Yes, you might want to cover up for that. But the price to pay is really
high. I don't know if any other wicket 2.0 users have experience with
no-nonsense applications, but in our case, the response times are really
unacceptable (sometimes more than 10 seconds). By overwriting the
initModel() method on just a few of the containers that have no own model,
we reduce that time to less than a second.

But I think a framework like Wicket can not leave that up to the user of the
framework.

Jan.


Eelco Hillenius wrote:
> 
>> But I think there is a conceptual error in the component implementation:
>> I
>> think initModel() should not be called from the method getModel(). Once
>> decided that there is no model for a component, it should not try and get
>> one each time we want to get it. I.e., initModel should be called only
>> once
>> for each component. Am I right ?
> 
> Theoretically, people can manually set model to null after the
> component had it's own model first. But that would be quite weird.
> 
> Eelco
> 
> 

-- 
View this message in context: 
http://www.nabble.com/Performance-problems-due-to-Component%3AinitModel%28%29-tf3362476.html#a9360881
Sent from the Wicket - Dev mailing list archive at Nabble.com.

Reply via email to