Santiago Pericas-Geertsen
Santiago Pericas-Geertsen
> Thanks for the clear and concise responses. It really helps with ensuring that the message came across well. > > > Yes, except the client can always use q...
> "thread-safe" _in general_ means that each thread MUST NOT see the changes of the other thread, so it looks like an isolated instance, hence allow sharing of one instance...
> @spericas Agreed in general, but in case we agree that `Response` MAY (instead of MUST) be thread-safe there is no _additional_ cost, at least for Jersey -- as per...
> > @spericas Agreed in general, but in case we agree that `Response` MAY (instead of MUST) be thread-safe there is no _additional_ cost, at least for Jersey -- as...
> > I agree it is not ideal, but you're not addressing the cost factor of making everything thread safe (there's no free lunch). Also, saying nothing or saying that...
IMO it doesn't seem worth adding any more complexity (no matter how much that is) to the algorithm to handle this use case. You may be able to get around...
First, the annotations in your sample are not following the proper case. Second, the spec clearly states in Section 3.5 that "An implementation MUST NOT invoke a method whose effective...
@chkal I'm always in favor for clarifications, especially the `application/octet-stream` part from the HTTP/1.1 spec.
@jansupol Having looked at the TCK changes, the `@Consumes` inheritance on GET methods is really problematic. The change proposed in this issue is technically sound, but presents some practical challenges....
> @jansupol I can say that RESTEasy does not default to `application/octet-stream` if the `Content-Type` header is not defined. Like it seems other implementations, RESTEasy uses `*/*` as the default...