in my app the user is able to move or rotate a box2d fixture. The collision detection works fine. But if I set the fixture as a sensor, with the SetSensor(true) method, the collision detection reacts weird.
In the following picture the red rectangle is the sensor, but a collision is also detected if an other fixture collides with the black border.
Image: http://img851.imageshack.us/img851/7292/rect.png
Is it possible that only the red rectangle reacts as a sensor?
BR
I found the help here:
Box2D Forum
I just forgot to check if the contact is touching (with the IsTouching() method)! Now it works fine!!!
b2BodyDef bodyDef;
bodyDef.type = b2_staticBody;
bodyDef.position.Set(position.x/PTM_RATIO, position.y/PTM_RATIO);
bodyDef.userData = NULL;
b2Body *body = _game.world->CreateBody(&bodyDef);
b2PolygonShape dynamicBox;
dynamicBox.SetAsBox(size.x/2/PTM_RATIO, size.y/2/PTM_RATIO);
b2FixtureDef fixtureDef;
fixtureDef.shape = &dynamicBox;
fixtureDef.isSensor = true;
body->CreateFixture(&fixtureDef);
body->SetTransform(body->GetPosition(), rotatingAngle);
Related
I'm currently trying to create a multiplayer game using mirror!
So far I've been successful in creating a lobby and set up a simple character model with a player locomotion script I've taken and learnt inside out from Sebastian Graves on YouTube (you may know him as the dark souls III tutorial guy)
This player locomotion script uses unity's package 'Input System' and is also dependent on using the camera's .forward and .right directions to determine where the player moves and rotates instead of using forces on the rigidbody. This means you actually need the camera free in the scene and unparented from the player.
Here is my HandleRotation() function for rotating my character (not the camera's rotation function):
private void HandleRotation()
{
// target direction is the way we want our player to rotate and move // setting it to 0
Vector3 targetDirection = Vector3.zero;
targetDirection = cameraManager.cameraTransform.forward * inputHandler.verticalInput;
targetDirection += cameraManager.cameraTransform.right * inputHandler.horizontalInput;
targetDirection.Normalize();
targetDirection.y = 0;
if (targetDirection == Vector3.zero)
{
// keep our rotation facing the way we stopped
targetDirection = transform.forward;
}
// Quaternion's are used to calculate rotations
// Look towards our target direction
Quaternion targetRotation = Quaternion.LookRotation(targetDirection);
// Slerp = rotation between current rotation and target rotation by the speed you want to rotate * constant time regardless of framerates
Quaternion playerRotation = Quaternion.Slerp(transform.rotation, targetRotation, rotationSpeed * Time.deltaTime);
transform.rotation = playerRotation;
}
It's also worth mentioning I'm not using cinemachine however I'm open to learning cinemachine as it may be beneficial in the future.
However, from what I've learnt and managed to find about mirror is that you have to parent the Main Camera under your player object's prefab so that when multiple people load in, multiple camera's are created. This has to happen on a Start() function or similar like OnStartLocalPlayer().
public override void OnStartLocalPlayer()
{
if (mainCam != null)
{
// configure and make camera a child of player
mainCam.orthographic = false;
mainCam.transform.SetParent(cameraPivotTransform);
mainCam.transform.localPosition = new Vector3(0f, 0f, -3f);
mainCam.transform.localEulerAngles = new Vector3(0f, 0f, 0f);
cameraTransform = GetComponentInChildren<Camera>().transform;
defaultPosition = cameraTransform.localPosition.z;
}
}
But of course, that means that the camera is no longer independent to the player and so what ends up happening is the camera rotates when the player rotates.
E.g. if I'm in game and looking to the right of my player model and then press 'w' to walk in the direction the camera is facing, my camera will spin with the player keeping the same rotation while my player spins in order to try and walk in the direction the camera is facing.
What my question here would be: Is there a way to create a system using mirror that doesn't require the camera to be parented to the player object prefab on start, and INSTEAD works by keeping it independent in the scene?
(I understand that using forces on a rigidbody is another method of creating a player locomotion script however I would like to avoid that if possible)
If anybody can help that would be greatly appreciated, Thank You! =]
With the constraint of your camera being attached to the player you will have to adapt the camera logic. What you can do is instantiate a proxy transform for the camera and copy that transforms properties to your camera, globally, every frame. The parent constraint component might do this for you. Then do all your transformations on that proxy.
Camera script with instantiation:
Transform proxy;
void Start(){
proxy = (new GameObject()).transform;
}
void Update(){
//transformations on proxy
transform.position = proxy.position;
transform.rotation = proxy.rotation;
}
I'm trying to do a little game on mobile using Unity and I've got a problem with the rotation of a maze.
To add context :
When your moving your finger on the screen, the maze is rotating on himself. There is a ball in it and you need to make it go on a cell to win the level.
