Packages and namespaces



Packages and namespaces are related concepts. Packages allow you to bundle class definitions together in a way that facilitates code sharing and minimizes naming conflicts. Namespaces allow you to control the visibility of identifiers, such as property and method names, and can be applied to code whether it resides inside or outside a package. Packages let you organize your class files, and namespaces let you manage the visibility of individual properties and methods.

Packages

Packages in ActionScript 3.0 are implemented with namespaces, but are not synonymous with them. When you declare a package, you are implicitly creating a special type of namespace that is guaranteed to be known at compile time. Namespaces, when created explicitly, are not necessarily known at compile time.

The following example uses the package directive to create a simple package containing one class:

package samples 
{ 
    public class SampleCode 
    { 
        public var sampleGreeting:String; 
        public function sampleFunction() 
        { 
            trace(sampleGreeting + " from sampleFunction()"); 
        } 
    } 
}

The name of the class in this example is SampleCode. Because the class is inside the samples package, the compiler automatically qualifies the class name at compile time into its fully qualified name: samples.SampleCode. The compiler also qualifies the names of any properties or methods, so that sampleGreeting and sampleFunction() become samples.SampleCode.sampleGreeting and samples.SampleCode.sampleFunction() , respectively.

Many developers, especially those with Java programming backgrounds, may choose to place only classes at the top level of a package. ActionScript 3.0, however, supports not only classes at the top level of a package, but also variables, functions, and even statements. One advanced use of this feature is to define a namespace at the top level of a package so that it will be available to all classes in that package. Note, however, that only two access specifiers, public and internal , are allowed at the top level of a package. Unlike Java, which allows you to declare nested classes as private, ActionScript 3.0 supports neither nested nor private classes.

In many other ways, however, ActionScript 3.0 packages are similar to packages in the Java programming language. As you can see in the previous example, fully qualified package references are expressed using the dot operator ( . ), just as they are in Java. You can use packages to organize your code into an intuitive hierarchical structure for use by other programmers. This facilitates code sharing by allowing you to create your own package to share with others, and to use packages created by others in your code.

The use of packages also helps to ensure that the identifier names that you use are unique and do not conflict with other identifier names. In fact, some would argue that this is the primary benefit of packages. For example, two programmers who wish to share their code with each other may have each created a class called SampleCode. Without packages, this would create a name conflict, and the only resolution would be to rename one of the classes. With packages, however, the name conflict is easily avoided by placing one, or preferably both, of the classes in packages with unique names.

You can also include embedded dots in your package name to create nested packages. This allows you to create a hierarchical organization of packages. A good example of this is the flash.xml package provided by ActionScript 3.0. The flash.xml package is nested inside the flash package.

The flash.xml package contains the legacy XML parser that was used in previous versions of ActionScript. One reason that it now resides in the flash.xml package is that the name of the legacy XML class conflicts with the name of the new XML class that implements the XML for ECMAScript (E4X) specification functionality available in ActionScript 3.0.

Although moving the legacy XML class into a package is a good first step, most users of the legacy XML classes will import the flash.xml package, which will generate the same name conflict unless you remember to always use the fully qualified name of the legacy XML class (flash.xml.XML). To avoid this situation, the legacy XML class is now named XMLDocument, as the following example shows:

package flash.xml 
{ 
    class XMLDocument {} 
    class XMLNode {} 
    class XMLSocket {} 
}

Most of ActionScript 3.0 is organized under the flash package. For example, the flash.display package contains the display list API, and the flash.events package contains the new event model.

Creating packages

ActionScript 3.0 provides significant flexibility in the way you organize your packages, classes, and source files. Previous versions of ActionScript allowed only one class per source file and required that the name of the source file match the name of the class. ActionScript 3.0 allows you to include multiple classes in one source file, but only one class in each file can be made available to code that is external to that file. In other words, only one class in each file can be declared inside a package declaration. You must declare any additional classes outside your package definition, which makes those classes invisible to code outside that source file. The name of the class declared inside the package definition must match the name of the source file.

