this release

Upgrade Steps

  • Make sure you have Go 1.9+.

Breaking Changes

  • Change hazelcast.Instance interface as hazelcast.Client. #357.
  • MessageListener's OnMessage method should return an error. #351.
  • AddHeartbeatListener is removed. #360.
  • GetMemberByUUID and GetMember methods shouldn't be in cluster service interface. They are removed. #315.
  • SetInvocationTimeout, SetHeartbeatInterval and SetHeartbeatTimeout are removed from config as they dont exist in Java Client. #281.

New Features

  • SSL/TLS Support With Mutual Authentication(Enterprise)
  • Custom Credentials
  • Reliable Topic
  • Hazelcast Cloud Discovery(Enterprise)
  • Statistics

Bug Fixes

  • Change hazelcast.Instance interface as hazelcast.Client. #357.
  • Possible member addresses that were tried to connect to server could have duplicate addresses. #363.
  • Update last write time of a connection if there wasnt any error during the writing process. #361.
  • Client connection should detect that a connection is reading but not sending any data to server to detect heartbeat idle case. #353.
  • AddHeartbeatListener is removed. #360.
  • When server returns an error client was returning it as a string. When they need to look up for a certain thing in the error message they needed to do string comparison. ServerError interface is added, which has error className, errorCode, stackTrace etc. #352.
  • SetInvocationTimeout, SetHeartbeatInterval and SetHeartbeatTimeout are removed from config as they dont exist in Java Client. #281.
  • MessageListener's OnMessage method should return an error. #351.
  • Default partition table update interval was 5 seconds instead of 10. #341.
  • There was a race condition in cluster service. It was possible that client would try to access reconnectChan when client is shutdown. shutdown and onConnectionClosed methods should be synced. #334.
  • Client was hanging because clusterService was using a global waitgroup. The waitgroup should be unique per cluster service. Initial fetch for partition table is added to partition service like Java Client.#334.
  • Map TryPutWithTimeout method was missing. #320.
  • Multimap Delete method was missing. #317.
  • GetMemberByUUID and GetMember methods shouldn't be in cluster service interface. They are removed. #315.
  • Lifecycle states (Starting, Started, Connected, Disconnected, Shuttingdown, Shutdown) werent public. #314.
  • Connection timeout wasn't used while dialing. #306.
  • If a data was serialized by an unknown serializer and if it was tried to be deserialized on Go side, it would panic..#275.

Improvements

  • Readme Enhancement
  • Hazelcast Property. #281.
  • ServerError interface is added, which has error className, errorCode, stackTrace etc. #352.
  • Godocs #314.