RE: Recommended way to clean up a component?

...or something more obvious.

To me, the differences between pages and components should be minimized
whereever possible, since in the course of refactoring your app you often
extract components from pages. Why not have components be
PageDetachListeners always and just have a detach() method like pages? Is
there a problem I'm not seeing, or is this just a usability tweak that has
been overlooked?

-----Original Message-----
From: Geoff Longman [mailto:[email protected]]
Sent: Monday, November 11, 2002 10:11 AM
To: Phil Surette; Tapestry Developer (E-mail)
Subject: Re: [Tapestry-developer] Recommended way to clean up a component?


By looking at the Vlib example your following the pattern.

I would suggest that explicity registering a DetachListener should not be
required. Could not the framework recognize that a component
implements PageDetachListener and just register it? That would save
developers from having to override finishLoad().

Geoff

----- Original Message ----- 
From: Phil Surette 
To: Tapestry Developer (E-mail) 
Sent: Monday, November 11, 2002 9:30 AM
Subject: [Tapestry-developer] Recommended way to clean up a component?


I have some components that retrieve and store a copy of a db object in a
local variable which is accessed many times by the component. The object is
lazily loaded like this:
Object getCachedDBObject() { 
        if (m_cachedDBObject == null) { 
                m_cachedDBObject = retrieveObjectFromDB(); 
        } 
        return m_cachedDBObject; 
} 
At the end of the current cycle, m_cachedDBObject needs to be nulled out so
the next time the component is used it'll get a new copy. 
What's the best way to do this? Right now I'm making my components implement
PageDetachListener and override finishLoad() like this:
public finishLoad() { 
        getPage().addPageDetachListener(this); 
} 
public void pageDetached(PageEvent pageEvent){ 
        m_cachedDBObject = null; 
} 
Is this the right way to do it? Is there a better way? I tried overriding
cleanUpAfterRender but that didn't seem to work... could easily have been
user error though.