Martin Kouba
Martin Kouba
> The experimental runtime also adds a data class based API so that you can avoid having to work with JSON directly. This is also possible in Java, using e.g....
Also the current return type (`com.google.gson.JsonObject`) of a Java action signature prevents from returning a JSON array. I think it should be `com.google.gson.JsonElement` instead. Of course, this is irrelevant if...
I'd try to avoid `main(String args[])` if possible. Instead, we could allow user to specify a `java.util.function.Function` implementation as a main class. And then when initializing action inspect the actual...
@rabbah Why is that? Is the contract for OW actions defined somewhere?
@kameshsampath Yep, I've seen this code already. We can wrap the resulting array/primitive if necessary right there, e.g. `{result : [...]}`
@rabbah Ok, thanks.
> I assume you also mean there is a a serializer/deserializer from Foo to JSON object? This functionality is provided by Gson. So the only thing we need is to...
I believe that a new well-defined scope would be more appropriate. I do remember the hard times trying to fix all the bugs related to the conversation context in Weld.
> @mkouba just so we can be aware, what sort of issues where there? Asking because I am sure similar problems will arise in this context Well, I don't remember...
> > the conversation context is bound to the Servlet API. > > Why is that ? If you refer to [Conversation Context in Jakarta EE](https://jakarta.ee/specifications/cdi/3.0/jakarta-cdi-spec-3.0#conversation_context_ee) is that not just...