java.beans

Class Introspector

public class Introspector extends Object

Introspector is the class that does the bulk of the design-time work in Java Beans. Every class must have a BeanInfo in order for an RAD tool to use it; but, as promised, you don't have to write the BeanInfo class yourself if you don't want to. All you have to do is call getBeanInfo() in the Introspector and it will use standard JavaBeans-defined method signatures to determine the information about your class.

Don't worry about it too much, though: you can provide JavaBeans with as much customized information as you want, or as little as you want, using the BeanInfo interface (see BeanInfo for details).

Order of Operations

When you call getBeanInfo(class c), the Introspector first searches for BeanInfo class to see if you provided any explicit information. It searches for a class named <bean class name>BeanInfo in different packages, first searching the bean class's package and then moving on to search the beanInfoSearchPath.

If it does not find a BeanInfo class, it acts as though it had found a BeanInfo class returning null from all methods (meaning it should discover everything through Introspection). If it does, then it takes the information it finds in the BeanInfo class to be canonical (that is, the information speaks for its class as well as all superclasses).

When it has introspected the class, calls getBeanInfo(c.getSuperclass) and adds that information to the information it has, not adding to any information it already has that is canonical.

Introspection Design Patterns

When the Introspector goes in to read the class, it follows a well-defined order in order to not leave any methods unaccounted for. Its job is to step over all of the public methods in a class and determine whether they are part of a property, an event, or a method (in that order). Properties:

  1. If there is a public boolean isXXX() method, then XXX is a read-only boolean property. boolean getXXX() may be supplied in addition to this method, although isXXX() is the one that will be used in this case and getXXX() will be ignored. If there is a public void setXXX(boolean) method, it is part of this group and makes it a read-write property.
  2. If there is a public <type> getXXX(int) method, then XXX is a read-only indexed property of type <type>. If there is a public void setXXX(int,<type>) method, then it is a read-write indexed property of type <type>. There may also be a public <type>[] getXXX() and a public void setXXX(<type>) method as well.
  3. If there is a public void setXXX(int,<type>) method, then it is a write-only indexed property of type <type>. There may also be a public <type>[] getXXX() and a public void setXXX(<type>) method as well.
  4. If there is a public <type> getXXX() method, then XXX is a read-only property of type <type>. If there is a public void setXXX(<type>) method, then it will be used for the property and the property will be considered read-write.
  5. If there is a public void setXXX(<type>) method, then as long as XXX is not already used as the name of a property, XXX is assumed to be a write-only property of type <type>.
  6. In all of the above cases, if the setXXX() method throws PropertyVetoException, then the property in question is assumed to be constrained. No properties are ever assumed to be bound (Spec Note: this is not in the spec, it just makes sense). See PropertyDescriptor for a description of bound and constrained properties.
Events:

If there is a pair of methods, public void addXXX(<type>) and public void removeXXX(<type>), where <type> is a descendant of java.util.EventListener, then the pair of methods imply that this Bean will fire events to listeners of type <type>.

If the addXXX() method throws java.util.TooManyListenersException, then the event set is assumed to be unicast. See EventSetDescriptor for a discussion of unicast event sets.

Spec Note: the spec seems to say that the listener type's classname must be equal to the XXX part of addXXX() and removeXXX(), but that is not the case in Sun's implementation, so I am assuming it is not the case in general.

Methods:

Any public methods (including those which were used for Properties or Events) are used as Methods.

