I am having a class inherited from QGraphicsScene. I have a mouseMoveEvent in this class. Based on the mouse move I am sending the co ordinates via signal to the main window. I have multiple GraphicViews in main window. Based on the scene from which the signal is received I am display the scene co ordinates using QGraphicsTextItem.
The problem is when I move out of the scene area I am unable to hide the QGraphicsTextItem.
Can someone give me a work around for this?
Class Scene::QGraphicsScene
{
void MouseMoveEvent(QGraphicsSceneMouseEvent *Event)
{
int XPos = event.x();
int YPos = event.y();
emit SignalPos(XPos,YPos);
}
}
//In Main Window
connect(scene1,SignalPos(int,int),this,SlotPos1(int,int);
//Similarly for scene2,scene3,scene4
void MainWindow::SlotPos(int X, int Y)
{
m_qgtxtItemX.setText(QString::x);
//I want to hide this once I am out of scene.It is a member variable. I tried
//taking local variable but it didn't work.
//Similarly for y and other slots
}
Install an event filter on the scene.
scene1->installEventFilter(this);
then implement the method in the class that is referenced by "this":
bool eventFilter(QObject *watched, QEvent *event) {
if (event->type() == QEvent::Leave)
{
qDebug() << "Mouse left the scene";
}
return false;
}
Just tried it and it worked! If you are installing the event filter on more than one object, please use "watched" to differentiate between them.
Best,
Related
I'm using MRTK 2.3.0 trying to catch a "Grip Press" event from the Mixed Reality motion controller.
I've setup the Input Action in the MRTK Toolkit in the Hierarchy of Unity. I've also assigned the action to the controller's grip button in the Controller Definitions. I'm using the following code and made sure the Grip variable is assigned to the Grip Press event. Nothing happens... I'm able to catch touchpad and joystick, menu button press, but not Grip? Why?
According to this documentation: https://microsoft.github.io/MixedRealityToolkit-Unity/Documentation/MixedRealityConfigurationGuide.html the Grip should be a "float" as it's designed as single axis (I wonder why, since it's a button and not a trigger...). However, I'm trying to catch the event where I can... not working...
Anyone understand what the heck I'm trying to say here? (sorry, the code below also includes other events that I can catch without a hickup).
using System.Collections;
using System.Collections.Generic;
using UnityEngine;
using Microsoft.MixedReality.Toolkit.Input;
using Microsoft.MixedReality.Toolkit;
public class TestInput : MonoBehaviour, IMixedRealityBaseInputHandler,
IMixedRealityInputHandler<Vector2>, IMixedRealityInputHandler
{
public MixedRealityInputAction Trigger, Grip, Scroll, Swipe, Joystick, DPad, TouchPadClicked, TouchPadTouched, MenuPressed;
public MixedRealityInputAction DPadUp, DPadDown, DPadLeft, DPadRight, JoystickUp, JoystickDown, JoystickLeft, JoystickRight;
private bool TouchpadPressed, MenuButtonPressed, GrabPressed, TouchDialogPadPressed, TouchpadClicked, Thouchpadtouched,
ThumbstickActive, ThumbstickPressed, VRManipulationStarted, ScrubStarted, PlayPause = false;
private void OnEnable()
{
IMixedRealityInputSystem inputSystem;
if (MixedRealityServiceRegistry.TryGetService<IMixedRealityInputSystem>(out inputSystem))
{
inputSystem?.RegisterHandler<IMixedRealityBaseInputHandler>(this);
inputSystem?.RegisterHandler<IMixedRealityInputHandler<Vector2>>(this);
inputSystem?.RegisterHandler<IMixedRealityInputHandler>(this);
inputSystem?.RegisterHandler<IMixedRealityInputHandler<float>>(this);
}
}
private void OnDisable()
{
IMixedRealityInputSystem inputSystem;
if (MixedRealityServiceRegistry.TryGetService<IMixedRealityInputSystem>(out inputSystem))
{
inputSystem?.UnregisterHandler<IMixedRealityBaseInputHandler>(this);
inputSystem?.UnregisterHandler<IMixedRealityInputHandler<Vector2>>(this);
inputSystem?.UnregisterHandler<IMixedRealityInputHandler>(this);
inputSystem?.UnregisterHandler<IMixedRealityInputHandler<float>>(this);
}
}
public void OnInputChanged(InputEventData<Vector2> ed)
{
Debug.Log("InputChanged");
if (ed.MixedRealityInputAction == DPad)
{
Debug.Log("Touched Touchpad at:" + ed.InputData.x.ToString() + "," + ed.InputData.y.ToString());
}
if (ed.MixedRealityInputAction == Joystick)
{
Debug.Log("Touched Joystick at:" + ed.InputData.x.ToString() + "," + ed.InputData.y.ToString());
}
}
public void OnInputChanged(InputEventData ed)
{
if (ed.MixedRealityInputAction == MenuPressed)
{
Debug.Log("Menu button pressed");
//ActionText.text = "Grab pressed";
}
}
public void OnInputChanged(InputEventData<float> ed)
{
Debug.Log("Float Changed");
if (ed.MixedRealityInputAction == Grip)
Debug.Log("Grab Pressed");
}
Ok... I got this..... It's a BUG in MRTK.