When the maze is rotating too fast, the ball falls down and go through the ground and I don't know how to fix it.
I tried to play with the gravity, colliders...
This is the same when the ball is jumping (when the maze is going up and down quickly).
For the moment I just reset the ball position when you're falling.
{
ball.transform.position = new Vector3(0, 2, 0);
maze.transform.position = Vector3.zero;
maze.transform.rotation = Quaternion.identity;
}
Do you guys have some ideas ? Thanks
I had a similar problem in a tilt maze mini-game I worked on. Ideally implementing jkimishere's solution will work but I assume the maze is moving too fast for the collisions to register properly. You'll need to smooth the maze's rotation with a Lerp. In our case we had pressure plates with a tilt value, so it doesn't directly translate to your mobile use but perhaps give you a nudge in the right direction. We used:
public GameObject maze;
private float _lerpTime;
private bool _isRotating;
private Quaternion _startingRot, _desiredRot;
private void Awake()
{
_startingRot = maze.transform.localRotation;
}
private void Update()
{
//Don't want to move the maze if we don't ask for it
if(!_isRotating)
return;
//Lerp the maze's rotation
_lerpTime = Mathf.Clamp(_lerpTime + Time.deltaTime * 0.5f, 0f, 1f);
maze.transform.localRotation = Quaternion.Lerp(_startingRot, _desiredRot, _lerpTime);
//Once the maze gets where it needs to be, stop moving it
if(affectedObject.transform.localRotation.Equals(_desiredRot)
_isRotating = false;
}
private void ActivateTilt()
{
//Set the new starting point of the rotation.
_startingRot = maze.transform.localRotation;
//However you want to calculate the desired rotation here
//Reset our lerp and start rotating again
_lerpTime = 0f;
_isRotating = true;
}
This will ease the rotation of your maze over time. So that the ball can adapt to the new collider positions.
In the rigidbody(for the ball), make the collision detection to continuous, and in the rigidbody for the maze(if you have one) set the collision detection to continuous dynamic. Hope this helps!
I think that is unavoidable if you allow the player to move the platform freely. I suggest you restrain the speed at wich the player can tilt the platform. This way, the ball will have more frames to adapt to the new slope
I'm trying to get the NPC to look at the main character when I'm talking to him. I need to make sure it looks natural and that he is facing me. I know I can do Transform.LookAt() but that is too instant and unnatural.
How do I rotate the navmeshagent to face my character when its stopped moving?
Try this out to control the body orientation - the slerp is adjustable to your desired rotation speed (https://docs.unity3d.com/ScriptReference/Quaternion.Slerp.html):
private void FaceTarget(Vector3 destination)
{
Vector3 lookPos = destination - transform.position;
lookPos.y = 0;
Quaternion rotation = Quaternion.LookRotation(lookPos);
transform.rotation = Quaternion.Slerp(transform.rotation, rotation, [fill in desired rotation speed]);
}
if(agent.remainingDistance < agent.stoppingDistance)
{
agent.updateRotation = false;
//insert your rotation code here
}
else {
agent.updateRotation = true;
}
This will rotate your agent when it's distance below the stoppingDistance variable. However it will look inhuman, so if you're going for a humanoid ai I'd recommend looking at the unity Mecanim demo (particularly the locomotion scene) as it has code and animations that will properly animate the agent.
Maybe try this Head Look Controller. It's very smooth and blends with animations!
Put the char in a game object and copy the navmesh from the char to the parent, uncheck enable in char. move any scripts up too.
Just spent 5 hours to find this.
I tried to set up a prismatic joint, but I get this error upon run time:
Assertion failed: (IsLocked() == false), function CreateBody, file /Users/Aether/Developer/JFRecode/JFRecode/libs/Box2D/Dynamics/b2World.cpp, line 109.
Here is the code for the screen boundaries since I am using it as bodyB.
CGSize s = [CCDirector sharedDirector].winSize;
//define screen boundaries
// Create edges around the entire screen
b2BodyDef groundBodyDef;
groundBodyDef.position.Set(0,0);
_groundBody = _world->CreateBody(&groundBodyDef);
b2EdgeShape groundBox;
b2FixtureDef groundBoxDef;
groundBoxDef.shape = &groundBox;
groundBox.Set(b2Vec2(0,0), b2Vec2(s.width/PTM_RATIO, 0));
_bottomFixture = _groundBody->CreateFixture(&groundBoxDef);
groundBox.Set(b2Vec2(0,0), b2Vec2(0, s.height/PTM_RATIO));
_groundBody->CreateFixture(&groundBoxDef);
groundBox.Set(b2Vec2(0, s.height/PTM_RATIO), b2Vec2(s.width/PTM_RATIO,
s.height/PTM_RATIO));
_groundBody->CreateFixture(&groundBoxDef);
groundBox.Set(b2Vec2(s.width/PTM_RATIO, s.height/PTM_RATIO),
b2Vec2(s.width/PTM_RATIO, 0));
_groundBody->CreateFixture(&groundBoxDef);
Here is the code for the prismatic joint itself
b2PrismaticJointDef jointDef;
b2Vec2 worldAxis(1.0f,0.0f);
jointDef.collideConnected = true;
jointDef.Initialize(_shipBody,_groundBody,
_shipBody->GetWorldCenter(), worldAxis);
_world->CreateJoint(&jointDef);
isLocked indicates that the physics world is being updated. That means you're running this code during the world step method, likely in or as a result of a contact callback method. Delay executing this code until after the world step.
