Basically, yes. :) I was just pointing out that your classes don't
/have/ to be abstract... there are ways to work around that, if that
happens to be an issue for you. It's nice for maintaining things like
your ide telling you when you haven't implemented all the methods of an
interface, etc. :)
Robert
Darío Vasconcelos wrote:
> Robert,
>
> just curious: why would you do a (MyObject)getProperty("foo") instead
> of just defining a public abstract MyObject getFoo() ? Is it just to
> save yourself from defining these abstract methods in your class?
>
> Dario
>
> On 7/20/05, Robert Zeigler <[EMAIL PROTECTED]> wrote:
>
>>Incidentally, just to point out to new users...
>>You don't /have/ to have an abstract accessor.
>>You can define the property in the .page or .jwc file,
>>and then access it using the "getProperty("foo")" method.
>>(like: MyObject obj = (MyObject) getProperty("myObj");)
>>There is a corresponding setProperty method, as well.
>>If you like having getters/setters, but don't like the abstract methods,
>>you can do something like:
>>1) Define property "prop_foo" in your .page/.jwc file
>>2) Define a concrete accessor in your .java file that look like:
>>public Foo getFoo() {
>> return (Foo) getProperty("prop_foo");
>>}
>>public void setFoo(Foo f) {
>> setProperty("prop_foo",f);
>>}
>>
>>Note that if you choose to do this, then the concrete property should be
>>different than the property defined in your .page/.jwc file.
>>
>>Robert
>>
>>Howard Lewis Ship wrote:
>>
>>>You are not making your properties abstract; you are therefore having
>>>problems because user data is stored in properties of pages (stored
>>>within the page pool), not within the HttpSession as they should be.
>>>
>>>When the pages are culled from the page pool, you lose your server side data.
>>>
>>>You MUST use abstract accessor methods and mark the properties as
>>>persistent. Only in this way will properties be stored into the
>>>HttpSession. This can be done in 3.0 using XML and in 4.0 using
>>>annotations (or XML).
>>>
>>>The fact that your application works at all in development is a
>>>feature. It will neither scale to more users on a single server, or
>>>work in properly in a cluster.
>>>
>>>Not a single example piece of code in TiA or in the Quick Start
>>>tutorial tries to do such an "end run" around Tapestry. Do so at your
>>>own peril.
>>>
>>>The only times I don't use abstract properties is when I want to cache
>>>an expensive object that is not client- or request- dependent. For
>>>example, I may cache an IPropertySelection model, lazily creating it
>>>as necessary.
>>>
>>>
>>>On 7/18/05, Edward Scanzano <[EMAIL PROTECTED]> wrote:
>>>
>>>
>>>>Hi all,
>>>>
>>>>I have a situtation where my application runs normally
>>>>as I click through it. Parameters are being passed to
>>>>components correctly, etc. However, if I let the
>>>>application sit idle for a short time (maybe 10
>>>>minutes, not sure of the exact time) then when I click
>>>>on those same links I get exceptions that are caused
>>>>by component parameters being null. The session has
>>>>not timed out yet.
>>>>
>>>>What is causing this?
>>>>
>>>>Thanks
>>>>E
>>>>
>>>>
>>>>
>>>>---------------------------------------------------------------------
>>>>To unsubscribe, e-mail: [EMAIL PROTECTED]
>>>>For additional commands, e-mail: [EMAIL PROTECTED]
>>>>
>>>>
>>>
>>>
>>>
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: [EMAIL PROTECTED]
>>For additional commands, e-mail: [EMAIL PROTECTED]
>>
>>
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
