Inconsistent ArrayIndexOutOfBoundsException using SwingWorker worker thread

Question

On occasions I will get the following exception thrown whenever I happen to be using SimpleBrowser.this.processor.processTask() method to run a SwingWorker<Void, Void>-based class Task worker thread (within doInBackground()) :
Exception in thread "AWT-EventQueue-0" java.lang.ArrayIndexOutOfBoundsException: -1
        at javax.swing.text.BoxView.getOffset(BoxView.java:1076)
        at javax.swing.text.BoxView.childAllocation(BoxView.java:670)
        at javax.swing.text.CompositeView.getChildAllocation(CompositeView.java:215)
        at javax.swing.text.BoxView.getChildAllocation(BoxView.java:428)
        at javax.swing.plaf.basic.BasicTextUI$UpdateHandler.calculateViewPosition(BasicTextUI.java:1978)
        at javax.swing.plaf.basic.BasicTextUI$UpdateHandler.layoutContainer(BasicTextUI.java:1954)
        at java.awt.Container.layout(Container.java:1432)
        at java.awt.Container.doLayout(Container.java:1421)
        at java.awt.Container.validateTree(Container.java:1519)
        at java.awt.Container.validateTree(Container.java:1526)
        at java.awt.Container.validateTree(Container.java:1526)
        at java.awt.Container.validate(Container.java:1491)
        at javax.swing.RepaintManager.validateInvalidComponents(RepaintManager.java:639)
        at javax.swing.SystemEventQueueUtilities$ComponentWorkRequest.run(SystemEventQueueUtilities.java:127)
        at java.awt.event.InvocationEvent.dispatch(InvocationEvent.java:209)
        at java.awt.EventQueue.dispatchEvent(EventQueue.java:597)
        at java.awt.EventDispatchThread.pumpOneEventForFilters(EventDispatchThread.java:273)
        at java.awt.EventDispatchThread.pumpEventsForFilter(EventDispatchThread.java:183)
        at java.awt.EventDispatchThread.pumpEventsForHierarchy(EventDispatchThread.java:173)
        at java.awt.EventDispatchThread.pumpEvents(EventDispatchThread.java:168)
        at java.awt.EventDispatchThread.pumpEvents(EventDispatchThread.java:160)
        at java.awt.EventDispatchThread.run(EventDispatchThread.java:121)I haven't found much reliable online information to illustrate this issue any further so I'm a bit in the dark. Following is my code that runs within the aforementioned worker thread which I believe throws the exception:
         * Use {@link #setWebBrowserURL} using a local {@link com.ppowell.tools.ObjectTools.SwingTools.Task}
        protected void processTask() {
            Task task = new Task() {
                public Void doInBackground() {
                    int progress = 0;
                    while (!SimpleBrowser.this.builder.hasLoadedWebpage && progress < 100) {
                        SimpleBrowser.this.statusBar.setMessage("Attempting to load " + SimpleBrowser.this.getURL().toString());
                        this.setProgress(progress);
                        progress++;
                    SimpleBrowser.this.setWebBrowserURL();
                    try {
                        Thread.sleep(2500);
                    } catch (InterruptedException ignore) {} // DO NOTHING
                    if (SimpleBrowser.this.builder.hasLoadedWebpage) {
                        SimpleBrowser.this.statusBar.setMessage("Done");
                    return null;
            task.addPropertyChangeListener(SimpleBrowser.this);
            task.execute();
        }Is there a way I might at least be able to suppress this error (the GUI application browser functions just fine in spite of it), or, even better, solve this inconsistent problem?
Thanks
Phil

Answer 1

I suspect that you need to "clean" your html priorto
calling super.setText(). My guess is that you're
yanking nodes from beaneath Swing while it'strying
to render.
In addition to that, you may want to use JTidy to
clean up your html, and convert it to xhtml --again
prior to calling super.setText().Problem is that however I clean up the HTML, the
moment I try to reinsert back into the JEditorPane
using setText(), I get EmptyStackException or I'll
get NullPointerException or it just might work - same
problem, different exception(s).Ok this time I am simply using setText() instead of setPage(), but the results are, while consistent, they are consistently completely wrong. The browser appears blank every time, no HTML can be retrieved (you get a NullPointerException if you try), all the while I can verify that that original HTML from the remote site is correct, it never, ever, inserts into JEditorPane.