I have 18 actions defined (when you have DPad left, up, down etc it goes by fast!).
When I try to assign my MixedRealityInputAction to an InputAction in the Inspector, when I try to select GrabPressed (which is the 18th action), I get
IndexOutOfRangeException: Index was outside the bounds of the array.
Microsoft.MixedReality.Toolkit.Input.Editor.InputActionPropertyDrawer.OnGUI
Therefore the action does not work....
I tried to use another InputAction instead and assign it to the grab button (I used Select in this case) and it works like a charm....
Beware....
Try
Input.GetAxisRaw(ControllerMappingLibrary.AXIS_11) > 0 || Input.GetAxisRaw(ControllerMappingLibrary.AXIS_12) > 0;
AXIS_11 is grip press for the left controller while 12 is for the right controller
My monogame game has stopped responding to mouse clicks. Prior to version 3.5, this was working fine. Here's how I'm currently getting the input:
protected override void Update (GameTime game_time)
{
Mouse_Input (game_time);
}
void Mouse_Input(GameTime game_time)
{
mouse_current = Mouse.GetState();
if (mouse_current.LeftButton == ButtonState.Pressed)
{
// click
}
}
Setting breakpoints in the function reveals all the code is being hit, but LeftButton is always ButtonState.Released.
I've tried with both a wired mouse and the trackpad. Keyboard input is working fine. Anyone else running into this?
I always use this way.
MouseState currentMouseState;
MouseState oldMouseState;
public bool checkClick()
{
oldMouseState = currentMouseState;
currentMouseState = Mouse.GetState();
if (Visible)
{
if (currentMouseState.LeftButton == ButtonState.Pressed && oldMouseState.LeftButton == ButtonState.Released)
{
return true;
}
}
}
If you want to check if the Mouse clicks on a Rectangle (Hud elements for example)
public bool checkClickRectangle(Rectangle rec)
{
oldMouseState = currentMouseState;
currentMouseState = Mouse.GetState();
if (Visible)
{
if (rec.Contains(new Vector2(Mouse.GetState().X, Mouse.GetState().Y)) && currentMouseState.LeftButton == ButtonState.Pressed && oldMouseState.LeftButton == ButtonState.Released)
{
return true;
}
}
}
This was actually a not a problem with Monogame, but a problem in my game logic that was very difficult to track down.
After upgrading to 3.5, I had to reconfigure how my Texture2D's were being loaded, which also meant refactoring some classes. I ended up with a class within a class which were both inheriting from Game.
public class Brush_Control : Game
{
public class Tile : Game
{
Process of elimination narrowed the search to this class. I believe this caused an infinite loop that interfered with the input somehow, but without throwing an error or causing an obvious freeze.
Removing the inner Game reference as a parent fixed the problem, and it turns out I no longer need to have it there anyway.
Could you please advise how i would go about using the input touch function in Unity to make an object changes its x direction every time the user tap on the screen. For example, for 2d setting game, an object is moving forward (to the right) in the x position, if the user tap then the object would move backward in the x position (to the left). Sorry no code is produced.
It's simple as your name "tony" :)
What you can do is to make a simple script which'd move your object to left and right. And on screen touch you can easily change its direction by just -1 multiplication.
Simple script that you can attach to your object.
using UnityEngine;
using System.Collections;
public class MoveObject : MonoBehaviour
{
float _limit = 5;
// 1 for right and -1 for left.
float _direction = 1;
// You can call it as speed
float _speed = 0.01f;
void Start ()
{
}
void Update ()
{
transform.position = Vector3.MoveTowards (transform.position, new Vector3 (transform.position.x + _direction, transform.position.y, transform.position.z), _speed);
if (Input.GetMouseButtonDown (0))
_direction *= -1;
}
}
Hope this helps :)
I've been reading various tutorials on how to write a contact listener, and I can't wrap my head around it.
