Should we create a JIRA issue to track this?
Stuart
On Sun, Oct 18, 2015 at 11:06 AM, Greg Wilkins <gregw_at_webtide.com> wrote:
>
> On 17 October 2015 at 23:38, Stuart Douglas <sdouglas_at_redhat.com> wrote:
>>
>> This seems reasonable, but we should probably include some kind of thread
>> safety warning in the docs. The listener will be invoked by a different
>> thread to the one that is currently processing, so the end user must take
>> care to make sure the listener/processing code is thread safe.
>
>
> +1
>
>>
>> I am also not convinced that ServletRequestListener is the best place for
>> this. Likely only specific long running requests will need this
>> functionality, and I think that is may make more sense for them to
>> explicitly register a listener with the request.
>
>
> I share this concern. Perhaps simplest to create a new Listener that has to
> be explicitly set via ServletRequest#addListener on a particular request:
>
> interface ServletRequestCancellationListener
> {
> void cancelled(ServletRequestCancellationEvent e);
> }
>
> Plus single method interfaces a sooo fashionable at the moment :)
>
>
>
> --
> Greg Wilkins <gregw@webtide.com> CTO http://webtide.com