ActionScript 3.0 also provides more flexibility in the way you declare packages. In previous versions of ActionScript, packages merely represented directories in which you placed source files, and you didn’t declare packages with the package statement, but rather included the package name as part of the fully qualified class name in your class declaration. Although packages still represent directories in ActionScript 3.0, packages can contain more than just classes. In ActionScript 3.0, you use the package statement to declare a package, which means that you can also declare variables, functions, and namespaces at the top level of a package. You can even include executable statements at the top level of a package. If you do declare variables, functions, or namespaces at the top level of a package, the only attributes available at that level are public and internal , and only one package-level declaration per file can use the public attribute, whether that declaration is a class, variable, function, or namespace.

Packages are useful for organizing your code and for preventing name conflicts. You should not confuse the concept of packages with the unrelated concept of class inheritance. Two classes that reside in the same package will have a namespace in common, but are not necessarily related to each other in any other way. Likewise, a nested package may have no semantic relationship to its parent package.

Importing packages

If you want to use a class that is inside a package, you must import either the package or the specific class. This differs from ActionScript 2.0, where importing classes was optional.

For example, consider the SampleCode class example from earlier in this chapter. If the class resides in a package named samples, you must use one of the following import statements before using the SampleCode class:

import samples.*;

or

import samples.SampleCode;

In general, import statements should be as specific as possible. If you plan to use only the SampleCode class from the samplespackage, you should import only the SampleCode class rather than the entire package to which it belongs. Importing entire packages may lead to unexpected name conflicts.

You must also place the source code that defines the package or class within your classpath . The classpath is a user-defined list of local directory paths that determines where the compiler will search for imported packages and classes. The classpath is sometimes called the build path or source path .

After you have properly imported the class or package, you can use either the fully qualified name of the class (samples.SampleCode) or merely the class name by itself (SampleCode).

Fully qualified names are useful when identically named classes, methods, or properties result in ambiguous code, but can be difficult to manage if used for all identifiers. For example, the use of the fully qualified name results in verbose code when you instantiate a SampleCode class instance:

var mySample:samples.SampleCode = new samples.SampleCode();

As the levels of nested packages increase, the readability of your code decreases. In situations where you are confident that ambiguous identifiers will not be a problem, you can make your code easier to read by using simple identifiers. For example, instantiating a new instance of the SampleCode class is much less verbose if you use only the class identifier:

var mySample:SampleCode = new SampleCode();

If you attempt to use identifier names without first importing the appropriate package or class, the compiler will not be able to find the class definitions. On the other hand, if you do import a package or class, any attempt to define a name that conflicts with an imported name will generate an error.

When a package is created, the default access specifier for all members of that package is internal , which means that, by default, package members are only visible to other members of that package. If you want a class to be available to code outside the package, you must declare that class to be public . For example, the following package contains two classes, SampleCode and CodeFormatter:

// SampleCode.as file 
package samples 
{ 
    public class SampleCode {} 
} 
 
// CodeFormatter.as file 
package samples 
{ 
    class CodeFormatter {} 
}

The SampleCode class is visible outside the package because it is declared as a public class. The CodeFormatter class, however, is visible only within the samples package itself. If you attempt to access the CodeFormatter class outside the samples package, you will generate an error, as the following example shows:

import samples.SampleCode; 
import samples.CodeFormatter; 
var mySample:SampleCode = new SampleCode(); // okay, public class 
var myFormatter:CodeFormatter = new CodeFormatter(); // error

If you want both classes to be available outside the package, you must declare both classes to be public . You cannot apply the public attribute to the package declaration.

Fully qualified names are useful for resolving name conflicts that may occur when using packages. Such a scenario may arise if you import two packages that define classes with the same identifier. For example, consider the following package, which also has a class named SampleCode:

package langref.samples 
{ 
    public class SampleCode {} 
}

If you import both classes, as follows, you will have a name conflict when referring to the SampleCode class:

import samples.SampleCode; 
import langref.samples.SampleCode; 
var mySample:SampleCode = new SampleCode(); // name conflict

The compiler has no way of knowing which SampleCode class to use. To resolve this conflict, you must use the fully qualified name of each class, as follows:

var sample1:samples.SampleCode = new samples.SampleCode(); 
var sample2:langref.samples.SampleCode = new langref.samples.SampleCode();
Note: Programmers with a C++ background often confuse the import statement with #include . The #include directive is necessary in C++ because C++ compilers process one file at a time, and will not look in other files for class definitions unless a header file is explicitly included. ActionScript 3.0 has an include directive, but it is not designed to import classes and packages. To import classes or packages in ActionScript 3.0, you must use the import statement and place the source file that contains the package in the class path.

