I set the gravity scale property of a Box2d body. I would like to change the gravity scale when the body reaches a specific position. Can this be done? If yes, how can it be achieved.
In my case I set constant speed for falling objects like this.
#define MIN_SPEED 2.0f
-(void)update:(ccTime) dt
{
b2Vec2 vel = self.body->GetLinearVelocity();
if( ABS(vel.x) > MIN_SPEED )
{
if(vel.x>0)
vel.x = MIN_SPEED;
else
vel.x = -(MIN_SPEED);
}
if( ABS(vel.y) > MIN_SPEED )
{
if(vel.y>0)
vel.y = MIN_SPEED;
else
vel.y = -(MIN_SPEED);
}
self.body->SetLinearVelocity(vel);
}
Check the position and use SetGravityScale() :
b2Vec2 pos = body->GetPosition();
if (pos.x > minPosX && pos.x < maxPosX
&& pos.y > minPosY && pos.y < maxPosY) {
body->SetGravityScale(theScalingFactor);
}
Related
I'm making a basic DOOM recreate in Bullet and would like to know how to move in the direction my object is facing.
Currently my framework is set-up where my camera sets its rotation and position to the player, who's position and rotation is set to the physics body created by Bullet.
Here's my current movement code for the player's movement:
GameObject::Update( deltatime );
if( m_pController == 0 )
return;
vec3 dir(0, 0, 0);
vec3 rot(0, 0, 0);
if( m_pController->IsButtonPressed( PCB_Up ) )
dir.z += 1;
if( m_pController->IsButtonPressed( PCB_Down ) )
dir.z -= 1;
if( m_pController->IsButtonPressed( PCB_Left ) )
dir.x -= 1;
if( m_pController->IsButtonPressed( PCB_Right ) )
dir.x += 1;
if (m_pController->IsButtonPressed(PCB_RotLeft))
rot.y -= 1;
if (m_pController->IsButtonPressed(PCB_RotRight))
rot.y += 1;
dir.Normalize();
rot.Normalize();
float speed = 10.0f;
btVector3 force = btVector3(dir.x, dir.y, dir.z) * speed;
btVector3 rotForce = btVector3(rot.x, rot.y, rot.z) * speed;
if( m_pBody )
{
m_pBody->applyForce( force, btVector3(0,0,0) );
m_pBody->applyTorque(rotForce);
m_pBody->forceActivationState( DISABLE_DEACTIVATION );
}
I know it can be done by creating a matrix from my camera and extracting my view but the thing is my camera is being set to my player's rotation and translation.
Can the same be done by creating an SRT matrix with my physics body's values and calculating a direction vector from those numbers and then have:
if( m_pController->IsButtonPressed( PCB_Up ) )
dir += ForwardFacing;
if( m_pController->IsButtonPressed( PCB_Down ) )
dir -= ForwardFacing;
I'm not sure how it'd be done for left and right keys though.
Anyways,
thanks for your time!
Yeah so I just ended up putting the Object in an SRT matrix and extracting its looking at position then multiplying it by the dir.z and adding the getRight for turning and moving on the x axis.
mat4 mat;
mat.CreateSRT(m_Scale, m_Rotation, m_Position);
vec3 forceDir = mat.GetAt() * dir.z + mat.GetRight()*dir.x;
btVector3 force = btVector3(forceDir.x, forceDir.y, forceDir.z) * speed;
btVector3 rotForce = btVector3(rot.x, rot.y, rot.z) * speed;
I have some sprites (2d-boxes) of the same type stored in a formation vector. Now i want to move them around with the mouse, that works well. But the code should prevent the player to move one sprite onto another already existing sprite of the vector.
My solution is quite ugly and does not work. Whenever a sprite is moved around, i test with the spriteoverlap function if the sprite is moved onto another. The Problem:
Whenever the sprite is directly close to the it stops moving, which is what wanted.
