I've looked at similar questions but they all are concerned with collision detection rather than preventing overlap. I've gotten most of it to work with the below code:
private final EventHandler<MouseEvent> onPress = mouseEvent -> {
xDrag = this.getCenterX() - mouseEvent.getX();
yDrag = this.getCenterY() - mouseEvent.getY();
};
private final EventHandler<MouseEvent> onDrag = mouseEvent -> {
for (Shape shape : getAllShapes()) {
if (!this.equals(shape)) {
Shape intersect = Shape.intersect(shape, this);
if (intersect.getLayoutBounds().getWidth() > 0) {
return;
}
}
}
this.setCenterX(mouseEvent.getX() + xDrag);
this.setCenterY(mouseEvent.getY() + yDrag);
};
However, the problem is, once there is a tiniest bit of overlap, the Shape is no longer draggable at all. Meaning, if I drag a shape to another, once they become essentially tangent, neither of them are draggable anymore. What I want to happen is just that, for example, if you try to drag a circle onto another, the circle won't follow the mouse position as long as the future position of the drag will cause an overlap.
I can't figure out exactly how to accomplish this.
EDIT: Minimum Reproducible Example:
Main.java
public class Main extends Application {
static Circle circle1 = new DraggableCircle(100, 200);
static Circle circle2 = new DraggableCircle(200, 300);
static Circle[] circleList = new Circle[]{circle1, circle2};
#Override
public void start(Stage primaryStage) {
primaryStage.setTitle("Hello World");
Pane pane = new Pane();
primaryStage.setScene(new Scene(pane, 300, 275));
primaryStage.show();
pane.getChildren().addAll(circle1, circle2);
}
public static void main(String[] args) {
launch(args);
}
}
DraggableCircle.java
public class DraggableCircle extends Circle {
private double xDrag, yDrag;
public DraggableCircle(double x, double y) {
super(x, y, 30);
this.setFill(Color.WHITE);
this.setStroke(Color.BLACK);
this.setStrokeWidth(1.5);
this.addEventHandler(MouseEvent.MOUSE_PRESSED, onPress);
this.addEventHandler(MouseEvent.MOUSE_DRAGGED, onDrag);
}
private final EventHandler<MouseEvent> onPress = (mouseEvent) -> {
xDrag = this.getCenterX() - mouseEvent.getX();
yDrag = this.getCenterY() - mouseEvent.getY();
};
private final EventHandler<MouseEvent> onDrag = mouseEvent -> {
for (Shape shape : Main.circleList) {
if (!this.equals(shape)) {
Shape intersect = Shape.intersect(shape, this);
if (intersect.getLayoutBounds().getWidth() > 0) {
return;
}
}
}
this.setCenterX(mouseEvent.getX() + xDrag);
this.setCenterY(mouseEvent.getY() + yDrag);
};
}
This also has an issue where dragging too quickly causes a noticeable overlap between the circles, before the drag detection ends.
A simple (imperfect) solution
The following algorithm will allow a node to continue to be dragged after an intersection has occurred:
Record the current draggable shape position.
Set the new position.
Check the intersection.
If an intersection is detected, reset the position to the original position.
An implementation replaces the drag handler in the supplied minimal example code from the question.
private final EventHandler<MouseEvent> onDrag = (mouseEvent) -> {
double priorCenterX = getCenterX();
double priorCenterY = getCenterY();
this.setCenterX(mouseEvent.getX() + xDrag);
this.setCenterY(mouseEvent.getY() + yDrag);
for (Shape shape : Main.circleList) {
if (!this.equals(shape)) {
Shape intersect = Shape.intersect(shape, this);
if (intersect.getLayoutBounds().getWidth() > 0) {
this.setCenterX(priorCenterX);
this.setCenterY(priorCenterY);
return;
}
}
}
};
This handler does work better than what you had, it does at least allow you to continue dragging after the intersection.
But, yes, if you drag quickly it will leave a visible space between nodes when it has detected that the drag operation would cause an intersection, which isn't ideal.
Also, the additional requirement you added in your comment about having the dragged shape glide along a border would require a more sophisticated solution.
Other potential solutions
I don't offer code for these more sophisticated solutions here.
One potential brute force solution is to interpolate the prior center with the new center and then, in a loop, slowly move the dragged object along the interpolated line until an intersection is detected, then just back it out to the last interpolated value to prevent the intersection. You can do this by calculating and applying a normalized (1 unit distance) movement vector. That might fix space between intersected nodes.
