The JavaTM Tutorial
Previous Page Lesson Contents Next Page Start of Tutorial > Start of Trail > Start of Lesson Search
Feedback Form

Trail: Creating a GUI with JFC/Swing
Lesson: Using Swing Components

Text Component Features

JTextComponent (in the API reference documentation) is the foundation for Swing's text components, and provides these customizable features for all of its descendants: This section uses the application shown below to explore these capabilities. Although the demo application contains a customized instance of JTextPane, the capabilities discussed in this section are inherited by all of JTextComponent's subclasses.

A snapshot of TextComponentDemo, which contains a customized text pane and a standard text area

The upper text component is the customized text pane. The lower text component is an instance of JTextArea, which serves as a log that reports all changes made to the contents of the text pane. The status line at the bottom of the window reports either the location of the selection or the position of the caret, depending on whether text is selected.

Try this: 
  1. Run TextComponentDemo using JavaTM Web Start (in the Creating a GUI with JFC/Swing trail). Or, to compile and run the example yourself, consult the example index.
  2. Use the mouse to select text and place the cursor in the text pane. Information about the selection and cursor is displayed at the bottom of the window.
  3. Enter text by typing at the keyboard. You can move the caret around using four emacs key bindings: CTRL-B (backward one character), CTRL-F (forward one character), CTRL-N (down one line), and CTRL-P (up one line).
  4. Bring up the Edit menu, and use its various menu items to perform editing on the text in the text pane. Make a selection in the text area at the bottom of the window. Because the text area is uneditable, only some of the Edit menu's commands, like copy-to-clipboard, work. It's important to note, though, that the menu operates on both text components.
  5. Use the items in the Style menu to apply different styles to the text in the text pane.

Using this example application as a reference point, this section covers these topics:

Associating Text Actions with Menus and Buttons

All Swing text components supports standard editing commands such as cut, copy, paste, and inserting characters. Each editing command is represented and implemented by an Action object. (You can learn about actions by reading How to Use Actions (in the Creating a GUI with JFC/Swing trail).) Actions makes it easy for you to associate a command with a GUI component, such as a menu item or button, and therefore build a GUI around a text component.

You can invoke the getActions method on any text component to get an array containing all of the actions supported by it. Often it's convenient to load the array of actions into a HashMap so your program can retrieve an action by name. Here's the code from TextComponentDemo that gets the actions from the text pane and loads them into a HashMap:

private void createActionTable(JTextComponent textComponent) {
    actions = new HashMap();
    Action[] actionsArray = textComponent.getActions();
    for (int i = 0; i < actionsArray.length; i++) {
        Action a = actionsArray[i];
        actions.put(a.getValue(Action.NAME), a);
    }
}    
Here's a convenient method for retrieving an action by its name from the hash map:
private Action getActionByName(String name) {
    return (Action)(actions.get(name));
}
You can use both methods verbatim in your programs.

Now let's look at how the Cut menu item is created and associated with the action of removing text from the text component:

protected JMenu createEditMenu() {
    JMenu menu = new JMenu("Edit");
    ...
    menu.add(getActionByName(DefaultEditorKit.cutAction));
    ...
This code gets the action by name using the handy method shown previously. It then adds the action to the menu. That's all you need to do. The menu and the action take care of everything else. You'll note that the name of the action comes from DefaultEditorKit (in the API reference documentation). This kit provides actions for basic text editing and is the superclass for all the editor kits provided by Swing. So its capabilities are available to all text components unless overridden by a customization.

For efficiency, text components share actions. The Action object returned by getActionByName(DefaultEditorKit.cutAction) is shared by the uneditable JTextArea at the bottom of the window. This has two important ramifications:

Here's the code that creates the Style menu and puts the Bold menu item in it:
protected JMenu createStyleMenu() {
    JMenu menu = new JMenu("Style");
 
    Action action = new StyledEditorKit.BoldAction();
    action.putValue(Action.NAME, "Bold");
    menu.add(action);
    ...
The StyledEditorKit provides Action subclasses to implement editing commands for styled text. You'll note that instead of getting the action from the editor kit, this code creates an instance of the BoldAction class. Thus, this action is not shared with any other text component, and changing its name won't affect any other text component.

Associating Text Actions with Key Strokes

In addition to associating an action with a GUI component, you can also associate an action with a key stroke, using a text component's input map. Input maps are described in How to Use Key Bindings (in the Creating a GUI with JFC/Swing trail).

The text pane in the TextComponentDemo supports four key bindings not provided by default.