Namespaces

Namespaces give you control over the visibility of the properties and methods that you create. Think of the public , private , protected, and internal access control specifiers as built-in namespaces. If these predefined access control specifiers do not suit your needs, you can create your own namespaces.

If you are familiar with XML namespaces, much of this discussion will not be new to you, although the syntax and details of the ActionScript implementation are slightly different from those of XML. If you have never worked with namespaces before, the concept itself is straightforward, but the implementation has specific terminology that you will need to learn.

To understand how namespaces work, it helps to know that the name of a property or method always contains two parts: an identifier and a namespace. The identifier is what you generally think of as a name. For example, the identifiers in the following class definition are sampleGreeting and sampleFunction() :

class SampleCode 
{ 
    var sampleGreeting:String; 
    function sampleFunction () { 
        trace(sampleGreeting + " from sampleFunction()"); 
    } 
}

Whenever definitions are not preceded by a namespace attribute, their names are qualified by the default internal namespace, which means they are visible only to callers in the same package. If the compiler is set to strict mode, the compiler issues a warning that the internal namespace applies to any identifier without a namespace attribute. To ensure that an identifier is available everywhere, you must specifically precede the identifier name with the public attribute. In the previous example code, both sampleGreeting and sampleFunction() have a namespace value of internal .

There are three basic steps to follow when using namespaces. First, you must define the namespace using the namespace keyword. For example, the following code defines the version1 namespace:

namespace version1;

Second, you apply your namespace by using it instead of an access control specifier in a property or method declaration. The following example places a function named myFunction() into the version1 namespace:

version1 function myFunction() {}

Third, once you’ve applied the namespace, you can reference it with the use directive or by qualifying the name of an identifier with a namespace. The following example references the myFunction() function through the use directive:

use namespace version1; 
myFunction();

You can also use a qualified name to reference the myFunction() function, as the following example shows:

version1::myFunction();

Defining namespaces

Namespaces contain one value, the Uniform Resource Identifier (URI), which is sometimes called the namespace name . A URI allows you to ensure that your namespace definition is unique.

You create a namespace by declaring a namespace definition in one of two ways. You can either define a namespace with an explicit URI, as you would define an XML namespace, or you can omit the URI. The following example shows how a namespace can be defined using a URI:

namespace flash_proxy = "http://www.adobe.com/flash/proxy";

The URI serves as a unique identification string for that namespace. If you omit the URI, as in the following example, the compiler will create a unique internal identification string in place of the URI. You do not have access to this internal identification string.

namespace flash_proxy;

Once you define a namespace, with or without a URI, that namespace cannot be redefined in the same scope. An attempt to define a namespace that has been defined earlier in the same scope results in a compiler error.

If a namespace is defined within a package or a class, the namespace may not be visible to code outside that package or class unless the appropriate access control specifier is used. For example, the following code shows the flash_proxy namespace defined within the flash.utils package. In the following example, the lack of an access control specifier means that the flash_proxy namespace would be visible only to code within the flash.utils package and would not be visible to any code outside the package:

package flash.utils 
{ 
    namespace flash_proxy; 
}

The following code uses the public attribute to make the flash_proxy namespace visible to code outside the package:

package flash.utils 
{ 
    public namespace flash_proxy; 
}

Applying namespaces

Applying a namespace means placing a definition into a namespace. Definitions that can be placed into namespaces include functions, variables, and constants (you cannot place a class into a custom namespace).

