I am a little ignorant on the Xdoclets side . So never knew anythign about the webdoclet part.
and now i understand why JAX-RPC lifecycle is not implemented. I know this is a hack and somethign that you all would have thought of .. but do you think we could just implement the lifeservicecyle for services with scope="application" alone. Since we know exactly the time to call the init() .. and the object is kept thorughout till the servlet container is shutdown probably the destroy method can be skipped .. like I said a hack .. but might a lot of people who want to get something initialized in their service. Vidyanand ps: btw .. ur ppt on the "when webservices go bad" was a gr8 read :-). -----Original Message----- From: Steve Loughran [mailto:steve_l@;iseran.com] Sent: Friday, November 08, 2002 2:49 PM To: [EMAIL PROTECTED] Subject: Re: Service initialization ----- Original Message ----- From: "Vidyanand Murunikkara" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Friday, November 08, 2002 1:05 PM Subject: RE: Service initialization >Hi Steve >Yes I did notice what you said and I wasnt disagreeing with you. I was >suggesting using JAX-RPC lifecycle because this would mean just >maintaining the source code and not worry about all the servlet specific >stuff - this to mewould mean maintaing a different web.xml and stuff >which sometimes can get irritating. not as hard as you think now that axis source is marked up to that Xdoclets <webdoclet> can generate web.xml with the addins for axis. So include axis and your webapp in <webdoclet> and all servlets, taglibs, etc, get included. >btw .. do you know whether the JAX-RPC lifecycle has been implemented in axis yet ? No. One problem Axis has is the need to work with servlet2.2 systems that dont have lifecycle events, so it cant just hook in to those. Actually I may add that stuff for the servlet2.3 and say 'upgrade'; the alternative is to write a timer driven housekeeping thread and have many callbacks.
