java.beans.beancontext
public interface BeanContext extends Collection, BeanContextChild, Visibility, DesignMode
BeanContexts.
Since I can't sprinkle the Collections interface
documentation with special information for BeanContext
implementors, I'll have to document special requirements for
implementors of those functions here.
add() or addAll():
Object into the hierarchy as well as a
BeanContextChild, BeanContext or
BeanContextProxy object.
This way, any Bean can be in the hierarchy.
BeanContext.globalHierarchyLock.
Object if it's already there (only once
per BeanContext).
BeanContextChild implementor, call
setBeanContext() on it. If it's a
BeanContextProxy implementor, call
getBeanContextProxy().setBeanContext() on it.
If setBeanContext() vetoes the change, back out
all changes so far and throw IllegalStateException.
Visibility, call
dontUseGui() or okToUseGui() on it,
depending on whether you (the BeanContext) feel like
allowing it to use the GUI or not.
BeanContextChild or
BeanContextProxy, register yourself (the
BeanContext) as both a
PropertyChangeListener and
VetoableChangeListener on the "beanContext"
property (it may also add itself on any other properties it wishes
to).
BeanContext) are interested in, you may register
yourself to it or register it to you.
java.beans.beancontext.BeanContextMembershipEvent
before exiting. addAll() should wait until everything
is done changing before firing the event (or events) so that if a
failure occurs, the backing-out process can proceed without any
events being fired at all.
remove() or removeAll():
BeanContext.globalHierarchyLock.
Object is not a child of this
BeanContext, just exit without performing any actions.
Object from your collection of children.
BeanContextChild implementor, call
setBeanContext(null) on it. If it's a
BeanContextProxy implementor, call
getBeanContextProxy().setBeanContext(null) on it.
If setBeanContext() vetoes the change, back out
all changes so far and throw IllegalStateException.
Object to listen to you or
registered yourself as a listener on the Object during
add() or addAll(), undo the registration
bycalling the appropriate removeListener() method.
java.beans.beancontext.BeanContextMembershipEvent
before exiting. removeAll() should wait until
everything is done changing before firing the event (or events) so
that if a failure occurs, the backing-out process can proceed
without any events being fired at all.
addAll(), removeAll(),
retainAll() and clear() do not need to be
implemented, but may be if so desired.
Similarly, Visibility and DesignMode methods
should propagate changed values to children that implement interfaces
of the same name.
A hierarchy of beans is mainly useful so that different sets of beans can be established, each with their own set of resources.
Since: JDK1.2
| Field Summary | |
|---|---|
| Object | globalHierarchyLock
The global lock on changing any BeanContext hierarchy.
|
| Method Summary | |
|---|---|
| void | addBeanContextMembershipListener(BeanContextMembershipListener listener)
Add a listener on changes to the membership of this
BeanContext object. |
| URL | getResource(String resourceName, BeanContextChild requestor)
Get a resource. |
| InputStream | getResourceAsStream(String resourceName, BeanContextChild requestor)
Get a resource as a stream. |
| Object | instantiateChild(String beanName)
Instantiate a Bean using this Bean's ClassLoader
and this BeanContext as the parent.
|
| void | removeBeanContextMembershipListener(BeanContextMembershipListener listener)
Remove a listener on changes to the membership of this
BeanContext object. |
Methods that must (or do) synchronize on the global lock:
BeanContext.add() and addAll()
UNKNOWN: fill in the rest of the methods which use the global lock.
BeanContext object.Parameters: listener the listener to add.
BeanContext will typically
call ClassLoader.getResource(), but may do it any
way it wants to. This allows a BeanContext to
have its own set of resources separate from the rest of the
system.
Beans should call this method on their parent rather than the
associated ClassLoader method.
I am assuming, but am not entirely sure, that if a
BeanContext cannot find a resource, its
responsibility is to call the getResource method
of its parent BeanContext.
Parameters: resourceName the name of the resource requested. requestor a reference to the child requesting the resource.
Returns: a URL to the requested resource.
See Also: getResource
BeanContext will
typically call ClassLoader.getResourceAsStream(),
but may do it any way it wants to. This allows a
BeanContext's children to have their own set of
resources separate from the rest of the system.
Beans should call this method on their parent rather than the
associated ClassLoader method.
I am assuming, but am not entirely sure, that if a
BeanContext cannot find a resource, its
responsibility is to call the getResourceAsStream
method of its parent BeanContext.
Parameters: resourceName the name of the resource requested. requestor a reference to the child requesting the resource.
Returns: the requested resource as a stream.
See Also: getResourceAsStream
ClassLoader
and this BeanContext as the parent.
This method exists mainly so that BeanContext
implementations can perform extra actions on Beans that are
created within them.
Parameters: beanName the name of the bean to instantiate
Returns: the created Bean
Throws: IOException if there is an I/O problem during instantiation. ClassNotFoundException if a serialized Bean's class is not found.
See Also: instantiate instantiate
BeanContext object.Parameters: listener the listener to remove.