Thank you Jan so much for giving the detailed explanation on the
new Servlet 3.0 APIs. Congratulation for completing the features.
I can already see a big smile on Jack's face.
Would like every one to help testing Servlet 3.0 APIs. You can
read the following spec and try it out.
Here is Servlet 3.0 spec, which is available from
http://jcp.org/aboutJava/communityprocess/pr/jsr315/index.html
Thanks,
Judy
Jan Luehe wrote:
> Hi Jack,
>
> On 01/ 9/09 06:48 PM, Jan Luehe wrote:
>> Hi Jack,
>>
>> On 12/23/08 18:33, Captain Jack Sparrow wrote:
>>> Hi Judy,
>>>
>>> Thank you for your answering.
>>>
>>>
>>>> Let's wait for some time for the answer if async support is in or not .
>>>> Here is the latest promoted build b29 to try.
>>>>
>>> I tried async on the glassfish-v3-web-ea-b29.
>>> Async support is not. You are right!
>>>
>>> java.lang.StackOverflowError has occured. Oh no....
>>>
>
> I wanted to follow up with you and let you know that the upcoming
> nightly build will provide a feature-complete implementation of the
> Servlet 3.0 APIs.
>
> There have been a few changes to the async APIs since you last looked at
> them. Most notably, the javax.servlet.AsyncContext#forward methods
> have been renamed to AsyncContext#dispatch, and a new ASYNC request
> dispatcher type has been added to identify any requests that have been
> dispatched using this API.
>
> Only filters that have been declared to support the ASYNC dispatcher type
> will be considered during an ASYNC dispatch. A filter that has been
> configured for ASYNC and other types of dispatches (e.g., FORWARD) may
> query the request for its current dispatcher type (by calling the new
> ServletRequest#getDispatcherType), and act accordingly.
>
> I should point out that there is one outstanding issue that the
> Servlet EG is trying to close on: the semantics of a zero-argument
> AsyncContext#dispatch. While our impl will perform an ASYNC dispatch
> of the request and response to the current request URI (as returned by
> HttpServletRequest#getRequestURI), which corresponds to the target of
> the most recent FORWARD dispatch (or the original URI if no FORWARD
> dispatches have occurred), some Servlet EG members have been
> arguing that an ASYNC dispatch should always go to the original
> URI. Hopefully, we'll be able to agree on the semantics soon within the
> Servlet EG.
>
> For now, you have to rely on the javax.servlet.ServletRequest and
> javax.servlet.AsycContext javadocs for documentation of the latest
> async features, but we will follow up with blogs in the near future.
>
> Please give our impl a try and let us know if you run into any
> problems.
>
> I also wanted to thank the Grizzly team for their integration support
> of the async APIs with the underlying Grizzly framework.
>
> Thanks,
>
> Jan
>