Node gets invalidated

[Former Member]

In my WD-application I populate a Value Node with new entrys when a user clicks on the AddToList-Button.

I use the addElementMethod of the Node-Object. This works for a couple of times but then the Node is invalidated somehow and the application hangs.

In the onActionAdd-Method I use

IPSPListElement element;

element = wdContext.createPSPListElement();

element.set....

wdContext.nodePSPList().addElement(element);

My ContextNode has just a flat structure with about 5 attributes (no child nodes).

I understood that the supplyFunction is only needed when we deal with hierarchical node-structures?

What could be the reason for this strange behaviour?

Regards

Michael

Message was edited by: Michael Barth

[Former Member]

Hi Stefan,

I don't know why, but your answer disappeared somehow from the forum. So did my reply from this morning.

When the node is invalid my onActionAddPosition-Method is not executed but the wdDoInit-Method. Then the wdDoModifyView and the supplyFunction.

In the SupplyFunction I repopulate the invalid Node now with the content of an extra Vector I created.Here I additionally store the data.

Also I would have to add the last element here since the onActionAddPosition method where I add the elements normally is not executed in the error-case.

This seems redundant to me. I'd like to get the data from the view and put it in the node and rely on the fact that it will be there when I need it. I don't need an extra collection in the first place. Or do I misunderstand the concept?

In my environment the error can be reproduced regularly.

Sometimes within the first couple of Add-Clicks.

Regards

Michael

[Former Member]

Hi Michael,

yes, the forum lacks a little bit on "remembering performance" sometimes.

Anyway, the addElement() you are using in the action handler does (or should) not invalidate the node, so the supply function should not be called again, unless you explicitly call node.invalidate() somewhere.

Since the context is the designated data store of a controller, providing a redundant collection like a vector is normally not necessary.

So you did not misunderstand the concept. It's simply a matter of misleading behaviour in your environment.

The problems begin, when the framework restarts your controller lifecycle and forgets the action.

Isn't there a possibility for you to upgrade your WebAS?

Regards

Stefan

[Former Member]

Hi Stefan,

I came to the same conclusion than you, that it must be a problem of the framework and opened an OSS-note meanwhile.

In the error-case my onActionAddPosition is not called but the wdDoInit Method:

MethodName: wdDoInit

MethodName: wdDoModifiyView

MethodName: supplyPSPList

===> First Click

Method Name: OnActionAddPosition

MethodName: wdDoModifiyView

===> Second Click (Error)

MethodName: wdDoInit

MethodName: wdDoModifiyView

MethodName: supplyPSPList

In the SupplyFunction I repopulate the invalid Node with the whole data that I stored in a Vector. Nevertheless the application hangs although the node should be ok again. The only entry missing is the new one I am adding in the addPositionMethod since this method is not called in the error-case.

Should I really do that in the supplyFunction as well, that means twice? Can't be the solution.

Regards

Michael

[Former Member]

Hi Michael,

the wdDoInit method shouldn't be called twice at all. It looks like a complete "restart" of the View controller lifecycle (and, consequently recalling the supply function also).

So we are not able to solve the problem here. Let's wait and see, what the OSS note results in.

Regards

Stefan

[Former Member]

Hi Michael,

i was able to reproduce the problem, but only once. From my point of view it has nothing to do with your code, it seems to be a problem of the client side framework, which get's confused somehow.

The supply function is not only useful for hierarchical node structures. There are three main benefits using a supply function:

1. You don't have to think about the "right" place and point of time where/when to fill the nodes. The framework will call you when the data is needed.

2. Contexts and context nodes are created in a lazy manner, when they are accessed the first time. If you prefill them in the wdDoInit() method for example without really using them in the ongoing application logic, it wastes runtime and ressources.

3. It's easier to structure and maintain your coding.

But this won't help you really in case of the current problem. The supply function is called, if the node is somehow invalid. Adding does <b>not</b> invalidate the node. And, if you invalidate the node by yourself, you will have to provide the complete list again.

Regards

Stefan