Martijn Dashorst wrote:
>
> If you don't want to make the session contain the objects explicitly,
> because the multistep process is only a small part of your
> application, you could use session metadata to store the data
> (temporarily). This does mean that it is serialized with the session,
> but IMO that is a small price to pay (session in a single server
> environment with low memory usage should not often be serialized, only
> when the server is shutdown properly).
>
> I haven't done multi window session storage yet, but that fails mostly
> with detecting the different browser windows. Or you could generate a
> unique number at the start of the wizard and use that to locate each
> session object. But this could lead to significant memory leaks when
> folks randomly start new wizard steps. You should limit the number of
> concurrent wizards per session I think.
>
Thanks Martijn, yes great idea to store my "unit of work" (containing all
changed objects) in the session meta-data with an id and a maximum number.
I was just considering storing the unit-of-work in the PageMap meta-data.
What exactly is the problem with detecting the different browser windows?
When does it break?
JD
--
View this message in context:
http://www.nabble.com/Custom-PageMap-tp19073327p19084492.html
Sent from the Wicket - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
