Can I get the value of MotionEvent.getRawX()/getRawY() of other pointers ?
MotionEvent.getRawX() api reference
The api says that uses getRawX/getRawY to get original raw X/Y coordinate, but it only for 1 pointer(the last touched pointer), is it possible to get other pointer's raw X/Y coordinate ?
Indeed, the API doesn't allow to do this, but you can compute it. Try that :
public boolean onTouch(final View v, final MotionEvent event) {
int rawX, rawY;
final int actionIndex = event.getAction() >> MotionEvent.ACTION_POINTER_ID_SHIFT;
final int location[] = { 0, 0 };
v.getLocationOnScreen(location);
rawX = (int) event.getX(actionIndex) + location[0];
rawY = (int) event.getY(actionIndex) + location[1];
}
A solution worth trying for most use cases is to add this to the first line of the onTouchEvent it simply finds the difference between the raw and processed, and shifts the location of the MotionEvent event, by that amount. So that all the getX(int) values are now the raw values. Then you can actually use the getX() getY() stuff as the Raw values.
event.offsetLocation(event.getRawX()-event.getX(),event.getRawY()-event.getY());
While Ivan's point is valid, it's simply the case that applying a matrix directly to the view itself sucks so bad you likely shouldn't do it. It's weird and inconsistent between devices, cause the touch events to fall out of view and get declined, etc. If you are moving a view around like that you are better off simply overloading the onDraw() and applying that matrix to the canvas, then applying the inverse matrix to the MotionEvent so everything meshes up right. Then you can properly react to the events with proper and fine grain control. And, if you do that, my solution here wouldn't be subject to Ivan's objection.
It might be not enough to just shift local coordinates by a view's location if the view is rotated. In this case you need something like this:
void getRowPoint(MotionEvent ev, int index, PointF point){
final int location[] = { 0, 0 };
getLocationOnScreen(location);
float x=ev.getX(index);
float y=ev.getY(index);
double angle=Math.toDegrees(Math.atan2(y, x));
angle+=getRotation();
final float length=PointF.length(x,y);
x=(float)(length*Math.cos(Math.toRadians(angle)))+location[0];
y=(float)(length*Math.sin(Math.toRadians(angle)))+location[1];
point.set(x,y);
}
The getLocationOnScreen answer works most of the time, but I was seeing it return incorrect values sometimes (when I was repositioning and re-parenting the view while the touch event was taking place), so I found an alternate approach that works more reliably.
If you look at the implementation of getRawX, it calls a private native function that accepts a pointerIndex, but the MotionEvent class only ever calls it with index 0:
public final float getRawX() {
return nativeGetRawAxisValue(mNativePtr, AXIS_X, 0, HISTORY_CURRENT);
}
Unfortunately, nativeGetRawAxisValue is private, but you can hack around that by using reflection to give yourself access to everything you need. Here's what the code looks like:
private Point getRawCoords(MotionEvent event, int pointerIndex) {
try {
Method getRawAxisValueMethod = MotionEvent.class.getDeclaredMethod(
"nativeGetRawAxisValue", long.class, int.class, int.class, int.class);
Field nativePtrField = MotionEvent.class.getDeclaredField("mNativePtr");
Field historyCurrentField = MotionEvent.class.getDeclaredField("HISTORY_CURRENT");
getRawAxisValueMethod.setAccessible(true);
nativePtrField.setAccessible(true);
historyCurrentField.setAccessible(true);
float x = (float) getRawAxisValueMethod.invoke(null, nativePtrField.get(event),
MotionEvent.AXIS_X, pointerIndex, historyCurrentField.get(null));
float y = (float) getRawAxisValueMethod.invoke(null, nativePtrField.get(event),
MotionEvent.AXIS_Y, pointerIndex, historyCurrentField.get(null));
return new Point((int)x, (int)y);
} catch (NoSuchMethodException|IllegalAccessException|InvocationTargetException|
NoSuchFieldException e) {
throw new RuntimeException(e);
}
}
Of course, the MotionEvent internals aren't documented, so this approach might crash on past or future versions of the SDK, but it seems to be working for me.
