如何使用分层窗格

A layered pane is a Swing container that provides a third dimension for positioning components: depth, also known as Z order. When adding a component to a layered pane, you specify its depth as an integer. The higher the number, closer the component is to the "top" position within the container. If components overlap, the "closer" components are drawn on top of components at a lower depth. The relationship between components at the same depth is determined by their positions within the depth.

Every Swing container that has a root pane — such as JFrame, JApplet, JDialog, or JInternalFrame — automatically has a layered pane. Most programs do not explicitly use the root pane's layered pane, so this section will not discuss it. You can find information about it in The Root Pane, which provides an overview, and The Layered Pane, which has further details. This section tells you how to create your own layered pane and use it anywhere you can use a regular Swing container.

Swing provides two layered pane classes. The first, JLayeredPane, is the class that root panes use and is the class used by the example in this section. The second, JDesktopPane, is a JLayeredPane subclass that is specialized for the task of holding internal frames. For examples of using JDesktopPane, see How to Use Internal Frames.

Here is a picture of an application that creates a layered pane and places overlapping, colored labels at different depths:

Here is the code from LayeredPaneDemo.java that creates the layered pane:

layeredPane = new JLayeredPane();
layeredPane.setPreferredSize(new Dimension(300, 310));
layeredPane.setBorder(BorderFactory.createTitledBorder(
                                    "Move the Mouse to Move Duke"));
layeredPane.addMouseMotionListener(new MouseMotionAdapter() {
    ...
});

The code uses JLayeredPane's only constructor — the no-argument constructor — to create the layered pane. The rest of the code uses methods inherited from superclasses to give the layered pane a preferred size and a border, and add a mouse-motion listener to it. The mouse-motion listener just moves the Duke image around in response to mouse movement. Although we do not show the code here, the example adds the layered pane to the frame's content pane.

As we will show you a bit later, you add components to a layered pane using an add method. When adding a component to a layered pane, you specify the component depth, and optionally, its position within its depth. The layered pane in the demo program contains six labels — the five colored labels and a sixth one that displays the Duke image. As the program demonstrates, both the depth of a component and its position within that depth can change dynamically.

The rest of this section covers these topics:

• Adding Components and Setting Component Depth
• Setting a Component Position Within Its Depth
• Laying Out Components in a Layered Pane
• The Layered Pane API
• Examples that Use Layered Panes

Adding Components and Setting Component Depth

Here is the code from the sample program that adds the colored labels to the layered pane:

for (int i = 0; i < ...number of labels...; i++) {
    JLabel label = createColoredLabel(...);
    layeredPane.add(label, new Integer(i));
    ...
}

You can find the implementation of the createColoredLabel method in the source code for the program. It just creates an opaque JLabel initialized with a background color, a border, some text, and a size.

The example program uses a two-argument version of the add method. The first argument is the component to add, the second is an Integer object, specifying the depth. This program uses the for loop iteration variable to specify depths. The actual values do not matter much. What matters is the relative value of the depths and that you are consistent within your program in how you use each depth.

As you can see from the example program, if components overlap, components at a higher depth are on top of components at a lower depth. To change a component depth dynamically, use the setLayer method. In the example, the user can change Duke's layer by making a selection from the combo box. Here is the actionPerformed method of the action listener registered on the combo box:

public void actionPerformed(ActionEvent e) {
    int position = onTop.isSelected() ? 0 : 1;
    layeredPane.setLayer(dukeLabel,
                         layerList.getSelectedIndex(),
                         position);
}

The setLayer method used here takes three arguments: the component whose depth is to be set, the new depth, and the position within the depth. JLayeredPane has a two-argument version of setLayer that takes only the component and the new depth. That method puts the component at the bottom position in its depth.

Setting a Component's Position Within Its Depth

The following code creates the label that displays Duke's image, and then adds the label to the layered pane.

final ImageIcon icon = createImageIcon("images/dukeWaveRed.gif");
...
dukeLabel = new JLabel(icon);
...
dukeLabel.setBounds(15, 225,
                    icon.getIconWidth(),
                    icon.getIconHeight());
...
layeredPane.add(dukeLabel, new Integer(2), 0);

This code uses the three-argument version of the add method. The third argument specifies the Duke label position within its depth, which determines the component's relationship with other components at the same depth.

Positions are specified with an int between -1 and (n - 1), where n is the number of components at the depth. Unlike layer numbers, the smaller the position number, the higher the component within its depth. Using -1 is the same as using n - 1; it indicates the bottom-most position. Using 0 specifies that the component should be in the top-most position within its depth. As the following figure shows, with the exception of -1, a lower position number indicates a higher position within a depth.

A component's position within its layer can change dynamically. In the example, you can use the check box to determine whether Duke label is in the top position at its depth. Here's the actionPerformed method for the action listener registered on the check box:

public void actionPerformed(ActionEvent e) {
    if (onTop.isSelected())
        layeredPane.moveToFront(dukeLabel);
    else
        layeredPane.moveToBack(dukeLabel);
}

When the user selects the check box, the moveToFront method moves Duke to the front (position 0). And when the user deselects check box, Duke gets moved to the back with the moveToBack method. You can also use the setPosition method or the three-argument version of setLayer to change a component's position.

Laying Out Components in a Layered Pane

By default a layered pane has no layout manager. This means that you typically have to write the code that positions and sizes the components you put in a layered pane.

The example uses the setBounds method to set the size and position of each of the labels:

dukeLabel.setBounds(15, 225,
                    icon.getIconWidth(),
                    icon.getIconHeight());
...
label.setBounds(origin.x, origin.y, 140, 140);

When the user moves the mouse around, the program calls setPosition to change Duke's position:

dukeLabel.setLocation(e.getX()-XFUDGE, e.getY()-YFUDGE);

Although a layered pane has no layout manager by default, you can still assign a layout manager to the layered pane. All of the layout managers provided by the Java platform arrange the components as if they were all on one layer. Here is a version of the previous demo that sets the layered pane's layout manager to an instance of GridLayout, using that layout manager to lay out six colored labels.

You can find the code for this program in LayeredPaneDemo2.java. You can run LayeredPaneDemo2 ( download JDK 7 or later). If you want to compile the example, consult the example index for a list of all necessary files.

Many programs use intermediate containers (such as panels) and their layout managers to lay out components on the same layer, but use absolute positioning to lay out components on different layers. For more information about absolute positioning, see Doing Without a Layout Manager (Absolute Positioning).

The Layered Pane API

The following tables list the commonly used JLayeredPane constructors and methods. Other methods you are most likely to invoke on a JLayeredPane object are those it inherits from its superclasses, such as setBorder, setPreferredSize, and so on. See The JComponent API for tables of commonly used inherited methods.

The API for using layered pane falls into these categories:

• Creating or Getting a Layered Pane
• Layering Components
• Setting Component's Intra-Layer Positions

Examples that Use Layered Panes

This table shows the examples that use JLayeredPane and where those examples are described.