Martin Kouba
Martin Kouba
> That's where I would like to see a websocket transport to avoid the complexity of streamable http. We added that transport to the roadmap, but it's not there yet....
> Right. And as streamable is the future I think we should just do that and be done with it Yes, until they decprecate it and introduce a new "shiny"...
> The main difference is the server does not care if there was an intialization request it just expects to receive the correct calls confirming to it's capabilities But this...
> They do say it's for simpler use cases, also nothing stops the client from performing the negotiation step, it's just not enforced. Where is it stated? The [Lifecycle part...
> I believe this is partially solved with the new protocol-version header in the latest spec, but it still requires protocol version and capabilities retrieval from initialization for cases where...
FTR there was https://issues.redhat.com/browse/CDI-535. > > We would also need to define behavior for beans that have no priority which can be majority of them - so either say those...
> > > adding a method to the beans which returns a priority and sorting the instances using that > > > > > > This sounds familiar; I think...
> The initializers would be rolled into every constructor right after the super call (this might be tricky to detect though). Hm, that's what `javac` does as well, right?
+1 for default `Set getInterceptorBindings() { return emptySet(); }` and the interceptors spec should state that this method returns an empty set if no interceptor bindings were used to associate...