Here is what I have so far:
In each of the classes that I have representing a physics object I do:
_body->SetUserData(self);
I write a contact listener class containing the following two methods:
void ContactListener::BeginContact(b2Contact* contact)
{
// Box2d objects that collided
b2Fixture* fixtureA = contact->GetFixtureA();
b2Fixture* fixtureB = contact->GetFixtureB();
// Sprites that collided
MyNode* actorA = (MyNode*) fixtureA->GetBody()->GetUserData();
MyNode* actorB = (MyNode*) fixtureB->GetBody()->GetUserData();
}
void ContactListener::EndContact(b2Contact* contact)
{
// Box2d objects that collided
b2Fixture* fixtureA = contact->GetFixtureA();
b2Fixture* fixtureB = contact->GetFixtureB();
// Sprites that collided
MyNode* actorA = (MyNode*) fixtureA->GetBody()->GetUserData();
MyNode* actorB = (MyNode*) fixtureB->GetBody()->GetUserData();
}
I don't know what to do next. I now have the two sprites which are colliding, but I want to do the following:
1) When they collide, I want to remove one of the sprites from the world, based on the type of object. (for example if one a cat object and the other is a mouse object, I want to remove the mouse object.
2) I want to let the cat object know it ate a mouse
3) I want the cat to continue moving as if it didn't contact with the mouse.
4) I still wan't the cat to collide normally with things like the terrain.
What do I do next ? I'm pretty clueless on what to do? How do I get the cat to continue to collide normally with the terrain, but not with the mouse? When do I remove the mouse?
Having an "Entity" class that hold the reference to the Box2D body and does the manipulations on it is definitely a good way to go. If you have a Spaceship class vs. a Meteor class, each of them can supply their own derived methods of controlling the body (AI), but each of them has common logic and code to support operations on "Things that have a body" (e.g. common "Entity" base class). I think you are on the right track.
It gets a little murky when the contacts start happening. This is where you start getting into the architecture of your overall system, not just the structure of the physics world or a single Coco2d Scene.
Here is how I have done this in the past:
First I set up the contact listener, listed below:
class EntityContactListener : public ContactListener
{
private:
GameWorld* _gameWorld;
EntityContactListener() {}
typedef struct
{
Entity* entA;
Entity* entB;
} CONTACT_PAIR_T;
vector<CONTACT_PAIR_T> _contactPairs;
public:
virtual ~EntityContactListener() {}
EntityContactListener(GameWorld* gameWorld) :
_gameWorld(gameWorld)
{
_contactPairs.reserve(128);
}
void NotifyCollisions()
{
Message* msg;
MessageManager& mm = GameManager::Instance().GetMessageMgr();
for(uint32 idx = 0; idx < _contactPairs.size(); idx++)
{
Entity* entA = _contactPairs[idx].entA;
Entity* entB = _contactPairs[idx].entB;
//DebugLogCPP("Contact Notification %s<->%s",entA->ToString().c_str(),entB->ToString().c_str());
msg = mm.CreateMessage();
msg->Init(entA->GetID(), entB->GetID(), Message::MESSAGE_COLLISION);
mm.EnqueueMessge(msg, 0);
msg = mm.CreateMessage();
msg->Init(entB->GetID(), entA->GetID(), Message::MESSAGE_COLLISION);
mm.EnqueueMessge(msg, 0);
}
_contactPairs.clear();
}
void PreSolve(b2Contact* contact, const b2Manifold* oldManifold)
{
b2Fixture* fixtureA = contact->GetFixtureA();
b2Body* bodyA = fixtureA->GetBody();
Entity* entityA = bodyA->GetUserData();
b2Fixture* fixtureB = contact->GetFixtureB();
b2Body* bodyB = fixtureB->GetBody();
Entity* entityB = bodyB->GetUserData();
if(test if entityA and entityB should not have collision response)
{
contact->SetEnabled(false);
}
// Do this if you want there to be collision notification, even if
// there is no response.
