Interesting performance tip...

[crystalsquid]

This code takes about 8 seconds (Java 1.4):

 

1   2   3   4   5

canvas = new Canvas();   canvas.setBackground(new Color(0, 0, 0));   canvas.setBounds(0, 0, 640, 480);   applet.add("Center", canvas);  // <- BIG STALL HERE

This code takes <1 sec:

 

1   2   3   4   5

canvas = new Canvas();   applet.add("Center", canvas);  // <- No stall here  canvas.setBackground(new Color(0, 0, 0));   canvas.setBounds(0, 0, 640, 480);

Bizarre...
Moral of the story: Add the canvas straight after you build it.

- Dom

modified - I forgot to change the 2nd code snippet - Oops!  

[blahblahblahh]

This code takes about 8 seconds (Java 1.4):

 

1   2   3   4   5

canvas = new Canvas();   canvas.setBackground(new Color(0, 0, 0));   canvas.setBounds(0, 0, 640, 480);   applet.add("Center", canvas);  // <- BIG STALL HERE

This code takes <1 sec:

 

1   2   3   4   5

canvas = new Canvas();  // <- No stall here  canvas.setBackground(new Color(0, 0, 0));   canvas.setBounds(0, 0, 640, 480);   applet.add("Center", canvas);

Bizarre...

Yeah; what's the difference, exactly?

Anyway, I have an outstanding bug on loading and starting of applets in 1.4.x where sometimes applets can crash on startup - and never get to display anything - under linux. I narrowed it down to a race condition somewhere in Sun's code. It's been accepted but not yet fixed. Who knows how much more screwed code there is in applet startup ?

[Jeff]

Yeah; what's the difference, exactly?

Hmm I dunno but you AIR take your  life  into your hands modifying live AWt components as it is not thread-safe...