20thAirhacksQ&A.md

20thAirhacksQ&A.md

Ask questions and see you at November, 2nd at 6.PM. CET: http://www.ustream.tv/channel/adambien Also see archives: airhacks.tv

Hi Adam,
What might be the reason that the SOAP request/response contains the duplicated namespaces? This it seems that appeared after an application server upgrade(weblogic10 -> weblogic12). thanks

Could you continue thread: "Encapsulation violation: with or without DTOs" from last airhacks and explain deeper why you prefer JSON in inter-application communication over strong types? Does it depend if the application has single client (typical frontend and backend separation) or many clients?
What do you think about http://cxf.apache.org/docs/jax-rs-client-api.html#JAX-RSClientAPI-Proxy-basedAPI? (client and server shares Java classes, REST+JSON layer is transparent)

Hello Mr. Bien, I am new to java ee and I am planning to create my first java ee application and I am thinking in go to production too with this application.
This applicatoin is my first java ee app, so what do you think wich application server should I use? it's find Glasfish? or Payara? or Wildfly? What the difference is with tomcat?
Thanks!!

Could you explain a little bit more about how to separate the Boundary of the Control (tier I supose?) ? It's a little bit hard to see without any experience with this pattern wich is the fundaments or when to create the control for the business tier, thanks.

Hi Adam.

I noticed in your repositories that you're using the Maven Release Plugin. Maven's concept of snapshot versioning somewhat conflicts with the idea of continuous delivery where every build is a potential release which again should have a proper version number without snapshot qualifier.

What are your thoughts on release versioning with Maven when it comes to configuring a build plan on your build server?
Do you create two build plans for every project - one that does the snapshots and another one that creates the tag and increments the version number?
Would you create a new release version number for every sprint that is getting tested by the QA team? Or would you rather ship a snapshot to the QA team and then tag it once it is approved?

Cheers
Michael

In java fx I need to change keyboard layout from english language to arabic language when user focus on textfield

Hello.
How to implement a "SessionId" mechanism between a web application (e.g. a single page app in Javascript) and a JAX-RS service. The two apps run in separate containers and the rest service provides somehow a login.

Thanks and kind regards
Markus

Hi,
I was playing around with JCashe and Hazelcast on Payara, works fine with Java SE classes, but not so nice with custom classes:
http://ironicprogrammer.blogspot.se/2015/10/playing-with-hazelcast-on-payara.html

My findings is using the internal HAzelcast gieves me classnotfoutex... due to Hazelcast not recognising my classes since it loads before.
This is a pain in the a** since it is then not possible to work with EntryProcessors, EntryListeners, Loares/Writers for either JCache or Hazelcasts distributed structures as IMap etc.

Do you have a better solution than to kick-start Hazelcast from the top to make it work in a Java EE environment?

Hi Adam,

We are building a Java EE web application which makes moderately to heavy use of DB resources.
A lot of the DB table is configuration stuff.
At the moment we are only able to have around 250 concurrent users in an 8 core 16 GB of RAM Server machine.
We have not done any JVM tuning, nor used batch processing with JPA, so far just plain JPA relationships with defaults.

My question is:
In your projects how many users do you support in a server with similar characteristics?
I know that this is difficult and changes from app to app but I would like to have some numbers on an average boring Java EE application.

How many concurrent users has it been able to support without/before any optimisation performed?

Thanks.

Hello again 😆
What is the best api to implement SSE? REST+Ajax or Websockets ?

Thank you so much 😄

Hello Adam

You've previously show neat tricks, like marshalling/unmarshalling json to pojo's.

Staying in the json realm :

Do you have an easy way to validate input/output (@XmlRootElements), when calling method annotated like this:

    @POST
    @Path("/")
    @Consumes(MediaType.APPLICATION_JSON)
    @Produces(MediaType.APPLICATION_JSON)
    public Out method(In in) {
...

Having looked at examples, I hope I've missed a trick - you can't have to jump through all those hoops.

It would be nice to know that when the method is invoked, all required fields are populated (doing null pointer checks, feels like you're doing it wrong).

See You