Similarly to get the gliding, on the intersection, you could just update either the interpolated x or y value rather than both.
There may be more sophisticated methods with geometry math applied, especially if you know shape geometry along with movement vectors and surface normals.
+1 for #jewelsea answer.
On top of #jewelsea answer, I would like to provide a fix for the "space between nodes" issue.
So you might have already observed that when you drag fast, it will not cover each and every pixel in the drag path. It varies with the speed of the drag. So when you decide to move it to the previous recorded point, we will do a quick math, to see if there is any gap between the two nodes, if yes:
We will do a math "to determine a point along a line which is at distance d" and move the drag circle to that point. Here..
start point of line is : previous recorded point
end point of line is : the intersected shape center
d is : the gap between the two shapes.
So the updated code to the #jewelsea answer is as below:
private final EventHandler<MouseEvent> onDrag = (mouseEvent) -> {
double priorCenterX = getCenterX();
double priorCenterY = getCenterY();
this.setCenterX(mouseEvent.getX() + xDrag);
this.setCenterY(mouseEvent.getY() + yDrag);
for (Circle shape : Main.circleList) {
if (!this.equals(shape)) {
Shape intersect = Shape.intersect(shape, this);
if (intersect.getLayoutBounds().getWidth() > 0) {
Point2D cx = new Point2D(priorCenterX, priorCenterY);
Point2D px = new Point2D(shape.getCenterX(), shape.getCenterY());
double d = cx.distance(px);
if (d > getRadius() + shape.getRadius()) {
cx = pointAtDistance(cx, px, d - getRadius() - shape.getRadius());
}
this.setCenterX(cx.getX());
this.setCenterY(cx.getY());
return;
}
}
}
};
private Point2D pointAtDistance(Point2D p1, Point2D p2, double distance) {
double lineLength = p1.distance(p2);
double t = distance / lineLength;
double dx = ((1 - t) * p1.getX()) + (t * p2.getX());
double dy = ((1 - t) * p1.getY()) + (t * p2.getY());
return new Point2D(dx, dy);
}
Related
I have a sample 3D application (built by taking reference from the Javafx sample 3DViewer) which has a table created by laying out Boxes and Panes:
The table is centered wrt (0,0,0) coordinates and camera is at -z position initially.
It has the zoom-in/out based on the camera z position from the object.
On zooming in/out the object's boundsInParent increases/decreases i.e. area of the face increases/decreases. So the idea is to put more text when we have more area (always confining within the face) and lesser text or no text when the face area is too less. I am able to to do that using this node hierarchy:
and resizing the Pane (and managing the vBox and number of texts in it) as per Box on each zoom-in/out.
Now the issue is that table boundsInParent is giving incorrect results (table image showing the boundingBox off at the top) whenever a text is added to the vBox for the first time only. On further zooming-in/out gives correct boundingBox and does not go off.
Below is the UIpane3D class:
public class UIPane3D extends Pane
{
VBox textPane;
ArrayList<String> infoTextKeys = new ArrayList<>();
ArrayList<Text> infoTextValues = new ArrayList<>();
Rectangle bgCanvasRect = null;
final double fontSize = 16.0;
public UIPane3D() {
setMouseTransparent(true);
textPane = new VBox(2.0)
}
public void updateContent() {
textPane.getChildren().clear();
getChildren().clear();
for (Text textNode : infoTextValues) {
textPane.getChildren().add(textNode);
textPane.autosize();
if (textPane.getHeight() > getHeight()) {
textPane.getChildren().remove(textNode);
textPane.autosize();
break;
}
}
textPane.setTranslateY(getHeight() / 2 - textPane.getHeight() / 2.0);
bgCanvasRect = new Rectangle(getWidth(), getHeight());
bgCanvasRect.setFill(Color.web(Color.BURLYWOOD.toString(), 0.10));
bgCanvasRect.setVisible(true);
getChildren().addAll(bgCanvasRect, textPane);
}
public void resetInfoTextMap()
