Hi David,
Thanks for your insight. I may be late to drop this Q in the WO Group. I have designed most of the app and is working fine. Basically this module is just one part of the my application.
I have no idea of JavaClient. Should it be installed in the individual users machines? Can I move my current code to this app? What all changes you anticipate? How can I sysnc it with my regular web app? Please advise.
Thank You, Shravan Kumar. M ------------------------------------
--- On Mon, 7/6/09, David Avendasora <email@hidden> wrote:
From: David Avendasora <email@hidden> Subject: Re: Maintain server state of a page + _javascript_ page navigation To: "shravan kumar"
<email@hidden> Cc: "WO Dev Group" <email@hidden> Date: Monday, July 6, 2009, 6:24 PM
On Jul 6, 2009, at 7:54 AM, shravan kumar wrote:
> Hi Group, > > Could you please advise me a solution to achieve below functionality: > > Scenario: > ------------- > > I have a page which gets all the data like text content [3000 lines or more], images info [about 500 or more] once when the user requests for this page. > > Once this data from the server has arrived, we will be displaying this whole data part by part by splitting this data into sub-pages through _javascript_. So, user actually sees the sub-page rather the whole page received from the server. > > There can be around 40 sub-pages or more. > > While navigating through each of these sub-pages user may perform a server action like edit
some text and save, rotate an image and save, ... and all these actions do not refresh the whole page/ sub-page (like we see in ajax implementations) and just update a particular field. > > I assume that there might be some time-gap say 10 minutes or more between each of the server actions performed by the user. > > During these times as you know, we have to maintain the session of this user in the server (this can be achieved through lets say session timeout value), state of this page in the server, ... > > Problem: > ------------- > How can I maintain the state of this page (generally whenever we do a submit WO app returns with a new context info, correct?), so all my actions are valid and there are issues of Broken pipe/ page not in the cache/ ...
Unless I'm missing something, you can just extend the length of the Session timeout to be what you anticipate the longest a user will need to
respond. 10 minutes I believe is the default, but you can make it anything. Editing Contexts (your user's pools of objects) are maintained as part of the user's Session and you control which editing context is used, or if a new one is created for each action.
Maintaining state is one of WO's core functions, you just have to determine exactly what state and for how long you want to maintain it for your particular situation
> I thought of just pinging the server every few seconds for some dummy data like: how are you doing? or something like gmail does? However, I do not know how to exactly implement this behavior.
You should not need to do this at all. WO will maintain state as long as the session hasn't timed out or you don't start using a different editing context.
With all this said, it sounds like you are trying to implement a rich client in a web UI. Have you considered Java Client at all for this functionality? It
would make managing the client-server interaction much easier since you will actually have EOs on the client that you can manipulate using normal Java instead of having to write _javascript_ to do it.
Dave
|