The following code adds the CTRL-B key binding to the text pane. The code for adding the other three is similar.
InputMap inputMap = textPane.getInputMap();

KeyStroke key = KeyStroke.getKeyStroke(KeyEvent.VK_B,
                                       Event.CTRL_MASK);
inputMap.put(key, DefaultEditorKit.backwardAction);

The code starts off by getting the text component's input map. Next, it gets a KeyStroke (in the API reference documentation) object representing the CTRL-B key sequence. Finally, the code binds the key stroke to the Action that moves the cursor backward.


Version note: Before 1.3, you needed to use a JTextComponent feature called a keymap to associate key strokes with actions. The keymap API still exists and works, but it is now implemented in terms of the key binding infrastructure added in 1.3. Here's the pre-1.3 equivalent of the preceding code:
//PRE-1.3 CODE
Keymap keymap = textPane.addKeymap("MyEmacsBindings",
				   textPane.getKeymap());

Action action = getActionByName(DefaultEditorKit.backwardAction);
KeyStroke key = KeyStroke.getKeyStroke(KeyEvent.VK_B,
                                       Event.CTRL_MASK);
keymap.addActionForKeyStroke(key, action);
...
textPane.setKeymap(keymap);

Implementing Undo and Redo

Implementing undo and redo has two parts:

Part 1: Remembering Undoable Edits
To support undo and redo, a text component must remember each edit that occurs, the order of edits, and what it takes to undo each edit. The example program uses an instance of the UndoManager (in the API reference documentation) class to manage its list of undoable edits. The undo manager is created where the member variables are declared:

protected UndoManager undo = new UndoManager();
Now, let's look at how the program finds out about undoable edits and adds them to the undo manager.

A document notifies interested listeners whenever an undoable edit occurs on its content. An important step in implementing undo and redo is to register an undoable edit listener on the document of the text component. The following code adds an instance of MyUndoableEditListener to the text pane's document:

lsd.addUndoableEditListener(new MyUndoableEditListener());
The undoable edit listener used in our example adds the edit to the undo manager's list:
protected class MyUndoableEditListener
          implements UndoableEditListener
{
    public void undoableEditHappened(UndoableEditEvent e) {
        //Remember the edit and update the menus
        undo.addEdit(e.getEdit());
        undoAction.updateUndoState();
        redoAction.updateRedoState();
    }
}  
Note that this method updates two objects: undoAction and redoAction. These are the action objects attached to the Undo and Redo menu items, respectively. The next step shows you how the menu items are created and the implementation of the two actions. For general information about undoable edit listeners and undoable edit events, see How to Write an Undoable Edit Listener.

Note: By default, undoable edits can be as fine-grained as single key presses. It is possible, with some effort, to group edits so that (for example) a series of key presses is combined into one undoable edit. The implementation would require defining a class that intercepts undoable edit events from the document, combining them if appropriate and forwarding the results to your undoable edit listener.

Part 2: Implementing the Undo and Redo Commands
The first step in this part of implementing undo and redo is to create the actions to put in the Edit menu.

JMenu menu = new JMenu("Edit");

//Undo and redo are actions of our own creation
undoAction = new UndoAction();
menu.add(undoAction);

redoAction = new RedoAction();
menu.add(redoAction);
...
The undo and redo actions are implemented by custom AbstractAction subclasses: UndoAction and RedoAction, respectively. These classes are inner classes of the example's primary class.

When the user invokes the Undo command, UndoAction's actionPerformed method, shown here, gets called:

public void actionPerformed(ActionEvent e) {
    try {
        undo.undo();
    } catch (CannotUndoException ex) {
        System.out.println("Unable to undo: " + ex);
        ex.printStackTrace();
    }
    updateUndoState();
    redoAction.updateRedoState();
}
This method calls the undo manager's undo method and updates the menu items to reflect the new undo/redo state.

Similarly, when the user invokes the Redo command, the actionPerformed method in RedoAction gets called:

public void actionPerformed(ActionEvent e) {
    try {
        undo.redo();
    } catch (CannotRedoException ex) {
        System.out.println("Unable to redo: " + ex);
        ex.printStackTrace();
    }
    updateRedoState();
    undoAction.updateUndoState();
}
This method is similar except that it calls the undo manager's redo method.

Much of the code in the UndoAction and RedoAction classes is dedicated to enabling and disabling the actions as appropriate for the current state, and changing the names of the menu items to reflect the edit to be undone or redone.


Note:  The implementation of undo and redo in TextComponentDemo was taken from the NotePad demo that comes with the Java 2 SDK. Many programmers will also be able to copy this implementation of undo/redo without modification.

Concepts: About Documents

Like other Swing components, a text component separates its data (known as the model) from its view of the data. If you are not yet familiar with the model-view split used by Swing components, refer to Using Models.