Here is my code:
* SimpleHTMLRenderableEditorPane.java
* Created on March 13, 2007, 3:39 PM
* To change this template, choose Tools | Template Manager
* and open the template in the editor.
package com.ppowell.tools.ObjectTools.SwingTools;
import java.io.*;
import java.net.*;
import javax.swing.JEditorPane;
import javax.swing.text.html.HTMLEditorKit;
* A safer version of {@link javax.swing.JEditorPane}
* @author Phil Powell
* @version JDK 1.6.0
public class SimpleHTMLRenderableEditorPane extends JEditorPane {
    //--------------------------- --* CONSTRUCTORS *--
    // <editor-fold defaultstate="collapsed" desc=" Constructors ">
    /** Creates a new instance of SimpleHTMLRenderableEditorPane */
    public SimpleHTMLRenderableEditorPane() {
        super();
     * Creates a new instance of SimpleHTMLRenderableEditorPane
     * @param url {@link java.lang.String}
     * @throws java.io.IOException Thrown if an I/O exception occurs
    public SimpleHTMLRenderableEditorPane(String url) throws
IOException {
        super(url);
     * Creates a new instance of SimpleHTMLRenderableEditorPane
     * @param type {@link java.lang.String}
     * @param text {@link java.lang.String}
    public SimpleHTMLRenderableEditorPane(String type, String text) {
        super(type, text);
     * Creates a new instance of SimpleHTMLRenderableEditorPane
     * @param url {@link java.net.URL}
     * @throws java.io.IOException Thrown if an I/O exception occurs
    public SimpleHTMLRenderableEditorPane(URL url) throws IOException
        super(url);
    // </editor-fold>
    //----------------------- --* GETTER/SETTER METHODS *--
    // <editor-fold defaultstate="collapsed" desc=" Getter/Setter
Methods ">
     * Retrieve HTML content
     * @return html {@link java.lang.String}
    public String getText() {
        try {
             * I decided to use {@link java.net.HttpURLConnection} to
retrieve the
             * HTML code from the remote site instead of using
super.getText() because
             * of the HTML code return constantly being stripped to
primitive HTML
             * template formatting irregardless of the original HTML
source code
            HttpURLConnection conn =
(HttpURLConnection)getPage().openConnection();
            conn.setUseCaches(false);
            conn.setDefaultUseCaches(false);
            conn.setDoOutput(false); // READ-ONLY
            BufferedReader in = new BufferedReader(
                    new InputStreamReader(
                    conn.getInputStream()));
            int data;
            StringBuffer sb = new StringBuffer();
            char[] ch = new char[512];
            while ((data = in.read(ch)) != -1) {
                sb.append(ch, 0, data);
            in.close();
            conn.disconnect();
            return sb.toString();
        } catch (IOException e) {
            return super.getText(); // DEFAULT TO USING
super.getText() IF NO I/O CONNECTION
     * Overloaded to fix HTML rendering bug Bug ID: 4695909.
     * @param text {@link java.lang.String}
    public void setText(String text) {
        // Workaround for bug Bug ID: 4695909 in java 1.4
        // JEditorPane does not handle the META tag in the html HEAD
        if (isJava14() && "text/
html".equalsIgnoreCase(getContentType())) {
            text = stripMetaTag(text);
        super.setText(text);
    // </editor-fold>
    //--------------------------- --* OTHER METHODS *--
    // <editor-fold defaultstate="collapsed" desc=" Methods ">
     * Clean HTML to remove things like <link>, <script>,
     * <style>, <object>, <embed>, and <!-- -->
     * Based upon <a href="http://bugs.sun.com/bugdatabase/view_bug.do?