Edit: It looks like the type of mNativePtr and the nativePtr param changed from int to long in API level 20, so if you're targeting API level 19 or earlier, the above code will crash because getDeclaredMethod won't find anything. To fix this in my code, I just fetched the method by name instead of full type signature, which happens to work in this case. There isn't a way to directly look up methods with a given name, so I looped through the declared methods at static init time and saved the matching one to a static field. Here's the code:
private static final Method NATIVE_GET_RAW_AXIS_VALUE = getNativeGetRawAxisValue();
private static Method getNativeGetRawAxisValue() {
for (Method method : MotionEvent.class.getDeclaredMethods()) {
if (method.getName().equals("nativeGetRawAxisValue")) {
method.setAccessible(true);
return method;
}
}
throw new RuntimeException("nativeGetRawAxisValue method not found.");
}
Then I used NATIVE_GET_RAW_AXIS_VALUE in place of the getRawAxisValueMethod in the above code.
There is no API to get pointer's specific RawX and RawY.
But you can calculate similar values with regards to View's position to the parent and its rotation.
In case if parent view occupies the entire touch area you are trying to handle, using Matrix will help you to solve your problem:
private float[] calcRawCoords(MotionEvent event, int pointerIndex) {
Matrix screenMatrix = new Matrix();
screenMatrix.postRotate(getRotation(), mPivotX, mPivotY);
screenMatrix.postTranslate(getLeft(), getTop());
float viewToScreenCoords[] = {event.getX(pointerIndex), event.getY(pointerIndex)};
screenMatrix.mapPoints(viewToScreenCoords);
return viewToScreenCoords;
}
Related
There are few actors on stage, lines are drawn connecting centers of them, like a graph, Nodes and edges. Nodes are gradable. On Drag I am using the following code
public void touchDragged(InputEvent event, float eventOffsetX, float
eventOffsetY, int pointer)
{
float deltaX = eventOffsetX - self.grabOffsetX;
float deltaY = eventOffsetY - self.grabOffsetY;
self.moveBy(deltaX, deltaY);
moveCoordinatesBy((int)deltaX,(int)deltaY);
}
the Method moveCoordinatesBy is updating the coordinate of the center of the node. which is used to draw lines (Edges) connecting the. This is working Fine.
My problem is - When I am using moveTo action for Nodes, I need to update edges at the same time. For this I need to update coordinates of there center. To do this I am using the following code in the act method of the Node -
public void act(float dt)
{
super.act(dt);
Vector2 loc = new Vector2();
loc.x = self.getX()+ self.getWidth()/2;
loc.y = self.getY() + self.getHeight()/2;
Vector2 v = new Vector2();
v = self.localToStageCoordinates(loc);
setCoordinates((int)v.x, (int)v.y);
}
The coordinates I am getting are not center of the node, they are almost multiplied by 2. Am I doing something wrong?
And also getWidth() and self.getHeight() returning double the size. When actor (node) is scaled getWidth() and self.getHeight() are returning arbitery values depending on how for it is located on stage. Further from (0,0) larger the returned value.
v = self.localToStageCoordinates(loc);
above line of code not needed, beacuse all the actors were directly added to the stage. By removing localToStageCoordinates(loc) its working fine now.
Im trying to implement the Drag And Drop method/class in libgdx, but what´s different for me is that I dont have a "Target" to drop my actor. I simply want to drag it around all over the screen. I haven´t really got any further than this.
dd = new DragAndDrop();
dd.addSource(new Source(stoneImage){
#Override
public Payload dragStart(InputEvent event, float x, float y, int pointer) {
Payload payload = new Payload();
payload.setObject("some payload");
payload.setDragActor(stoneImage);
return payload;
}
});
Here I set the Image as an actor:
stageMove.addActor(stoneImage);
At the moment it won´t move when I try to drag it.
I suggest you 2 easy ways to solve your problem.
1) Add a DragListener to your actor:
actor.addListener(new DragListener() {
public void drag(InputEvent event, float x, float y, int pointer) {
actor.moveBy(x - actor.getWidth() / 2, y - actor.getHeight() / 2);
}
});
2) Create a generic actor which size is filled to the stage, and use it as the target actor.
