This taken from comp.lang.java.gui:

-----Original Message-----
From: John Fisher <###@###.###OSPAM>
Newsgroups: comp.lang.java.gui
Date: Sunday, December 13, 1998 11:45 AM
Subject: CPU 100%: setbackground call in TableCellRender
>
>The code below demonstrates the following problem on JDK 1.1.7 on Linux,
JDK
>1.1.7B on Win98, and JDK 1.2 on Win98, using the release version of Swing.
>
>The problem is that a table with multiple colors (as implemented below)
>immediately sends the CPU usage to 100%, and slows the entire system. Note,
>if I uncomment the line below, and comment the line following it,
everything
>seems to work fine. The CPU usage only goes 100% when more than one color
is
>displayed at once.
>
>I've submitted this as a bug to JavaSoft, but I was wondering if anyone has
>had similar problems and found a workaround (or can explain to me what I
>might be doing wrong).
>
>Thanks,
>John
>###@###.###OSPAM
>
>==============================================
>
>import java.awt.*;
>import javax.swing.*;
>import javax.swing.table.*;
>
>
>public class JTableTest extends JFrame
>{
>
>
> public static void main(String args[]) {
> new JTableTest();
> }
>
> public JTableTest() {
>
> TableModel dataModel = new AbstractTableModel() {
> public int getColumnCount() { return 10; }
> public int getRowCount() { return 10;}
> public Object getValueAt(int row, int col) { return new
>Integer(row*col); }
> };
>
> JTable mytable = new JTable(dataModel) {
> public TableCellRenderer getCellRenderer(int row, int column) {
> TableCellRenderer renderer = super.getCellRenderer(row,
column);
> if (row > 5) {
> // setBackground(Color.blue);
> setBackground(Color.red);
> } else {
> setBackground(Color.blue);
> }
> return renderer;
> }
> };
>
> JScrollPane scrollpane = new JScrollPane(mytable);
>
> mytable.setToolTipText(null);
> this.getContentPane().add(mytable);
>
> pack();
> show();
> }
>
> }
>

-----Original Message-----
From: John Bridges <###@###.###>
Newsgroups: comp.lang.java.gui
To: John Fisher <###@###.###OSPAM>
Date: Monday, December 14, 1998 3:13 AM
Subject: Re: CPU 100%: setbackground call in TableCellRender


>Hi John,
>
>yup this is completely mad.
>
>The problem is not with the JDK per se but with the Swing1.1
implementation.
>
>(see also my "Is this mad ?" posting to this group).
>
>The problem lies with the fact (as far as I can determine), that a bug was
>posted about Swing that complained setting the background/foreground colors
etc
>did not immediately repaint the control.
>
>JavaSoft in their GROSS stupidity fixed this (which isn't a bug at all in
my
>opinion) by placing a repaint call at the end of the method if the value of
the
>property changes (you can try this with setFont, setBounds, blah, blah
etc) -
>these are in the JComponent overrides.
>
>If you set the Color.blue then the background (in your example) hasn't
changed.
>If you set it to red, then this creates a repaint call ...
>
>Which enter's your method determines the color has changed which causes a
>repaint call ...
>Which enter's your method determines the color has changed ...
>etc
>etc
>
>Daft isn't it. So the long and short of all this is. If you want to change
>properties of a table cell renderer, it mustn't be a JComponent derivative
(I
>have'nt checked 1.2 they have been really dumb and moved this fix into the
>Component class ... ).
>
>This is definitely a bug.
>
>More than a bug, its gross stupidity.
>
>Regards
>
>John