{
if (infoTextKeys != null || infoTextValues != null)
{
try
{
infoTextKeys.clear();
infoTextValues.clear();
} catch (Exception e){e.printStackTrace();}
}
}
public void updateInfoTextMap(String pKey, String pValue)
{
int index = -1;
boolean objectFound = false;
for (String string : infoTextKeys)
{
index++;
if(string.equals(pKey))
{
objectFound = true;
break;
}
}
if(objectFound)
{
infoTextValues.get(index).setText(pValue.toUpperCase());
}
else
{
if (pValue != null)
{
Text textNode = new Text(pValue.toUpperCase());
textNode.setFont(Font.font("Consolas", FontWeight.BLACK, FontPosture.REGULAR, fontSize));
textNode.wrappingWidthProperty().bind(widthProperty());
textNode.setTextAlignment(TextAlignment.CENTER);
infoTextKeys.add(pKey);
infoTextValues.add(textNode);
}
}
}
}
The code which get called at the last after all the manipulations:
public void refreshBoundingBox()
{
if(boundingBox != null)
{
root3D.getChildren().remove(boundingBox);
}
PhongMaterial blueMaterial = new PhongMaterial();
blueMaterial.setDiffuseColor(Color.web(Color.CRIMSON.toString(), 0.25));
Bounds tableBounds = table.getBoundsInParent();
boundingBox = new Box(tableBounds.getWidth(), tableBounds.getHeight(), tableBounds.getDepth());
boundingBox.setMaterial(blueMaterial);
boundingBox.setTranslateX(tableBounds.getMinX() + tableBounds.getWidth()/2.0);
boundingBox.setTranslateY(tableBounds.getMinY() + tableBounds.getHeight()/2.0);
boundingBox.setTranslateZ(tableBounds.getMinZ() + tableBounds.getDepth()/2.0);
boundingBox.setMouseTransparent(true);
root3D.getChildren().add(boundingBox);
}
Two things:
The table3D's boundsInParent is not updated properly when texts are added for the first time.
What would be the right way of putting texts on 3D nodes? I am having to manipulate a whole lot to bring the texts as required.
Sharing code here.
For the first question, about the "jump" that can be noticed just when after scrolling a new text item is laid out:
After digging into the code, I noticed that the UIPane3D has a VBox textPane that contains the different Text nodes. Every time updateContent is called, it tries to add a text node, but it checks that the vbox's height is always lower than the pane's height, or else the node will be removed:
for (Text textNode : infoTextValues) {
textPane.getChildren().add(textNode);
textPane.autosize();
if (textPane.getHeight() > getHeight()) {
textPane.getChildren().remove(textNode);
textPane.autosize();
break;
}
}
While this is basically correct, when you add a node to the scene, you can't get textPane.getHeight() immediately, as it hasn't been laid out yet, and you have to wait until the next pulse. This is why the next time you scroll, the height is correct and the bounding box is well placed.
One way to force the layout and get the correct height of the textNode is by forcing css and a layout pass:
for (Text textNode : infoTextValues) {
textPane.getChildren().add(textNode);
// force css and layout
textPane.applyCss();
textPane.layout();
textPane.autosize();
if (textPane.getHeight() > getHeight()) {
textPane.getChildren().remove(textNode);
textPane.autosize();
break;
}
}
Note that:
This method [applyCss] does not normally need to be invoked directly but may be used in conjunction with Parent.layout() to size a Node before the next pulse, or if the Scene is not in a Stage.
For the second question, about a different solution to add Text to 3D Shape.
Indeed, placing a (2D) text on top of a 3D shape is quite difficult, and requires complex maths (that are done quite nicely in the project, by the way).
There is an alternative avoiding the use of 2D nodes directly.
Precisely in a previous question, I "wrote" into an image, that later on I used as the material diffuse map of a 3D shape.
The built-in 3D Box places the same image into every face, so that wouldn't work. We can implement a 3D prism, or we can make use of the CuboidMesh node from the FXyz3D library.
