Simon,
I hear you. I think the rationale behind the client-side is the heavy
server load. Looks like one of those situations where there's no clear
winner -- highly dependant on the situation... ie, when does the memory
usage of server-side become more trouble than sending all that data on the
wire.
Fortunately, switching back to server-side is trivial -- at least for this
particular situation.
Matt
Simon Kitching-3 wrote:
>
> Matt Tyson wrote:
>> Got it. Here's the params I needed to send with the ajax request:
>>
>> var clientTree = document.getElementById("jsf_tree_64").value;
>> var clientState = document.getElementById("jsf_state_64").value;
>> var viewId = document.getElementById("jsf_viewid").value;
>>
>> dojo.io.bind({
>> url: url,
>> load: function(type, data, evt){
>> treeTools.handleAjaxResponse(data) },
>> mimetype: "text/xml",
>> content: { "ajaxTreeRequest":
>> treeTools.getThisTreeId(clickedNode),
>> "id": id,
>> "name": name,
>> "provider":
>> treeTools.getNodeDataProvider(clickedNode),
>> "jsf_tree_64": clientTree,
>> "jsf_state_64":clientState,
>> "jsf_viewid":viewId}
>> });
>>
>> Only took a day. Thanks for pointing me in the right direction Simon.
>
> The only problem with this is that an AJAX request is supposed to be
> "light", but the saved state trees can be quite large. In the case of
> the app I'm currently working on, often around 100kb :-).
>
> You might want to consider if client-side state is appropriate in your
> case.
>
> Regards,
>
> Simon
>
>
--
View this message in context:
http://www.nabble.com/resotring-state-tf2816777.html#a7864156
Sent from the MyFaces - Users mailing list archive at Nabble.com.