But i can't move it anymore afterwards because the overlapfunction sets the bool always to false.
while (App.pollEvent(Event))
{
//Moving the playerbuttons on the formation screen
for (size_t k = 0; k < formation.size(); k++)
{
bool drag_onto_otherplayer = false;
if (isMouseOver(formation[k], App) == true)
{
//The next loop tests if the sprite being moved with the mouse overlaps with another sprite from the formation vector
for (size_t j = 0; j < formation.size(); j++)
{
if (spriteOverlap(formation[j], formation[k], App) == true && k!=j)
{
std::cout << drag_onto_otherplayer << std::endl;
drag_onto_otherplayer = true;
std::cout << drag_onto_otherplayer <<std::endl;
}
if (drag_onto_otherplayer == false)
//(If the sprite does not overlap getting the new sprite position from the mouseposition)
{
Mouseposition =sf::Vector2f(sf::Mouse::getPosition(App));
Mouseposition.x = Mouseposition.x - formation[k].getLocalBounds().width / 2;
Mouseposition.y = Mouseposition.y - formation[k].getLocalBounds().height / 2;
formation[k].setPosition(sf::Vector2f(Mouseposition));
Formation_playernames.clear();
Formation_playerinformation.clear();
Formation_Playernames(Font, Formation_playernames, formation, playerlist);
Formation_Playerinformation(Font, Formation_playerinformation, formation, playerlist);
}
So the problem are my loops and the bool test i guess, but i don't know how to solve it.
Any ideas ?
Here is my spriteoverlap function:
bool spriteOverlap(sf::Sprite &sprite1, sf::Sprite &sprite2, sf::RenderWindow &App)
{
float x_min1 = sprite1.getPosition().x;
float x_max1 = sprite1.getPosition().x + sprite1.getLocalBounds().width;
float y_min1 = sprite1.getPosition().y;
float y_max1 = sprite1.getPosition().y + sprite1.getLocalBounds().height;
float x_min2 = sprite2.getPosition().x;
float x_max2 = sprite2.getPosition().x + sprite2.getLocalBounds().width;
float y_min2 = sprite2.getPosition().y;
float y_max2 = sprite2.getPosition().y + sprite2.getLocalBounds().height;
if (x_min1 > x_max2 | x_max1 < x_min2 | y_min1 > y_max2 | y_max1 < y_max2)
return false;
else
return true;
};
And my isMouseover function:
bool isMouseOver(sf::Sprite &sprite, sf::RenderWindow &App)
{
float pos_x = sprite.getPosition().x;
float pos_y = sprite.getPosition().y;
if (sf::Mouse::getPosition(App).x > pos_x && sf::Mouse::getPosition(App).x < pos_x+sprite.getLocalBounds().width &&
sf::Mouse::getPosition(App).y >pos_y && sf::Mouse::getPosition(App).y < pos_y + sprite.getLocalBounds().height)
{
return true;
}
else
return false;
};
Check for collision is already included somewhat in sfml:
bool spriteOverlap(sf::Sprite& sprite1, sf::Sprite& sprite2) // possibly const, dunno
{
return sprite1.getGlobalBounds().intersects(sprite2.getGlobalBounds());
}
Generally try this: Only move, if the position of the next frame is valid. This prevents objects being stuck, because you already moved them into an invalid position, thus preventing any further movement
edit:
//untested
bool spritesWillOverlap(sf::Sprite& sprite1, sf::Sprite& sprite2, sf::Vector2f vel)
{
top1 = sprite1.getGobalBounds().top + vel.y;
left1 = sprite1.getGlobalBounds().left + vel.x;
right1 = left1 + sprite1.getGlobalBounds().width;
bottom1 = top1 + sprite1.getGlobalBounds().height;
top2 = sprite2.getGobalBounds().top + vel.y;
left2 = sprite2.getGlobalBounds().left + vel.x;
right2 = left2 + sprite1.getGlobalBounds().width;
bottom2 = top2 + sprite1.getGlobalBounds().height;
sf::FloatRect rect1(top1, left1, right1 - left1, bottom1 - top1);
sf::FloatRect rect2(top2, left2, right2 - left2, bottom2 - top2);
return rect1.intersects(rect2);
}
vel: velocity -> an object is moved by this 2D vector every frame
If the concept of "frames" is unfamiliar, read up on framerates/fixed framerate and/or timestep. here's an example article to get started: Fix Your Timestep!
