On Tue, 2016-01-19 at 13:46 -0700, tourili wrote: > Thanks a lot for your input Robbie > > Robbie Gemmell wrote > > I'm afraid I don't have any real experience of using Event Hubs, > > Android, Messenger, or combinations thereof, instead mostly > > using/developing proton-j as a pure protocol engine used within > > other > > components such as the JMS client or brokers like ActiveMQ. I can > > however try to answer some of your questions. > > You are already helping me a lot > > > Robbie Gemmell wrote > > The 'AndroidProton' you referred to in your original post looks to > > be > > a wrapper of a JNI based Java binding for proton-c which existed in > > the main repo previously. That JNI binging was removed perhaps a > > couple of years ago or longer. > > You are right, it is a full JNI implementation thanks to the authors' > job > > > Robbie Gemmell wrote > > The reactive work is more developed in the other > > languages/bindings, > > but there is a Reactor impl in proton-j that could form the basis > > ... > > That sounds good. I'm complete newb in the proton project, so I have > to do a > little digging in the project. > Having android native implementation in proton sound interesting in > this > nice project. A lot of devices out there are android based, and Amqp > will be > required for a lot IoT backends (Azure in instance) on such device. > > @tourili >
Possibly not relevant to Java but FYI I'm working on a "minimized" reactor the connection_engine. The idea is to break down the reactor into individual connections that can be managed by other IO/threading frameworks. The problem I'm addressing is having a C based rector dictate how IO is handled and how events are dispatched, since Java has a native reactor you may not have these problems and if you do can probably fix them directly in the java reactor. However I am building experience in breaking apart a "monolithic" reactor so if that sounds like something of interest give me a shout. Cheers, Alan.