I am overriding onInterceptTouch(MotionEvent) for the purpose of allowing horizontal scrolling. What I am finding however is that I cannot detect when the user has touched the embedded v. The x,y on the view are like 2000, 2400 while the MotionEvent.getX(),getY() are like 400,500
View v = findViewById(R.id.myView);
Rect r = new Rect();
v.getHitRect(r);
int[] loc = new int[2];
v.gtLocationOnScreen(loc);
int x = loc[0];
int y = loc[1];
// Result x,y are huge numbers 2400, etc
// event.getX() is 30, event.getY() == 500 nothing close to 2400.
if (r.contains((int)event.getX(), (int)event.getY())
{
return false; // this is never true even when I click right on View v.
}
I know this is an old question and is mostly answered in the linked post, but I just came across this problem so I thought I'd fill in my solution.
private boolean isTouchInView(View view, MotionEvent event) {
Rect hitBox = new Rect();
view.getGlobalVisibleRect(hitBox);
return hitBox.contains((int) event.getRawX(), (int) event.getRawY());
}
Try using getRawX() and getRawY(). These will give you the absolute positions you need.
See:
How do I know if a MotionEvent is relative or absolute?
You also have to adjust the location of your destination view to account for any displacement by other views like so:
int[] screenLocation = new int[2];
view.getLocationOnScreen(screenLocation);
hitRect.offset(screenLocation[0] - view.getLeft(), screenLocation[1] - view.getTop());
//Then check if source view is contained in target view
x=event.getRawX();
y=event.getRawY();
if (hitRect.contains(x, y)) {
//do your stuff
}
Im making a program about the game Go and im trying to write some code to allow a person to place a bead, which really just shows an invisible bead. But after an hour of going back and forth with it i just cant seem to figure out why the Y coordinate of my motionevent always seems 2 rows down from where it needs to be.
My code for filling my array
for(int y=0; y< 9;y++)
{
for(int x=0;x<9;x++)
{
String name="Whitebead" + ((9*y)+x+2); //The +2 is because of a naming problem
WhitePieces[x][y]=(ImageView) findViewById(getResources().getIdentifier(name, "id", getPackageName()));
}
}
My code for handling motion events
#Override
public boolean onTouchEvent(MotionEvent e)
{
if(e.getAction() == MotionEvent.ACTION_DOWN)
{
float X=e.getRawX();
float Y=e.getRawY();
if(X>Board.getLeft() && X<Board.getRight())
if(Y< Board.getTop() && Y>Board.getBottom());
player1.placepiece(X, Y);
}
return super.onTouchEvent(e);
}
And finally my code that resolves which bead to what coordinate
public void placepiece(float X, float Y)
{
int[] pieceindex=resolvePiece(X,Y);
pieces[pieceindex[0]][pieceindex[1]].setVisibility(ImageView.VISIBLE);
}
private int[] resolvePiece(float x, float y) {
int Xindex=0;
int[] oldcoords= new int[2]; //cordinates are like so {xCoord, Ycoord}
int[] newcoords= new int[2];
oldcoords[0]=pieces[0][0].getLeft(); //set the oldcoordinates to the first index
oldcoords[1]=pieces[0][0].getTop();
for(int i=1; i<9;i++) //go through the 9 indexs to find the closest X value
{
newcoords[0]=pieces[i][0].getLeft();
newcoords[1]=pieces[i][0].getTop();
if(Math.abs((int)x-newcoords[0])<Math.abs((int)x-oldcoords[0]))
{
Xindex=i;
oldcoords[0]=newcoords[0];
}
}
int Yindex=0;
oldcoords[0]=pieces[0][0].getLeft(); //Reset oldcoords again
oldcoords[1]=pieces[0][0].getTop();
for(int n=1; n<9;n++) //Go through the 9 indexes for the closest Y value
{
newcoords[0]=pieces[0][n].getLeft();
newcoords[1]=pieces[0][n].getTop();
if(Math.abs((int)y-newcoords[1])<Math.abs((int)y-oldcoords[1]))
{
Yindex=n;
oldcoords[1]=newcoords[1];
}
}
int[] rInt= new int[]{Xindex,Yindex};
return rInt;
}
//////EDIT: Fixed
I figured it out, at the top of the android window is about and inch of space where the title and battery life and stuff go, when you get motion coordinates it takes in the whole screen, where .getTop() only gets from where linear layout Starts. SO instead of using .getTop or.getLeft i used .getLocationInWindow(oldcoord[]) and it places the info i needed into my oldcoord array.