My game world space is a rectangle of 4 screen sizes (2 rows x 2 columns )
I am moving the camera based on the movement of a ball. The ball position is set according to the physics engine.
I notice that, sometimes the ball changes direction slightly as the camera is moving to follow the ball.
Initially I thought it was sort of an illusion based on the movement of the camera, but when I draw a path, the line is of crooked at times.
I used some of the code behind the CCFollow implementation code (not CCFollow itself, but the code behind it), for following the ball
(I had issues with CCfollow)
Below is my Tick() function:
————————————————–
- (void)tick:(ccTime) dt {
bool ballMov = false;
CGPoint pos;
CCSprite *ballData;
dt = 1.0 / 60.0f;
world->Step(dt,8, 3 );
for(b2Body *b = world->GetBodyList(); b; b=b->GetNext()) {
if (b->GetUserData() != NULL) {
CCSprite *tempData = (CCSprite *)b->GetUserData();
if (tempData == ball){
ballData = tempData;
ballMov = true;
ballData.position = ccp(b->GetPosition().x * PTM_RATIO, b->GetPosition().y * PTM_RATIO);
}
}
}
//
// Move the layer so we can keep looking at the travelling ball
// Code is similar to CCFollow code with some adjustments
pos = ballPosition;
if(ballMov){
#define CLAMP(x,y,z) MIN(MAX(x,y),z)
pos = ccp(CLAMP(pos.x,leftBoundary,rightBoundary), CLAMP(pos.y,bottomBoundary,topBoundary));
#undef CLAMP
pos = ccpSub(pos, halfScreenSize );
pos.x = - pos.x ;
pos.y = - pos.y ;
//calculate delta move towards the new position
if(gunShot){
CGPoint moveVect;
CGPoint oldPos = [self position];
double dist = ccpDistance(pos, oldPos);
if (dist > 1){
moveVect = ccpMult(ccpSub(pos,oldPos),0.4); //0.05 is the smooth constant.
oldPos = ccpAdd(oldPos, moveVect);
pos = oldPos;
}
}
}
//move towards the ball
[self setPosition:(pos)];
lastPosition = pos;
ballPosition = CGPointMake([ball body]->GetPosition().x * PTM_RATIO, [ball body]->GetPosition().y * PTM_RATIO );
//if ball is out of gameworld space then reset
if (!CGRectContainsPoint ( outerBoundaryRect, ballPosition)){
ball.visible = false ;
[self recreateBody]; //Ball is out of Gameworld, recreate it at the default position
}
}
I want cocos2d to use the nodes' position as a normalized factor when coordinates are less than 1, and otherwise when coordinates are greater than 1, I want cocos2d to treat coordinates as it would normaly do
For instace position (1,1) would locate the node at the top-right corner of the screen, while (.5,.5) would locate the node on the center of the screen.
I was thinking on rewriting the code just where cocos2d pass on the coordinates to the vertex buffer like the following pseudo code
if (Position.x <= 1)
BufferPosition.x = Screen.Width * Position.x;
else
BufferPosition.x = Position.x;
if (Position.y <= 1)
BufferPosition.y = Screen.Height * Position.y;
else
BufferPosition.y = Position.y;
So my question is where would you put this code so that cocos2d works that way with no problem, or what approach would you take istead if it's different from the shown above
Update
I changed every single call/assignment of positionInPixels_ in the entire cocos2d library for positionInPixels
This means that each time, the corresponding getter and setter is caled instead of the CCNode's positionInPixels_ member variable.