AddContactPair(entA,entB);
}
void AddContactPair(Entity* entA, Entity* entB)
{
for(uint32 idx = 0; idx < _contactPairs.size(); idx++)
{
if(_contactPairs[idx].entA == entA && _contactPairs[idx].entB == entB)
return;
// Not sure if this is needed...
if(_contactPairs[idx].entA == entB && _contactPairs[idx].entA == entB)
return;
}
CONTACT_PAIR_T pair;
pair.entA = entA;
pair.entB = entB;
_contactPairs.push_back(pair);
}
// BEWARE: You may get multiple calls for the same event.
void BeginContact(b2Contact* contact)
{
Entity* entA = (Entity*)contact->GetFixtureA()->GetBody()->GetUserData();
Entity* entB = (Entity*)contact->GetFixtureB()->GetBody()->GetUserData();
assert(entA != NULL);
assert(entB != NULL);
// Not sure this is still needed if you add it in the pre-solve.
// May not be necessary...
AddContactPair(entA, entB);
}
// BEWARE: You may get multiple calls for the same event.
void EndContact(b2Contact* contact)
{
}
};
Because of the way the engine works, you can get multiple contact hits for the same bodies. This listener filters them so if two Entities collide, you only get one message.
The Listener only stores the collisions that have occurred. It could be modified to further separate them into "begins" and "ends" for other purposes. Here, contact meant "you've been hit by something". I didn't need to know if it stopped contacting.
The call to NotifyCollisions is the "secret sauce". It sends out a message to both the contacted entities (via a message system) that they hit something and the other Entity is what they hit. Bullet hit ship. Bullet destroys self. Ship damages self based on bullet properties (GetDamageInflicted() method). This in turn signals the graphics system to remove the bullet from the display. If the ship was destroyed, it is destroyed as well.
From an overall execution standpoint:
Before you begin, assign the contact listener.
Each Cycle of your game:
Call "Update" on all your Entities. This updates their physics forces, etc.
Update Box2d world.
Call NotifyCollisions on Listener.
Remove dead Entities from system.
Was this helpful?
I'm at a loss here. I've got a simple terrain generation algorithm working, and I've got some simple keyboard navigation working by extending ViewPlatformAWTBehavior and handling my own events. That's all well and good, and I can follow terrain. Hooray!
What I'd like to do is get some simple "mouse look" working. MouseRotate is close, but I'm looking for something more like an FPS ... where you aren't going to roll the camera, you're limited to 90 degrees vertical (up or down), and the mouse cursor is captured by the JFrame (with an escape).
I just can't seem to get it to work. I can capture the mouse event just fine, and mouseMoved works. I could probably just use an invisible cursor, and that would get me a large part of the way there (maybe), but I'd be stuck when trying to keep the mouse in the screen ... as soon as you're out of the frame, the mouse would be visible and stop rotating the view.
I keep thinking I must be going about this wrong, because this has to be a fairly common implementation, but I can't find anything on it.
Does anyone have some pointers or references?
If you want this is the code to make a full screen window which will help you with mouse problems
package fullscreen;
import java.awt.*;
import java.awt.event.ActionEvent;
import java.awt.event.ActionListener;
import javax.swing.*;
public class FullScreen extends JFrame{
private JPanel contentPane = new JPanel();
private JButton fullscreenButton = new JButton("Fullscreen Mode");
private boolean Am_I_In_FullScreen = false;
private int PrevX,PrevY,PrevWidth,PrevHeight;
public static void main(String[] args) {
FullScreen frame = new FullScreen();
frame.setSize(600,500);
frame.setVisible(true);
}
public FullScreen(){
super("My FullscreenJFrame");
setContentPane(contentPane);
//From Here starts the trick
FullScreenEffect effect = new FullScreenEffect();
fullscreenButton.addActionListener(effect);
contentPane.add(fullscreenButton);
fullscreenButton.setVisible(true);
}
private class FullScreenEffect implements ActionListener{
#Override
public void actionPerformed(ActionEvent arg0) {
if(Am_I_In_FullScreen == false){
PrevX = getX();
PrevY = getY();
PrevWidth = getWidth();
PrevHeight = getHeight();
dispose(); //Destroys the whole JFrame but keeps organized every Component
//Needed if you want to use Undecorated JFrame
//dispose() is the reason that this trick doesn't work with videos
setUndecorated(true);
setBounds(-10,-100,getToolkit().getScreenSize().width+30,getToolkit()
.getScreenSize().height+110);
setVisible(true);
Am_I_In_FullScreen = true;
}
else{
setVisible(true);
setBounds(PrevX, PrevY, PrevWidth, PrevHeight);
dispose();
setUndecorated(false);
setVisible(true);
Am_I_In_FullScreen = false;
}
}
}
}