A text component's model is known as a document and is an instance of a class that implements the Document (in the API reference documentation) interface. A document provides these services for a text component:

The Swing text package contains a subinterface of Document, StyledDocument (in the API reference documentation), that adds support for marking up the text with styles. One JTextComponent subclass, JTextPane, requires that its document be a StyledDocument rather than merely a Document.

The javax.swing.text package provides the following hierarchy of document classes, which implement specialized documents for the various JTextComponent subclasses:

The hierarchy of document classes that javax.swing.text provides.

A PlainDocument is the default document for text fields, password fields, and text areas. PlainDocument provides a basic container for text where all the text is displayed in the same font. Even though an editor pane is a styled text component, it uses an instance of PlainDocument by default. The default document for a standard JTextPane in an instance of DefaultStyledDocument — a container for styled text in no particular format. However, the document instance used by any particular editor pane or text pane depends on the type of content bound to it. If you use setPage to load text into an editor pane or text pane, the document instance used by the pane might change. Refer to How to Use Editor Panes and Text Panes for details.

Although you can set the document of a text component, it's usually easier to let it be set automatically, and use a document filter if necessary to change how the text component's data is set. You can implement certain customizations either by installing a document filter or by replacing a text component's document with one of your own, For example, the text pane in TextComponentDemo has a document filter that limits the number of characters the text pane can contain.

Implementing a Document Filter

To implement a document filter, you create a subclass of DocumentFilter (in the API reference documentation) and then attach it to a document using the setFilter method defined in AbstractDocument. Although it's possible to have documents that don't descend from AbstractDocument, by default Swing text components use AbstractDocument subclasses for their documents.

The TextComponentDemo application has a document filter, DocumentSizeFilter (in a .java source file), that limits the number of characters that the text pane can contain. Here's the code that creates the filter and attaches it to the text pane's document:

...//Where member variables are declared:
JTextPane textPane;
AbstractDocument doc;
static final int MAX_CHARACTERS = 300;
...
textPane = new JTextPane();
...
StyledDocument styledDoc = textPane.getStyledDocument();
if (styledDoc instanceof AbstractDocument) {
    doc = (AbstractDocument)styledDoc;
    doc.setDocumentFilter(new DocumentSizeFilter(MAX_CHARACTERS));
} 
To limit the characters allowed in the document, DocumentSizeFilter overrides DocumentFilter's insertString method, which is called each time text is inserted into the document. It also overrides the replace method, which is most likely to be called when the user changes text by pasting it in. In general, text insertion can be the result of the user typing or pasting text in, or because of a call to setText. Here is DocumentSizeFilter's implementation of insertString:
public void insertString(FilterBypass fb, int offs,
                         String str, AttributeSet a)
    throws BadLocationException {

    if ((fb.getDocument().getLength() + str.length()) <= maxCharacters)
        super.insertString(fb, offs, str, a);
    else
        Toolkit.getDefaultToolkit().beep();
}

The code for replace is similar. The FilterBypass (in the API reference documentation) parameter to the methods defined by DocumentFilter is simply an object that enables updating the document in a thread-safe way.

Because the preceding document filter is only concerned about additions to the document's data, it only overrides the insertString and replace methods. Most document filters would override DocumentFilter's remove method, as well.

Listening for Changes on a Document

You can register two different types of listeners on a document: document listeners and undoable edit listeners. This subsection covers document listeners. For information about undoable edit listeners, refer to Implementing Undo and Redo.

A document notifies registered document listeners of changes to the document. Use a document listener to react when text is inserted or removed from a document, or when the style of some of the text changes.

The TextComponentDemo program uses a document listener to update the change log whenever a change is made to the text pane. The following line of code registers an instance of MyDocumentListener as a listener on the text pane's document:

doc.addDocumentListener(new MyDocumentListener());
Here's the implementation of MyDocumentListener:
protected class MyDocumentListener implements DocumentListener {
    public void insertUpdate(DocumentEvent e) {
        displayEditInfo(e);
    }
    public void removeUpdate(DocumentEvent e) {
        displayEditInfo(e);
    }
    public void changedUpdate(DocumentEvent e) {
        displayEditInfo(e);
    }
    private void displayEditInfo(DocumentEvent e) {
            Document document = (Document)e.getDocument();
            int changeLength = e.getLength();
            changeLog.append(e.getType().toString() + ": "
                + changeLength + " character"
                + ((changeLength == 1) ? ". " : "s. ")
                + " Text length = " + document.getLength()
                + "." + newline);
    }
} 
The listener implements three methods for handling three different types of document events: insertion, removal, and style changes. StyledDocuments can fire all three types of events. PlainDocuments fire events only for insertion and removal. For general information about document listeners and document events, see How to Write a Document Listener.

