Lab assignment 6 - Netscrape

Three advanced swing components should be used: JEditorPane, JScrollPane and JTable. Also, the swing class HTML that parses web pages will play a major role.

These are the imports you'll probably use.

   import java.io.*;
   import java.awt.*;
   import java.awt.event.*;
   import java.net.*;
   import javax.swing.*;
   import javax.swing.event.*;
   import javax.swing.text.html.*;
   import javax.swing.text.*;

The graphical interface

When writing a graphical program that is not trivially small, it's often a good idea to start with a prototype, i.e. a program that looks like the complete program but lacks most of the functionality. In this case the prototype should show a JFrame with three components: a text field at the top (position NORTH), a JEditorPane in the center and a JTable to the right (EAST), all without contents. The table should have 50 lines and two columns, so it will not fit in the window. Put the table inside a scrolling box, for instance with the following code:

   JScrollPane links=new JScrollPane(new JTable(50,2));

Run the program and check that it's possible to write text in all three components. This is an accidental feature (or bug?) as far as the web page and link table components are concerned. TextFields are frequently used for input so they are expected to be editable. The user enters the web address to go to in the TextField. When the user presses Enter, the ActionListener for the text field is triggered. It's possible to let the Netscrape class itself be an ActionListener and make a call like browser.showPage(...) when the ActionEvent takes place.

The web browser

Your task is now to write a class WebBrowser which extends JEditorPane and has a method showPage(address). This method actually only needs the single statement setPage(new URL(address)); but the compiler requires this to be wrapped in a try-catch to handle the cases when the page cannot be found etc.
When this happens, the user should be alerted of the error by a dialog box. This component is called JOptionPane. Look up its usage in the Java documentation (hint: the first parameter should be this).

When the class is ready, Netscrape can do WebBrowser browser = new WebBrowser();.

The surfer - extra assignment

When you try your web browser out you will be annoyed by the fact that the browser does nothing when you click the links. This is not necessary to fix to pass the lab assignment, but counts as an extra assignment. Two things must be done: you have to call setEditable(false) so that your browser does not interpret the click as a "start editing"-action. Also, you need to add a HyperlinkListener. The details of this areshown in the web summary of lecture 10.
The call getURL().toString() for a HyperlinkEvent returns a web address which can be used in the address bar.

So, write a class Surfer which inherits from WebBrowser and takes (and saves) a reference to the address bar as a parameter in its constructor. Just writing the address to the address bar is not enough, however. You also need to generate an event by callin postActionEvent() for the address bar. When the class is ready, Netscrape can do WebBrowser browser = new Surfer(addressBar);

The link table

Leave the Netscrape class for a while and write a new program that outputs all links from a web page. It can be called WebCheck and consist of, at least for now, just a main-method. A web page can be printed with the following code

   String webbsida="http://www.nada.kth.se/~henrik";
   InputStream in=new URL(webbsida).openConnection().getInputStream();   
   InputStreamReader reader= new InputStreamReader(in);
   while(reader.ready()) System.out.print((char)reader.read());

   doc.putProperty("IgnoreCharsetDirective", Boolean.TRUE);
   new HTMLEditorKit().read(reader,doc,0);

An HTML page consists of different kinds of tags. The tag used for links is called A and has an attribute HREF which has the web address of the link as value. Your task is to iterate through the document from A-tag to A-tag, read the HREF attribute and print its value. To do this, create a Tag.A iterator and a HREF attribute object. Use a while loop to iterate through the A-tags. For each A-tag, print the HREF-attribute.

When this works, change the program to also output the text between the A-tag and its closing tag by using it.getStartOffset() and it.getEndOffset().

Instead of printing links and link text, WebCheck should now add them in a JTable, which can be retrieved by Netscrape with the call

   links=new JScrollPane(WebbCheck.linkTable(addressBar.getText()));

While you are testing the program, it might be a good idea to have a main method in WebCheck that calls linkTable, puts the table in a JFrame and shows it. Once this works, you can be sure that Netscrape too will show the table properly.

The last problem to solve is how to get rid of the old table before adding a new one when a new URL is requested. One way is to first call remove(links), and then create a new JScrollPane links before adding it in the right part of the browser as before.

Finally, you can go web smurfing among the links to celebrate that the last prutt assignment is completed.