I see, that's however a regression for tomcat users. could the async listener wrapper impl this if a tomcat flag is switched on?
Romain Manni-Bucau @rmannibucau <https://twitter.com/rmannibucau> | Blog <http://rmannibucau.wordpress.com> | Github <https://github.com/rmannibucau> | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Tomitriber <http://www.tomitribe.com> | JavaEE Factory <https://javaeefactory-rmannibucau.rhcloud.com> 2016-05-17 13:03 GMT+02:00 Violeta Georgieva <miles...@gmail.com>: > Hi, > > 2016-05-17 13:46 GMT+03:00 Romain Manni-Bucau <rmannibu...@gmail.com>: > > > > the real issue is adding a listener there trigger onStartAsync()*ONLY* > (no > > onCompleted(), onTimeout(), onError() etc... cause > > org.apache.catalina.core.AsyncContextImpl#setStarted does > listeners.clear(); > > ) > > > > That's behaviour is by spec: > > "public void addListener(AsyncListener) - Registers the given listener for > notifications of onTimeout, onError, onComplete or onStartAsync. The first > three are associated with the most recent asynchronous cycle started by > calling > one of the ServletRequest.startAsync methods. The onStartAsync is > associated to a new asynchronous cycle via one of the > ServletRequest.startAsync > methods." > > so you can do the following: > > public void onStartAsync(AsyncEvent event) throws IOException { > // Re-add this listener to the new AsyncContext > event.getAsyncContext().addListener(this); > } > > Regards, > Violeta > > > > > Romain Manni-Bucau > > @rmannibucau <https://twitter.com/rmannibucau> | Blog > > <http://rmannibucau.wordpress.com> | Github < > https://github.com/rmannibucau> | > > LinkedIn <https://www.linkedin.com/in/rmannibucau> | Tomitriber > > <http://www.tomitribe.com> | JavaEE Factory > > <https://javaeefactory-rmannibucau.rhcloud.com> > > > > 2016-05-17 12:43 GMT+02:00 Rémy Maucherat <r...@apache.org>: > > > > > 2016-05-17 12:37 GMT+02:00 Romain Manni-Bucau <rmannibu...@gmail.com>: > > > > > > > Only found https://bz.apache.org/bugzilla/show_bug.cgi?id=59261 > which > > > > doesn't mean this code would be invalid > > > > > > > > Also if you add a listener eagerly you will have the start event and > > > loose > > > > all "end" (completed, timeout etc...) ones. This doesn't sound very > > > > reliable. > > > > > > > > Can you point me out the related issue please? > > > > > > > > The bugzilla title says: "Request getAsyncContext should throw > > > IllegalStateException if async is not started". So that's what happens > now. > > > > > > Rémy > > > >