bug_id=4695909">bug report</a>
    public void cleanHTML() {
        try {
            setText(cleanHTML(getText()));
        } catch (Exception e) {} // DO NOTHING
     * Clean HTML
     * @param html {@link java.lang.String}
     * @return html {@link java.lang.String}
    public String cleanHTML(String html) {
        String[] tagArray = {"<LINK", "<SCRIPT", "<STYLE", "<OBJECT",
"<EMBED", "<!--"};
        String upperHTML = html.toUpperCase();
        String endTag;
        int index = -1, endIndex = -1;
        for (int i = 0; i < tagArray.length; i++) {
            index = upperHTML.indexOf(tagArray);
endTag = "</" + tagArray[i].substring(1,
tagArray[i].length());
endIndex = upperHTML.indexOf(endTag, index);
while (index >= 0) {
if (endIndex >= 0) {
html = html.substring(0, index) +
html.substring(html.indexOf(">", endIndex)
+ 1,
html.length());
upperHTML = upperHTML.substring(0, index) +
upperHTML.substring(upperHTML.indexOf(">",
endIndex) + 1,
upperHTML.length());
} else {
html = html.substring(0, index) +
html.substring(html.indexOf(">", index) +
1,
html.length());
upperHTML = upperHTML.substring(0, index) +
upperHTML.substring(upperHTML.indexOf(">",
index) + 1,
upperHTML.length());
index = upperHTML.indexOf(tagArray[i]);
endIndex = upperHTML.indexOf(endTag, index);
// REF: http://forum.java.sun.com/thread.jspa?threadID=213582&messageID=735120
html = html.substring(0, upperHTML.indexOf(">",
upperHTML.indexOf("</HTML")) + 1);
// REF: http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=5042872
return html.trim();
* This actually only obtains the URL; this serves as a retriever
for cleanHTML(String html)
* @param url {@link java.net.URL}
* @return html {@link java.lang.String}
public String cleanHTML(URL url) {
try {
HttpURLConnection conn =
(HttpURLConnection)url.openConnection();
conn.setUseCaches(false);
conn.setDefaultUseCaches(false);
conn.setDoOutput(false); // READ-ONLY
BufferedReader in = new BufferedReader(
new InputStreamReader(
conn.getInputStream()));
int data;
StringBuffer sb = new StringBuffer();
char[] ch = new char[512];
while ((data = in.read(ch)) != -1) {
sb.append(ch, 0, data);
in.close();
conn.disconnect();
return cleanHTML(sb.toString());
} catch (IOException e) {
e.printStackTrace();
return null;
* Determine if java version is 1.4.
* @return true if java version is 1.4.x....
private boolean isJava14() {
if (System.getProperty("java.version") == null) return false;
return System.getProperty("java.version").startsWith("1.4");
* Workaround for Bug ID: 4695909 in java 1.4, fixed in 1.5
* JEditorPane fails to display HTML BODY when META tag included
in HEAD section.
* Code modified by Phil Powell
* <html>
* <head>
* <META http-equiv="Content-Type" content="text/html;
charset=UTF-8">
* </head>
* <body>
* @param text html to strip.
* @return same HTML text w/o the META tag.
private String stripMetaTag(String text) {
// String used for searching, comparison and indexing
String textUpperCase = text.toUpperCase();
int indexHead = textUpperCase.indexOf("<HEAD ");
int indexMeta = textUpperCase.indexOf("<META ");
int indexBody = textUpperCase.indexOf("<BODY ");
// Not found or meta not inside the head nothing to strip...
if (indexMeta == -1 || indexMeta < indexHead || indexMeta >
indexBody) {
return text;
// Find end of meta tag text.
int indexHeadEnd = textUpperCase.indexOf(">", indexMeta);
// Strip meta tag text
return text.substring(0, indexMeta - 1) +
text.substring(indexHeadEnd + 1);
// </editor-fold>
Instead if you try
browser.getText()
You will get a NullPointerException
If you try
    public void setText(String text) {
        // Workaround for bug Bug ID: 4695909 in java 1.4
        // JEditorPane does not handle the META tag in the html HEAD
        if (isJava14() && "text/
html".equalsIgnoreCase(getContentType())) {
            text = stripMetaTag(text);
        System.out.println(text); // YOU WILL SEE CNN'S HTML
        super.setText(text);
        System.out.println(super.getText()); // SEE BELOW
    }You see only this:
<html>
  <head>
  </head>
  <body>
    <p style="margin-top: 0">
    </p>
  </body>
</html>