ServiceRepaints throws Throwable on 6630

[petrusmi]

On the Nokia 6630 my run loop dies after a while.
After some investigating I found that serviceRepaints was throwing a Throwable.
It takes about 30 seconds before serviceRepaints freak out and starts throwing Throwables. Once it has started, each call to serviceRepaints throws a new Throwable. And needless to say, no repaints are beeing performed.
Unless I press a key. Then keyPressed calls repaint (not serviceRepaints) and somehow the screen is repainted.
It's just like if serviceRepaints empties the queue and then throws a Throwable without repainting.

The same code works on all other phones I've tried (about 10 different nokias and 10-20 other phones).
Does anyone recognize the problem? Does anyone have a solution?

Cheers,

Petrus

[dcrocha]

Can you post the section of the code you're using for your animation loop?

ServiceRepaints() is a heavy operation, if you're using that for a game, you could use the GameCanvas.flushGraphics() method instead.

Daniel

[petrusmi]

Hi Daniel,

It was a good suggestion to use flushGraphics instead. Unfortunately it worked just as bad.
Now (after about 30 seconds) flushGraphics throws a Throwable instead. And every once in a while getGraphics throws a Throwable.
So in the end I get almost exactly the same behavior as when using serviceRepaints.

I also have some try-catch blocks in there. But besides that, this is pretty much what my code looks like;

public void run() {

do {

//* a short delay
//* update stuff for animation

Graphics g = getGraphics();
paint(g);
flushGraphics();

} while(!paused);

}

The old version with serviceRepaints (which I much rater use since I’ve tested it on many devices) lookes something like this;

public void run() {

do {

//* a short delay
//* update stuff for animation

repaint();
serviceRepaints();

} while(!paused);

}

Thanks,

Petrus

[petrusmi]

Problem 'solved'..
I was accidentally using drawRegion with the TRANS_MIRROR transform.. for Series60 that's a big No-No!
I've seen this problem before, but then it manifested itself as a memory leak. That's why I didn't catch this error sooner.

Cheers,

Petrus