knutwalker
commented
8 years ago This is a new feature that was introduced in 2.4.9, I haven't even checked it out by myself yet. From a quick look at the docs and code, it seems like EntityStreamingSupport.json() provides basically the same functionality as this project, with the following differences:
- Akka uses a flexible framing based on a custom parser to identify complete json objects and dispatches to the existing unmarshalling infrastructure for the actual parsing (which might use jackson, spray, jawn, ...)
- this project uses Jawns async parsing facilities and can be configured accordingly
- EntityStreamingSupport requires akka-http as a dependency, which is optional for akka-stream-json (though, I suspect most usages will have akka-http regardless)
- In terms of backpressure and integration into streams, both solutions seem to have similar capabilities, while EntityStreamingSupport has an explicit API for parallel unmarshalling
So, I don't think you can profit from combining these two options but have to choose one over the other. I agree though, that this project can do with some more examples.
schrepfler
It's not clear to me if akka-stream-json can be used together with Source Streaming/JSON streaming as documented by http://doc.akka.io/docs/akka/2.4/scala/http/routing-dsl/source-streaming-support.html. In particular I don't understand if the framing done by akka-stream-json can work together with EntityStreamingSupport.json() (and the configuration of it) or is the framing done explicitly by akka-stream-json? I think an example with a web socket client which opens a connection and streams entities would be useful to illustrate concretely the API's use, either with a concrete case class object or a general Json object, especially as there are not many example around (example, let's not just do the canonical Sink.foreach(println), can the example illustrate working with the case class)?