//////EDIT: Fixed I figured it out, at the top of the android window is about and inch of space where the title and battery life and stuff go, when you get motion coordinates it takes in the whole screen, where .getTop() only gets from where linear layout Starts. SO instead of using .getTop or.getLeft i used .getLocationInWindow(oldcoord[]) and it places the info i needed into my oldcoord array.
I'm trying to implement scroll animation for gallery programmatically.
Tried with setSelection(int pos, Boolean animate) and it's not working.
Is there anyway to override setSelection() method.
Just now I have got this problem. I was need to move just one element of the gallery, so the best solution to me was to emulate key down event
myGallery.onKeyDown(KeyEvent.KEYCODE_DPAD_RIGHT, null);
or
myGallery.onKeyDown(KeyEvent.KEYCODE_DPAD_LEFT, null);
Gallery.setSelection(int position, boolean animate);
Refer below URL:
http://groups.google.com/group/android-developers/browse_thread/thread/9140fd6af3061cdf/7f89e53ae53e455b?lnk=gst&q=setselection#7f89e53ae53e455b
Soln:
If you're still looking, I have two possible solutions for you, both
slightly unpleasant:
(1) You can make the gallery do a fling with a chosen velocity, thus:
myGallery.onFling(null, null, velocity, 0);
By tweaking the velocity, you can set up values to move the selection
by one or two in either direction. As the Gallery self-centers, you do
not need to get the destination exactly right.
(2) As the Gallery source is available, you can modify it in order to
implement your own Gallery. It doesn't look as though you need to add
much code to be able to control a fling so as to end at your chosen
selection.
I thought I was going to have to do (2), but found I could get away
with (1) for my problem.
Based on Kurru's excellent thinking of simulating clicking next or previous view.
//scroll forward or backward
private void scroll(int type){
View selectedV = mG.getSelectedView();
int idx = mG.indexOfChild(selectedV);
switch(type){
case FORWARD:
default:
if(idx<mG.getChildCount()-1)
idx++;
break;
case BACKWARD:
if(idx>0)
idx--;
break;
}
//now scrolled view's child idx in gallery is gotten
View nextView = mG.getChildAt(idx);
//(x,y) in scrolled view is gotten
int x = nextView.getLeft()+nextView.getWidth()/2;
int y = nextView.getTop()+nextView.getHeight()/2;
String out = String.format("x=%d, y=%d", x, y);
Log.i(TAG+".scroll", out);
//Kurru's simulating clicking view
MotionEvent event = MotionEvent.obtain(100, 100, MotionEvent.ACTION_DOWN, x, y, 0);
mG.onDown(event);
boolean res = mG.onSingleTapUp(null);
Log.i(TAG+".scroll", "onSingleTapUp return =" + res);
}
I was looking through the Gallery source to see if I could get this feature. It looks like something is possible with this code. However I gave up before I could get it working. It seems like I wasn't passing in the correct coordinates so res always returned false. Would return true if it worked.
Just posting this here in-case someone else wants to give a go at fixing it! (Please post your solution if you manage it!)
Rect rect = new Rect();
gallery.getHitRect(rect);
int x = rect.centerX()+getWindowManager().getDefaultDisplay().getWidth();
int y = rect.centerY();
MotionEvent event = MotionEvent.obtain(100, 100, MotionEvent.ACTION_DOWN, x, y, 0);
timesGallery.onDown(event);
boolean res = timesGallery.onSingleTapUp(null);
I made little change in code given by "Kurru". this is working now
Rect rect = new Rect();
gallery.getHitRect(rect);
int width = Math.abs(rect.width());
if(!isForwardScroll){
width = width * -1;
}
int x = rect.centerX()+width/2;
int y = rect.centerY();
MotionEvent event = MotionEvent.obtain(100, 100, MotionEvent.ACTION_DOWN, x, y, 0);
gallery.onDown(event);
boolean res = gallery.onSingleTapUp(null);