Looking through the issues - I see that I sent this mail to the EG and
no one else seems to have an opinion. I think that the thin wrapper
approach that Mark mentions is what I would like to go with and throw
the ISE. This allows re-use of the request and response objects while
helping avoid bugs that occur from references. At this time I would like
to restrict it to Async only and not encompass the wider use of ISE for
request / response. We can file another bug to track that for a future
release. Please let me know what you'll think.
- Rajiv
On 12/11/12 4:40 PM, Rajiv Mordani wrote:
> All,
> There is an issue filed in JIRA -
> http://java.net/jira/browse/SERVLET_SPEC-6 on which we have had some
> discussions. To summarize the issue - the semantics of
> AsyncContext#getRequest and getResponse isn't absolutely clear after a
> AsyncContext#complete() or dispatch(). I would like to close out the
> issue one way or another and while I do see both the points of view of
> Mark and Greg, I wanted to see what the rest of the EG thought about
> the issue. I am inclined to clarify the behavior rather than leaving
> it unspecified. Please let me know your thoughts.
>
> Thanks
>
> - Rajiv
>