Replacing the Box in UIPaneBoxGroup:
final CuboidMesh contentShape;
UIPane3D displaypane = null;
PhongMaterial shader = new PhongMaterial();
final Color pColor;
public UIPaneBoxGroup(final double pWidth, final double pHeight, final double pDepth, final Color pColor) {
contentShape = new CuboidMesh(pWidth, pHeight, pDepth);
this.pColor = pColor;
contentShape.setMaterial(shader);
getChildren().add(contentShape);
addInfoUIPane();
}
and adding the generateNet method:
private Image generateNet(String string) {
GridPane grid = new GridPane();
grid.setAlignment(Pos.CENTER);
Label label5 = new Label(string);
label5.setFont(Font.font("Consolas", FontWeight.BLACK, FontPosture.REGULAR, 40));
GridPane.setHalignment(label5, HPos.CENTER);
grid.add(label5, 3, 1);
double w = contentShape.getWidth() * 10; // more resolution
double h = contentShape.getHeight() * 10;
double d = contentShape.getDepth() * 10;
final double W = 2 * d + 2 * w;
final double H = 2 * d + h;
ColumnConstraints col1 = new ColumnConstraints();
col1.setPercentWidth(d * 100 / W);
ColumnConstraints col2 = new ColumnConstraints();
col2.setPercentWidth(w * 100 / W);
ColumnConstraints col3 = new ColumnConstraints();
col3.setPercentWidth(d * 100 / W);
ColumnConstraints col4 = new ColumnConstraints();
col4.setPercentWidth(w * 100 / W);
grid.getColumnConstraints().addAll(col1, col2, col3, col4);
RowConstraints row1 = new RowConstraints();
row1.setPercentHeight(d * 100 / H);
RowConstraints row2 = new RowConstraints();
row2.setPercentHeight(h * 100 / H);
RowConstraints row3 = new RowConstraints();
row3.setPercentHeight(d * 100 / H);
grid.getRowConstraints().addAll(row1, row2, row3);
grid.setPrefSize(W, H);
grid.setBackground(new Background(new BackgroundFill(pColor, CornerRadii.EMPTY, Insets.EMPTY)));
new Scene(grid);
return grid.snapshot(null, null);
}
Now all the 2D related code can be removed (including displaypane), and after a scrolling event get the image:
public void refreshBomUIPane() {
Image net = generateNet(displaypane.getText());
shader.setDiffuseMap(net);
}
where in UIPane3D:
public String getText() {
return infoTextKeys.stream().collect(Collectors.joining("\n"));
}
I've also removed the bounding box to get this picture:
I haven't played around with the number of text nodes that can be added to the VBox, the font size nor with an strategy to avoid generating images on every scroll: only when the text changes this should be done. So with the current approach is quite slow, but it can be improved notably as there are only three possible images for each box.
I am developing a program for video annotation. Objects seen in a video may be marked as objects of interest and if they are interacting, the user may draw a line between two annotations. On the visible level, bject annotations are basically transparent Rectangles and relations between them are Lines. At this point it is easy to compute the center of a rectangle, but I am not able to bind the start and end of a Line to the center of the corresponding Rectangle. I have tried the following approaches:
Creating two DoubleBindings inside the rectangle class that computes the center x and y:
private DoubleBinding centerXBinding = new DoubleBinding() {
#Override
protected double computeValue() {
return getX() + getWidth() / 2;
}
};
and then bound it to the newly created Line:
`currentRelation.startXProperty().bind(startShape.centerXBinding());`
in the controller…
The result is ok at first, the line start and end points are exactly where I want to haven them, but when a Rectangle gets dragged to another position, the line end does not move anywhere!
Does anyone see the problem?
UPDATE:
The movement of a Rectangle is done by computing an offset and updating the translation values like translateX:
public class MyRectangle extends Rectangle {
private double orgSceneX;
private double orgSceneY;
private double orgTranslateX;
private double orgTranslateY;
private void initEventHandling() {
this.setOnMousePressed(mousePress -> {
if (mousePress.getButton() == MouseButton.PRIMARY) {
orgSceneX = mousePress.getSceneX();
orgSceneY = mousePress.getSceneY();
orgTranslateX = ((MyRectangle) mousePress.getSource()).getTranslateX();
orgTranslateY = ((MyRectangle) mousePress.getSource()).getTranslateY();
mousePress.consume();
} else if (mousePress.getButton() == MouseButton.SECONDARY) {
System.out.println(LOG_TAG + ": right mouse button PRESS on " + this.getId() + ", event not consumed");
}
});
this.setOnMouseDragged(mouseDrag -> {
if (mouseDrag.getButton() == MouseButton.PRIMARY) {
double offsetX = mouseDrag.getSceneX() - orgSceneX;
double offsetY = mouseDrag.getSceneY() - orgSceneY;
double updateTranslateX = orgTranslateX + offsetX;
double updateTranslateY = orgTranslateY + offsetY;
this.setTranslateX(updateTranslateX);
this.setTranslateY(updateTranslateY);
mouseDrag.consume();
}
});
}
}
Your binding needs to invalidate when either the xProperty or widthProperty are invalidated (so that anything bound to it knows to recompute). You can do this by calling the bind method in the custom binding's constructor:
private DoubleBinding centerXBinding = new DoubleBinding() {
{
bind(xProperty(), widthProperty());
}
#Override
protected double computeValue() {
return getX() + getWidth() / 2;
}
};
Note you can also do
private DoubleBinding centerXBinding = xProperty().add(widthProperty().divide(2));
The choice between the two is really just a matter of which style you prefer.