Remember that the document filter for this text pane limits the number of characters allowed in the document. If you try to add more text than the document filter allows, the document filter blocks the change and the listener's insertUpdate method is not called. Document listeners are notified of changes only if the change has already occurred.

Sometimes, you might be tempted to change the document's text from within a document listener. However, you should never modify the contents of a text component from within a document listener. In fact, if you do, your program will likely deadlock! Instead, you can use a formatted text field or provide a document filter.

Listening for Caret and Selection Changes

The TextComponentDemo program uses a caret listener to display the current position of the caret or, if text is selected, the extent of the selection.

The caret listener class in this example is a JLabel subclass. Here's the code that creates the caret listener label and makes it a caret listener of the text pane:

//Create the status area
CaretListenerLabel caretListenerLabel = new CaretListenerLabel(
						"Caret Status");
...
textPane.addCaretListener(caretListenerLabel);
A caret listener must implement one method, caretUpdate, which is called each time the caret moves or the selection changes. Here's the CaretListenerLabel implementation of caretUpdate:
public void caretUpdate(CaretEvent e) {
    //Get the location in the text
    int dot = e.getDot();
    int mark = e.getMark();
    if (dot == mark) {  // no selection
        try {
            Rectangle caretCoords = textPane.modelToView(dot);
            //Convert it to view coordinates
            setText("caret: text position: " + dot +
                    ", view location = [" +
                    caretCoords.x + ", " + caretCoords.y + "]" +
                    newline);
        } catch (BadLocationException ble) {
            setText("caret: text position: " + dot + newline);
        }
     } else if (dot < mark) {
        setText("selection from: " + dot + " to " + mark + newline);
     } else {
        setText("selection from: " + mark + " to " + dot + newline);
     }
}

As you can see, this listener updates its text label to reflect the current state of the caret or selection. The listener gets the information to display from the caret event object. For general information about caret listeners and caret events, see How to Write a Caret Listener.

As with document listeners, a caret listener is passive. It reacts to changes in the caret or in the selection but does not change the caret or the selection. If you want to change the caret or selection, then you should use a navigation filter or a custom caret instead.

Implementing a navigation filter is similar to implementing a document filter. First, you write a subclass of NavigationFilter (in the API reference documentation). You then attach an instance of it to a text component with the setNavigationFilter method.

You might create a custom caret to customize the appearance of a caret. To create a caret, write a class that implements the Caret (in the API reference documentation) interface — perhaps by extending the DefaultCaret (in the API reference documentation) class. Then provide an instance of your class as an argument to setCaret on a text component.

Concepts: About Editor Kits

Under the hood, text components use an EditorKit to tie the various pieces of the text component together. It provides the view factory, document, caret, and actions. An editor kit also reads and writes documents of a particular format. Although all text components use editor kits, some components hide theirs. You can't set or get the editor kit used by a text field or text area. Editor panes and text panes provide the getEditorKit method to get the current editor kit and the setEditorKit method to change it.

For all components, JTextComponent provides API for you to indirectly invoke or customize some editor kit capabilities. For example, JTextComponent provides read and write methods, which invoke the editor kit's read and write methods. JTextComponent also provides a method, getActions, which returns all of the actions supported by a component.

The Swing text package provides these editor kits:

DefaultEditorKit (in the API reference documentation)
Reads and writes plain text, and provides a basic set of editing commands. Details of how the text system treats newlines are in the DefaultEditorKit API documentation. (To summarize: The '\n' character is used internally, but the document or platform line separators are used when writing files.) All the other editor kits are descendants of DefaultEditorKit.
StyledEditorKit (in the API reference documentation)
Reads and writes styled text, and provides a minimal set of actions for styled text. This class is a subclass of DefaultEditorKit and is the editor kit used by JTextPane by default.
HTMLEditorKit (in the API reference documentation)
Reads, writes, and edits HTML. This is a subclass of StyledEditorKit.
RTFEditorKit
Reads, writes, and edits RTF. This is a subclass of StyledEditorKit.
Each of the editor kits above has been registered with the JEditorPane class and associated with the text format that the kit reads, writes, and edits. When a file is loaded into an editor pane, the pane checks the format of the file against its registered kits. If a registered kit is found that supports that file format, the pane uses the kit to read the file, display, and edit it. Thus, the editor pane effectively transforms itself into an editor for that text format. You can extend JEditorPane to support your own text format by creating an editor kit for it, and then using JEditorPane's registerEditorKitForContentType to associate your kit with your text format.

Previous Page Lesson Contents Next Page Start of Tutorial > Start of Trail > Start of Lesson Search
Feedback Form

Copyright 1995-2004 Sun Microsystems, Inc. All rights reserved.