Consider, for example, a function declared using the public access control namespace. Using the public attribute in a function definition places the function into the public namespace, which makes the function available to all code. Once you have defined a namespace, you can use the namespace that you defined the same way you would use the public attribute, and the definition will be available to code that can reference your custom namespace. For example, if you define a namespace example1 , you can add a method called myFunction() using example1 as an attribute, as the following example shows:

namespace example1; 
class someClass 
{ 
    example1 myFunction() {} 
}

Declaring the myFunction() method using the namespace example1 as an attribute means that the method belongs to the example1 namespace.

You should bear in mind the following when applying namespaces:

  • You can apply only one namespace to each declaration.

  • There is no way to apply a namespace attribute to more than one definition at a time. In other words, if you want to apply your namespace to ten different functions, you must add your namespace as an attribute to each of the ten function definitions.

  • If you apply a namespace, you cannot also specify an access control specifier because namespaces and access control specifiers are mutually exclusive. In other words, you cannot declare a function or property as public , private , protected, or internal in addition to applying your namespace.

Referencing namespaces

There is no need to explicitly reference a namespace when you use a method or property declared with any of the access control namespaces, such as public , private , protected , and internal . This is because access to these special namespaces is controlled by context. For example, definitions placed into the private namespace are automatically available to code within the same class. For namespaces that you define, however, such context sensitivity does not exist. In order to use a method or property that you have placed into a custom namespace, you must reference the namespace.

You can reference namespaces with the use namespace directive or you can qualify the name with the namespace using the name qualifier ( :: ) punctuator. Referencing a namespace with the use namespace directive “opens” the namespace, so that it can apply to any identifiers that are not qualified. For example, if you have defined the example1 namespace, you can access names in that namespace by using use namespace example1 :

use namespace example1; 
myFunction();

You can open more than one namespace at a time. Once you open a namespace with use namespace , it remains open throughout the block of code in which it was opened. There is no way to explicitly close a namespace.

Having more than one open namespace, however, increases the likelihood of name conflicts. If you prefer not to open a namespace, you can avoid the use namespace directive by qualifying the method or property name with the namespace and the name qualifier punctuator. For example, the following code shows how you can qualify the name myFunction() with the example1 namespace:

example1::myFunction();

Using namespaces

You can find a real-world example of a namespace that is used to prevent name conflicts in the flash.utils.Proxy class that is part of ActionScript 3.0. The Proxy class, which is the replacement for the Object.__resolve property from ActionScript 2.0, allows you to intercept references to undefined properties or methods before an error occurs. All of the methods of the Proxy class reside in the flash_proxy namespace in order to prevent name conflicts.

To better understand how the flash_proxy namespace is used, you need to understand how to use the Proxy class. The functionality of the Proxy class is available only to classes that inherit from it. In other words, if you want to use the methods of the Proxy class on an object, the object’s class definition must extend the Proxy class. For example, if you want to intercept attempts to call an undefined method, you would extend the Proxy class and then override the callProperty() method of the Proxy class.

You may recall that implementing namespaces is usually a three-step process of defining, applying, and then referencing a namespace. Because you never explicitly call any of the Proxy class methods, however, the flash_proxy namespace is only defined and applied, but never referenced. ActionScript 3.0 defines the flash_proxy namespace and applies it in the Proxy class. Your code only needs to apply the flash_proxy namespace to classes that extend the Proxy class.

The flash_proxy namespace is defined in the flash.utils package in a manner similar to the following:

package flash.utils 
{ 
    public namespace flash_proxy; 
}

The namespace is applied to the methods of the Proxy class as shown in the following excerpt from the Proxy class:

public class Proxy 
{ 
    flash_proxy function callProperty(name:*, ... rest):* 
    flash_proxy function deleteProperty(name:*):Boolean 
    ... 
}

As the following code shows, you must first import both the Proxy class and the flash_proxy namespace. You must then declare your class such that it extends the Proxy class (you must also add the dynamic attribute if you are compiling in strict mode). When you override the callProperty() method, you must use the flash_proxy namespace.

package 
{ 
    import flash.utils.Proxy; 
    import flash.utils.flash_proxy; 
 
    dynamic class MyProxy extends Proxy 
    { 
        flash_proxy override function callProperty(name:*, ...rest):* 
        { 
            trace("method call intercepted: " + name); 
        } 
    } 
}