Binding to the x and width properties assumes, obviously, that you are moving the rectangle by changing one or both of those properties. If you are moving the rectangle by some other means (e.g. by changing one of its translateX or translateY properties, or by altering its list of transformations), then you need to observe the boundsInParentProperty instead:
private DoubleBinding centerXBinding = new DoubleBinding() {
{
bind(boundsInParentProperty());
}
#Override
protected double computeValue() {
Bounds bounds = getBoundsInParent();
return (bounds.getMinX() + bounds.getMaxX()) / 2 ;
}
}
This binding will give the x-coordinate of the center of the rectangle in the parent's coordinate system (which is usually the coordinate system you want).
As mentioned in the title, i have two Circle 's the first is draggable and the second is fixed, I would rotate (with the drag) the first one around the second without overlapping them but my Circle reacts oddly, I'm sure the error comes from the drag condition but I don't know how to solve it, that's why I need your help, here is a minimal and testable code :
import javafx.application.Application;
import javafx.scene.Scene;
import javafx.scene.layout.Pane;
import javafx.scene.paint.Color;
import javafx.scene.shape.Circle;
import javafx.stage.Stage;
public class Collision extends Application{
private Pane root = new Pane();
private Scene scene;
private Circle CA = new Circle(20);
private Circle CB = new Circle(20);
private double xOffset = 0;
private double yOffset = 0;
#Override
public void start(Stage stage) throws Exception{
initCircles();
scene = new Scene(root,500,500);
stage.setScene(scene);
stage.show();
}
private void initCircles(){
CA.setCenterX(100);
CA.setCenterY(100);
CA.setFill(Color.rgb(255, 0, 0,0.2));
CA.setStroke(Color.BLACK);
CB.setCenterX(250);
CB.setCenterY(200);
CB.setFill(Color.rgb(255, 0, 0,0.2));
CB.setStroke(Color.BLACK);
CA.setOnMousePressed(evt->{
xOffset = CA.getCenterX() - evt.getSceneX();
yOffset = CA.getCenterY() - evt.getSceneY();
});
CA.setOnMouseDragged(evt->{
//get Scene coordinate from MouseEvent
drag(evt.getSceneX(),evt.getSceneY());
});
root.getChildren().addAll(CA,CB);
}
private void drag(double x, double y){
/* calculate the distance between
* the center of the first and the second circle
*/
double distance = Math.sqrt (Math.pow(CA.getCenterX() - CB.getCenterX(),2) + Math.pow(CA.getCenterY() - CB.getCenterY(),2));
if (!(distance < (CA.getRadius() + CB.getRadius()))){
CA.setCenterX(x + xOffset);
CA.setCenterY(y + yOffset);
}else{
/**************THE PROBLEM :Condition to drag************/
CA.setCenterX(CA.getCenterX() - (CB.getCenterX()-CA.getCenterX()));
CA.setCenterY(CA.getCenterY() - (CB.getCenterY()-CA.getCenterY()));
/*What condition must be established for the
* circle to behave correctly
*/
/********************************************************/
}
}
public static void main(String[] args) {
launch(args);
}
}
Here is a brief overview :
Note:
for my defense, i searched and found several subject close to mine but which have no precise or exact solution, among which:
-The circle remains blocked at the time of the collision
-Two circle that push each other
-JavaScript, Difficult to understand and convert to java
Thank you for your help !
Point2D can be interpreted as a 2D vector, and has useful methods for creating new vectors from it, etc. You can do:
private void drag(double x, double y){
// place drag wants to move circle to:
Point2D newCenter = new Point2D(x + xOffset, y+yOffset);
// center of fixed circle:
Point2D fixedCenter = new Point2D(CB.getCenterX(), CB.getCenterY());
// minimum distance between circles:
double minDistance = CA.getRadius() + CB.getRadius() ;
// if they overlap, adjust newCenter:
if (newCenter.distance(fixedCenter) < minDistance) {
// vector between fixedCenter and newCenter:
Point2D newDelta = newCenter.subtract(fixedCenter);
// adjust so that length of delta is distance between two centers:
Point2D adjustedDelta = newDelta.normalize().multiply(minDistance);
// move newCenter to match adjusted delta:
newCenter = fixedCenter.add(adjustedDelta);
}
CA.setCenterX(newCenter.getX());
CA.setCenterY(newCenter.getY());
}
Obviously, you could do all this without using Point2D and just doing the computation, but I think the API calls make the code easier to understand.
