Xenon 2.0 suggestions

[georgechrysanthakopoulos]

(feel free to close this and track in Tracker, there might be an item on this already)'

Here some suggestions for next (possibly backwards incompatible) version of Xenon:

1. Move to and require Java 9.0. Lots of perf benefits (reduction in memory due to new handling of char[] arrays that back strings, others). Also, lots of new syntactic sugar to eliminate for() loops.
   Ahead of time compilation nice bonus for embedded systems that need to faster boot up
2. Eliminate Netty. Use new, asynchronous Http/2, web socket, server push HTTP client, built in Java!
3. Add examples using Kotlin. Kotlin runs inside the JVM so will just work with existing Xenon JARs. We just need to show samples. Also, it becomes a viable language for native (no JVM) xenon!

we can meet in person, next time the team is all in Seattle, to discuss this. Or on email.

fyi @asafka @toliaqat @ttddyy

[asafka]

Hey @georgechrysanthakopoulos I'd like to take your offer to meet. I'm actually going to be in Seattle a few days next week, maybe we can meet on Tue (10/24)? anyways, how do I contact you? I can potentially get your contact details from Sufian.

[jvassev]

Hi George,
We can have Kotlin even now. It is flexible enough to add a xenon-kotlin module in xenon-utils repo.
I've done a POC with xenon and Kotlin's async extensions: https://kotlinlang.org/docs/reference/coroutines.html

I will submit an initial version soon!

Thanks,
Julian

[jvassev]

Hi @georgechrysanthakopoulos
Have a look at vmware-archive/xenon-utils@f0f8a96

It still needs a lot of love (for example operationContext is not propagated to coroutines) but we can start adding more Kotlin idioms and/or make Xenon more Kotlin-friendly without to much extension methods.