I want to achieve a 'responsive' servlet and perform some heavy task without resolving 
to threads.
As a definition for 'responsive' I take the visual hints the browser (IE in my case) 
shows:
running logo stopped and download bar empty.
I resumed to the pattern described in 
http://www.theserverside.com/patterns/thread.jsp?thread_id=9597.
Basically I commit the response and then, still being inside the service method, I 
perform the heavy task.
This way I trade container responsiveness using http handlers avoiding thread creation.
Now doubts.
Servlet spec. paragraph 5.5 states the conditions signaling the satisfaction to a 
request.
I'm interested only in the termination of the service method, a signal of request 
satisfaction.
I conducted an empirical study: tomcat 3.2.3/4.0 , jo 1.0alpha11, resin2.0.0 all 
stand-alone
with IE starting from 5.0 to 6.0.
As a result I obtained the responsive behaviour as defined before.
>From the spec. response closure implies response flush. The converse isn't however 
>stated.
So this is a call for clarification: has anybody any experience to share ?
I'm interested in some proof of wrongdoing of specs and yes I already searched the 
archive
looking for flushBuffer.

Marco

___________________________________________________________________________
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff SERVLET-INTEREST".

Archives: http://archives.java.sun.com/archives/servlet-interest.html
Resources: http://java.sun.com/products/servlet/external-resources.html
LISTSERV Help: http://www.lsoft.com/manuals/user/user.html

Reply via email to