Since: JDK1.1

See Also: BeanInfo

Field Summary
static intIGNORE_ALL_BEANINFO
static intIGNORE_IMMEDIATE_BEANINFO
static intUSE_ALL_BEANINFO
Method Summary
static Stringdecapitalize(String name)
A helper method to convert a name to standard Java naming conventions: anything with two capitals as the first two letters remains the same, otherwise the first letter is decapitalized.
static voidflushCaches()
Flush all of the Introspector's internal caches.
static voidflushFromCaches(Class<?> clz)
Flush the Introspector's internal cached information for a given class.
static BeanInfogetBeanInfo(Class<?> beanClass)
Get the BeanInfo for class beanClass, first by looking for explicit information, next by using standard design patterns to determine information about the class.
static BeanInfogetBeanInfo(Class<?> beanClass, int flag)
Returns a {@BeanInfo} instance for the given Bean class where a flag controls the usage of explicit BeanInfo class to retrieve that information.
static BeanInfogetBeanInfo(Class<?> beanClass, Class<?> stopClass)
Get the BeanInfo for class beanClass, first by looking for explicit information, next by using standard design patterns to determine information about the class.
static String[]getBeanInfoSearchPath()
Get the search path for BeanInfo classes.
static voidsetBeanInfoSearchPath(String[] beanInfoSearchPath)
Set the search path for BeanInfo classes.

Field Detail

IGNORE_ALL_BEANINFO

public static final int IGNORE_ALL_BEANINFO

IGNORE_IMMEDIATE_BEANINFO

public static final int IGNORE_IMMEDIATE_BEANINFO

USE_ALL_BEANINFO

public static final int USE_ALL_BEANINFO

Method Detail

decapitalize

public static String decapitalize(String name)
A helper method to convert a name to standard Java naming conventions: anything with two capitals as the first two letters remains the same, otherwise the first letter is decapitalized. URL = URL, I = i, MyMethod = myMethod.

Parameters: name the name to decapitalize.

Returns: the decapitalized name.

flushCaches

public static void flushCaches()
Flush all of the Introspector's internal caches.

Since: 1.2

flushFromCaches

public static void flushFromCaches(Class<?> clz)
Flush the Introspector's internal cached information for a given class.

Parameters: clz the class to be flushed.

Throws: NullPointerException if clz is null.

Since: 1.2

getBeanInfo

public static BeanInfo getBeanInfo(Class<?> beanClass)
Get the BeanInfo for class beanClass, first by looking for explicit information, next by using standard design patterns to determine information about the class.

Parameters: beanClass the class to get BeanInfo about.

Returns: the BeanInfo object representing the class.

getBeanInfo

public static BeanInfo getBeanInfo(Class<?> beanClass, int flag)
Returns a {@BeanInfo} instance for the given Bean class where a flag controls the usage of explicit BeanInfo class to retrieve that information.

You have three options:

With {@link #USE_ALL_BEANINFO} the result is the same as {@link #getBeanInfo(Class)}.

Calling the method with flag set to {@link #IGNORE_IMMEDIATE_BEANINFO} will let it use all explicit BeanInfo classes for the beans superclasses but not for the bean class itself. Furthermore eventset, property and method information is retrieved by introspection if the explicit BeanInfos did not provide such data (ie. return null on {@link BeanInfo.getMethodDescriptors}, {@link BeanInfo.getEventSetDescriptors} and {@link BeanInfo.getPropertyDescriptors}.)

When the method is called with flag

Note: Any unknown value for flag is interpreted as {@link #IGNORE_ALL_BEANINFO}

.

Parameters: beanClass The class whose BeanInfo should be returned. flag Controls the usage of explicit BeanInfo classes.

Returns: A BeanInfo object describing the class.

Throws: IntrospectionException If something goes wrong while retrieving the bean data.

getBeanInfo

public static BeanInfo getBeanInfo(Class<?> beanClass, Class<?> stopClass)
Get the BeanInfo for class beanClass, first by looking for explicit information, next by using standard design patterns to determine information about the class. It crawls up the inheritance tree until it hits topClass.

Parameters: beanClass the Bean class. stopClass the class to stop at.

Returns: the BeanInfo object representing the class.

getBeanInfoSearchPath

public static String[] getBeanInfoSearchPath()
Get the search path for BeanInfo classes.

Returns: the BeanInfo search path.

setBeanInfoSearchPath

public static void setBeanInfoSearchPath(String[] beanInfoSearchPath)
Set the search path for BeanInfo classes.

Parameters: beanInfoSearchPath the new BeanInfo search path.