I think it is wrong that the getWrappedModel() returns null
it is pretty much the contract that it has to return the model that it wraps
around
it is used in more places. (set model, inner model or getInnerMostModel)
and no if the model is not a inherited model but a component assign aware
model
then i want to handle the detach directly else every kind of implementation
has to
do it themselfs that will be forgotton now and then.
johan
On 7/9/07, Jan Vermeulen <[EMAIL PROTECTED]> wrote:
Following code provokes a NPE the wrappedModel is null:
// also detach the wrapped model of a component assignet wrap (not
inherited)
if (model instanceof IWrapModel &&
!getFlag(FLAG_INHERITABLE_MODEL))
{
((IWrapModel)model).getWrappedModel().detach();
}
We have that case because we have an IContextWrapModel that wraps models
that are stored in a context outside of the component graph. It lazily
looks
up the model by key in the context, and in its detach() it does the
following:
public void detach() {
if (wrappedModel instanceof IContextModel)
wrappedModel = null;
}
Two questions:
(1) Would it not be more correct to leave the 'detach' responsability to
the
wrapModel implementation ?
(2) If not, can there be a check for null pointers in the code above ?
Jan
--
View this message in context:
http://www.nabble.com/NPE-in-Component.detachModel%28%29-when-wrappedModel-is-null-tf4048326.html#a11499084
Sent from the Wicket - Dev mailing list archive at Nabble.com.
