Re: [picocontainer-dev] Picoweb sessionContainerholder

Paul Hammant <[email protected]> Tue, 16 Sep 2008 12:15:19 -0700
Newsgroups gmane.comp.java.picocontainer.devel
Message-ID <[email protected]>
Look in PicoServletContainerListener dude.

  .start() .stop() and .dispose() are called for new/dead context and  
new/dead session at the right moments

ThreadLocal is used to during the web request only.  At the start of  
the web request a the right session set of components are set up.

- Paul

(am back from vacation now)

On Sep 6, 2008, at 5:59 PM, Michael Rimov wrote:

> Hi guys!
>
> I'm still spelunking picocontainer-web trying to get my head around  
> the code.  I was wondering something:
>
> Why is SessionContainerHolder.lifecycleState threadlocal?
>
> Wouldn't that mean that the lifecycle state got lost in subsequent  
> requests by the client?  (Assuming that the client
> gets serviced by a different thread.)
>
> Since the SessionContainerHolder instance is stored within the  
> Session, isn't that sufficient identity?
> Thanks,
>
>      			-Mike (R)

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email