I am trying to get a node's point to connect to another node's point depending on the orientation of the first node. The problem is, I am having trouble thinking of the way a node call tell whether it is left, right , north or south given its X and Y positions and Width and Height.
Here is an image of the problem:
I have tried comparing the X and Y values, for example, if (child.Y > parent.Y) -> snap to southPoint, but it is too vague and will snap to north or south if it is placed right or left. Any idea on more specific conditions I can use to get the correct orientation? Thank you
You could do it by defining anchor points on your box:
public class UMLClassBox {
private Point2D north;
private Point2D east;
private Point2D south;
private Point2D west;
public List<Point> getAnchorPoints() {
...
}
}
As a second step you then can have a method to find the two nodes that are closest:
public Point2D[] findClosestNodes(UMLClassBox box1, UMLClassBox box2) {
Point2D[] closest = new Point2D[2];
double shortestDistance = Double.MAX_VALUE;
for (Point2D anchorBox1 : box1.getAnchorPoints()) {
for (Point2D anchorBox2 : box2.getAnchorPoints()) {
double distance = anchorBox1.distance(anchorBox2);
if (distance < shortestDistance) {
shortestDistance = distance;
closest[0] = anchorBox1;
closest[1] = anchorBox2;
}
}
}
return closest;
}
While there is a chance that two anchor points pairs have the same distance, this should not matter in this case because either of those would be equally valid.
This is a follow up to my last question:
How can I draw a circle to the screen with PlayN?
For my simple case, I want to programmatically create a single colored circle and move it across a 2-D plain (doesn't need to use box2d lib).
A real-world example would likely involve animating several circles. Two real-world examples for this case (sorry, I had to remove the links -- not enough karma!):
Browsmos for Chrome
Ants AI Challenge
It was suggested in response to my last question that I would want to use the ImmediateLayer class, so I am looking to understand how to properly incorporate this into my game loop.
Here's is my code sample:
public class SimpleCircleAnimation implements Game {
// Surface
private GroupLayer rootLayer;
private ImmediateLayer surface;
private Canvas canvas;
private Circle circle;
private CanvasImage circleImage;
#Override
public void init() {
// create root layer
rootLayer = graphics().rootLayer();
// a simple circle object
int circleX = 0; int circleY = 0;
int circleRadius = 20;
circle = new Circle(circleX, circleY, circleRadius);
// create an immediate layer and add to root layer
ImmediateLayer circleLayer = graphics().createImmediateLayer(new ImmediateLayer.Renderer() {
public void render (Surface surf) {
circleImage = graphics().createImage(circle.radius*2, circle.radius*2);
canvas = circleImage.canvas();
canvas.setFillColor(0xff0000eb);
canvas.fillCircle(circle.radius, circle.radius, circle.radius);
surf.drawImage(circleImage, circle.x, circle.y);
}
});
rootLayer.add(circleLayer);
}
#Override
public void paint(float alpha) {
}
#Override
public void update(float delta) {
// move circle
int newX = circle.x + 4; int newY = circle.y + 4;
circle.setPoint(newX, newY);
}
#Override
public int updateRate() {
return 25;
}
}
This successfully moves the circle diagonally down the screen from left to right. A couple questions:
Is this implemented properly?
In the case of multiple animated circles, is the idea with ImmediateLayer that you would create a circle image for each circle within the Renderer callback? Or would you perhaps create an Immediate Layer for each circle and add those to the root layer?
I would not use ImmediateLayer wiht render (Surface surf) adapter. Here u have, inside the render method creation of an image
circleImage = graphics().createImage(circle.radius*2, circle.radius*2);
just put this in the paint method
surf.drawImage(circleImage, circle.x, circle.y);
using the normal layer and u should be fine
Painting is done in paint method, and do not put calculations there
Update is for calculations, and physics oriented stuff
I discovered a detailed practical example of ImmediateLayer usage in the Cute Game source within the PlayN Samples:
CuteGame.java (code.google.com)