If you create an instance of the MyProxy class and call an undefined method, such as the testing() method called in the following example, your Proxy object intercepts the method call and executes the statements inside the overridden callProperty() method (in this case, a simple trace() statement).

var mySample:MyProxy = new MyProxy(); 
mySample.testing(); // method call intercepted: testing

There are two advantages to having the methods of the Proxy class inside the flash_proxy namespace. First, having a separate namespace reduces clutter in the public interface of any class that extends the Proxy class. (There are about a dozen methods in the Proxy class that you can override, all of which are not designed to be called directly. Placing all of them in the public namespace could be confusing.) Second, use of the flash_proxy namespace avoids name conflicts in case your Proxy subclass contains instance methods with names that match any of the Proxy class methods. For example, you may want to name one of your own methods callProperty() . The following code is acceptable, because your version of the callProperty() method is in a different namespace:

dynamic class MyProxy extends Proxy 
{ 
    public function callProperty() {} 
    flash_proxy override function callProperty(name:*, ...rest):* 
    { 
        trace("method call intercepted: " + name); 
    } 
}

Namespaces can also be helpful when you want to provide access to methods or properties in a way that cannot be accomplished with the four access control specifiers ( public , private , internal , and protected ). For example, you may have a few utility methods that are spread out across several packages. You want these methods available to all of your packages, but you don’t want the methods to be public. To accomplish this, you can create a new namespace and use it as your own special access control specifier.

The following example uses a user-defined namespace to group together two functions that reside in different packages. By grouping them into the same namespace, you can make both functions visible to a class or package through a single use namespace statement.

This example uses four files to demonstrate the technique. All of the files must be within your classpath. The first file, myInternal.as, is used to define the myInternal namespace. Because the file is in a package named example, you must place the file into a folder named example. The namespace is marked as public so that it can be imported into other packages.

// myInternal.as in folder example 
package example 
{ 
    public namespace myInternal = "http://www.adobe.com/2006/actionscript/examples"; 
}

The second and third files, Utility.as and Helper.as, define the classes that contain methods that should be available to other packages. The Utility class is in the example.alpha package, which means that the file should be placed inside a folder named alpha that is a subfolder of the example folder. The Helper class is in the example.beta package, which means that the file should be placed inside a folder named beta that is also a subfolder of the example folder. Both of these packages, example.alpha and example.beta, must import the namespace before using it.

// Utility.as in the example/alpha folder 
package example.alpha 
{ 
    import example.myInternal; 
     
    public class Utility 
    { 
        private static var _taskCounter:int = 0; 
     
        public static function someTask() 
        { 
            _taskCounter++; 
        } 
         
        myInternal static function get taskCounter():int 
        { 
            return _taskCounter; 
        } 
    } 
} 
 
// Helper.as in the example/beta folder 
package example.beta 
{ 
    import example.myInternal; 
     
    public class Helper 
    { 
        private static var _timeStamp:Date; 
         
        public static function someTask() 
        { 
            _timeStamp = new Date(); 
        } 
         
        myInternal static function get lastCalled():Date 
        { 
            return _timeStamp; 
        } 
    } 
}

The fourth file, NamespaceUseCase.as, is the main application class, and should be a sibling to the example folder. In Adobe Flash CS4 Professional, this class would be used as the document class for the FLA. The NamespaceUseCase class also imports the myInternal namespace and uses it to call the two static methods that reside in the other packages. The example uses static methods only to simplify the code. Both static and instance methods can be placed in the myInternal namespace.

// NamespaceUseCase.as 
package 
{ 
    import flash.display.MovieClip; 
    import example.myInternal; // import namespace 
    import example.alpha.Utility;// import Utility class 
    import example.beta.Helper;// import Helper class 
     
    public class NamespaceUseCase extends MovieClip 
    { 
        public function NamespaceUseCase() 
        { 
            use namespace myInternal; 
             
            Utility.someTask(); 
            Utility.someTask(); 
            trace(Utility.taskCounter); // 2 
             
            Helper.someTask(); 
            trace(Helper.lastCalled); // [time someTask() was last called] 
        } 
    } 
}