On 11/1/07, mfs <[EMAIL PROTECTED]> wrote: > > Thanks igor, that really clarifies a lot of things..a couple of followup > question though.. > > So as you said versioning is implemented to handle the back button (and > subsequently "forward" properly, but i wonder if wicket enforces a hit to > the server on a browser back button by setting every response header as > no-cache/no-store (or by setting the page expiry etc)..otherwise wouldnt it > be the browser cache showing the last viewed page, instead of wicket showing > the page from the session (or the serialized version of the same if an older > page)..
have you pressed a back button in a wicket app? it works exactly as you describe, the previous page is retrieved from browser cache. but, when you click a link - which has the page id/version encoded - wicket can pull out the proper page version from the disk to handle the click. > Also when you say just the page objects with their component graph are > stored, i assume also the models associated with the components are also > stored (thats where detach comes into play)...right ? yes. but components can also have state that is shed in detaching, so detaching applies to components and models. -igor > > Thanks alot again and Regards, > > Farhan. > > > > > > igor.vaynberg wrote: > > > > On 11/1/07, mfs <[EMAIL PROTECTED]> wrote: > >> > >> PageMap stores a single instance of a particular page ? OR could it have > >> multiple instances of the same page (browsed by user at different > >> instances > >> against 2 different use-case lets say), > > > > yes, it can have multiple instances of the same page-class > > > > > > > >> and if there could be muliple > >> instances of the same page, its just the page being viewed which is in > >> the > >> PageMap (in session) > >> and the rest are all serialized on the disk, database etc... > > > > in 1.2.x they would all be stored in session, at least the last X > > instances. in 1.3 only the current page is in session (last viewed), > > the rest are swapped to disk. > > > >> ARE they all > >> rendered instances or the components and models are stored seperately...? > > > > they are not rendered instances, they are the page object with its > > component graph, stored in the detached state (detach() is called on > > page and all its children before serialization) > > > >> i am just familiarizing myself with the concept, i mean wouldnt we want > >> to > >> have latest data against each page displayed every time, why version them > >> then ? is it there to handle the back button problem.. > > > > yes, versioning is done to support the back button, so an instance of > > the page can be retrieved in the "correct" state. > > > > -igor > > > > > > > >> > >> ALOT of questions at teh same time..just want to clarify the concept/idea > >> so > >> if someone could be kind enough to shed some light ? > >> > >> Farhan. > >> -- > >> View this message in context: > >> http://www.nabble.com/pageMap-question---tf4735205.html#a13541168 > >> Sent from the Wicket - User mailing list archive at Nabble.com. > >> > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: [EMAIL PROTECTED] > >> For additional commands, e-mail: [EMAIL PROTECTED] > >> > >> > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > > > -- > View this message in context: > http://www.nabble.com/pageMap-question---tf4735205.html#a13542377 > Sent from the Wicket - User mailing list archive at Nabble.com. > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]