* changing Plugin API for asyncWriteMessages and writeMessages
* passing explicit AtomicWrite that represents the events of
persistAll, or a single event from persist
* journal may reject events before storing them, and that
will result in onPersistRejected (logging) and continue in the
persistent actor
* clarified the semantics with regards to batches and atomic writes,
and failures and rejections in the api docs of asyncWriteMessages
and writeMessages
* adjust the Java plugin API, asyncReplayMessages, doLoadAsync
+ per plugin scoped adapters
+ could be swapped during runtime
+per EventAdapter now has manifest and is configurable ai la serializers
+ json examples in docs
+ including "completely manual" example in case one wants to add
metadata TO the persisted event
+ better error reporting when misconfigured bindings
+ manifest is handled by in memory plugin
- did not check if it works with LevelDB plugin yet
> TODO: json example uses Gson, as that's simplest to do, can we use
+per allows 1:n adapters, multiple adapters can be bound to 1 class
Previously know as [patriknw/akka-data-replication](https://github.com/patriknw/akka-data-replication),
which was originally inspired by [jboner/akka-crdt](https://github.com/jboner/akka-crdt).
The functionality is very similar to akka-data-replication 0.11.
Here is a list of the most important changes:
* The package name changed to `akka.cluster.ddata`
* The extension was renamed to `DistributedData`
* The keys changed from strings to classes with unique identifiers and type information of the data values,
e.g. `ORSetKey[Int]("set2")`
* The optional read consistency parameter was removed from the `Update` message. If you need to read from
other replicas before performing the update you have to first send a `Get` message and then continue with
the ``Update`` when the ``GetSuccess`` is received.
* `BigInt` is used in `GCounter` and `PNCounter` instead of `Long`
* Improvements of java api
* Better documentation
* avoid using Down and Exiting member from being used for joining
* delay shut down of Down member until the information is spread
to all reachable members, e.g. downing several nodes via one node
* akka.cluster.down-removal-margin setting
Margin until shards or singletons that belonged to a
downed/removed partition are created in surviving partition.
Used by singleton and sharding.
* remove the retry count parameters/settings for singleton in
favor of deriving those from the removal-margin
* I think it originated from channels, or some idea that
the sender should be revived (as good as possible) during replay,
but that is pretty useless
* It must still be in PersistentRepr for remote serialization
* I didn't want to change to the built in sender when looping to the
journal because keeping it together with the message makes it easier
to do batching (queueing)
* adjust tck