I am trying to make a game in Cocos 2D & Box 2D
but when i try to put a "setAsEdge" it just said me that the function dosn't exist
my code:
// Define the ground body.
b2BodyDef groundBodyDef;
groundBodyDef.position.Set(0, 0); // bottom-left corner
// Call the body factory which allocates memory for the ground body
// from a pool and creates the ground box shape (also from a pool).
// The body is also added to the world.
groundBody = world->CreateBody(&groundBodyDef);
// Define the ground box shape.
b2PolygonShape groundBox;
// bottom
groundBox.SetAsEdge(b2Vec2(0,FLOOR_HEIGTH/PTM_RATIO), b2Vec2(screenSize.width*2.0f/PTM_RATIO,FLOOR_HEIGTH/PTM_RATIO));
groundBody->CreateFixture(&groundBox,0);
// top
groundBox.SetAsEdge(b2Vec2(0,screenSize.height/PTM_RATIO), b2Vec2(screenSize.width*2.0f/PTM_RATIO,screenSize.height/PTM_RATIO));
groundBody->CreateFixture(&groundBox,0);
// left
groundBox.SetAsEdge(b2Vec2(0,screenSize.height/PTM_RATIO), b2Vec2(0,0));
groundBody->CreateFixture(&groundBox,0);
// right
groundBox.SetAsEdge(b2Vec2(screenSize.width*2.0f/PTM_RATIO,screenSize.height/PTM_RATIO), b2Vec2(screenSize.width*2.0f/PTM_RATIO,0));
groundBody->CreateFixture(&groundBox,0);
Error: No member named 'SetAsEdge' in 'b2PolygonShape'. I am using Cocos 2D 1.0.1 Stable
Thanks to everyone
You are working with Box2D 2.2.1 whose API has changed compared to the Box2D version in Cocos2D v1.0.1 (uses Box2D v2.1.2). I take it you're using Cocos2D 2.0 alpha? If so, it supposedly comes with updated Xcode project templates, so you might want to install these and create a project based on the new templates.
Here's how I set the border bounding box by using a body with 4 shapes attached to it:
// for the screenBorder body we'll need these values
CGSize screenSize = [CCDirector sharedDirector].winSize;
float widthInMeters = screenSize.width / PTM_RATIO;
float heightInMeters = screenSize.height / PTM_RATIO;
b2Vec2 lowerLeftCorner = b2Vec2(0, 0);
b2Vec2 lowerRightCorner = b2Vec2(widthInMeters, 0);
b2Vec2 upperLeftCorner = b2Vec2(0, heightInMeters);
b2Vec2 upperRightCorner = b2Vec2(widthInMeters, heightInMeters);
// Define the static container body, which will provide the collisions at screen borders.
b2BodyDef screenBorderDef;
screenBorderDef.position.Set(0, 0);
b2Body* screenBorderBody = world->CreateBody(&screenBorderDef);
b2EdgeShape screenBorderShape;
// Create fixtures for the four borders (the border shape is re-used)
screenBorderShape.Set(lowerLeftCorner, lowerRightCorner);
screenBorderBody->CreateFixture(&screenBorderShape, 0);
screenBorderShape.Set(lowerRightCorner, upperRightCorner);
screenBorderBody->CreateFixture(&screenBorderShape, 0);
screenBorderShape.Set(upperRightCorner, upperLeftCorner);
screenBorderBody->CreateFixture(&screenBorderShape, 0);
screenBorderShape.Set(upperLeftCorner, lowerLeftCorner);
screenBorderBody->CreateFixture(&screenBorderShape, 0);
Alternatively, get Kobold2D Preview 5 (available by tomorrow) and have a look at the Physics-Box2D template project. I made all the necessary changes to run a Box2D v2.2.1 app, including the changes needed for the Box2D debug drawing class. I'm also hosting the latest Box2D API reference.
Use b2EdgeShape instead of b2PolygonShape, for example:
b2EdgeShape groundBox;
groundBox.Set( b2Vec2(0,0), b2Vec2(winSize.width/PTM_RATIO,0));
use
setAsBox(0,0,b2vec2(0,0),0);