Then I changed setPosition, and setPositionInPixels in CCNode.m like this
setPosition
-(void) setPosition: (CGPoint)newPosition
{
CGSize winSize = [[CCDirector sharedDirector] winSize];
position_ = newPosition;
if (position_.x <= 1.0) {
positionInPixels_.x = position_.x * winSize.width;
}
else
{
if( CC_CONTENT_SCALE_FACTOR() == 1 )
positionInPixels_.x = position_.x;
else
positionInPixels_.x = newPosition.x * CC_CONTENT_SCALE_FACTOR();
}
if (position_.y <= 1.0) {
positionInPixels_.y = position_.y * winSize.height;
}
else
{
if( CC_CONTENT_SCALE_FACTOR() == 1 )
positionInPixels_.y = position_.y;
else
positionInPixels_.y = newPosition.y * CC_CONTENT_SCALE_FACTOR();
}
isTransformDirty_ = isInverseDirty_ = YES;
#if CC_NODE_TRANSFORM_USING_AFFINE_MATRIX
isTransformGLDirty_ = YES;
#endif
}
setPositionInPixels
-(void) setPositionInPixels:(CGPoint)newPosition
{
CGSize winSize = [[CCDirector sharedDirector] winSize];
positionInPixels_ = newPosition;
if (position_.x <= 1.0) {
position_.x = positionInPixels_.x / winSize.width;
}
else
{
if( CC_CONTENT_SCALE_FACTOR() == 1 )
position_.x = positionInPixels_.x;
else
position_.x = newPosition.x * 1/CC_CONTENT_SCALE_FACTOR() / winSize.width;
}
if (position_.y <= 1.0) {
position_.y = positionInPixels_.y / winSize.height;
}
else
{
if( CC_CONTENT_SCALE_FACTOR() == 1 )
position_.y = positionInPixels_.y;
else
position_.y = newPosition.y * 1/CC_CONTENT_SCALE_FACTOR() / winSize.height;
}
isTransformDirty_ = isInverseDirty_ = YES;
#if CC_NODE_TRANSFORM_USING_AFFINE_MATRIX
isTransformGLDirty_ = YES;
#endif
}
This is working so far, even when using normalized coordinates with CCActionInterval
So you really think that this variant is the right one? To change the whole framework instead of just adding one method to the CCNode's extension? I think you will see a lot of problem with positioning very soon.
Also it is not good, that (1.f, 1.f) is always have coordinates (screenWidth, screenHeight). You will have a lot of troubles with small nodes. IMHO if you need such positioning, it is much more appropriate to connect this "factor" coordinates to the node's parent size.
Just do restore cocos2d framework and add your own extension to your project. For example,
#interface CCNode(FactorPositioning)
- (void) setPositionFactor:(CGFloat) factor;
#end
#implementation CCNode(FactorPositioning)
- (void) setPositionFactor:(CGFloat) factor
{
assert(self.parent != nil);
// do your calculations relatieve to the parent's content size
}
#end
Im having trouble getting the accelerometer to work correctly in landscape mode with cocos2d. Everything works perfectly in portrait, but cant figure out how to flip it to correctly work in landscape. here is my code:
-(void) accelerometer:(UIAccelerometer *)accelerometer didAccelerate:(UIAcceleration *)acceleration
{
float deceleration = 0.0f;
float sensativity = 15.0f;
float maxVelocity = 100;
playerVelocity.x = playerVelocity.x * deceleration + acceleration.x * sensativity;
playerVelocity.y = playerVelocity.y * deceleration + acceleration.y * sensativity;
if (playerVelocity.x > maxVelocity)
{
playerVelocity.x = maxVelocity;
}
else if (playerVelocity.x < - maxVelocity)
{
playerVelocity.x = - maxVelocity;
}
if (playerVelocity.y > maxVelocity)
{
playerVelocity.y = maxVelocity;
}
else if (playerVelocity.y < - maxVelocity)
{
playerVelocity.y = - maxVelocity;
}
}
-(void) update:(ccTime) delta
{
CGPoint pos = player.position;
pos.x += playerVelocity.x;
pos.y += playerVelocity.y;
CGSize screeSize = [[CCDirector sharedDirector]winSize];
float imageHalf = [player texture].contentSize.width * 0.5f;
float leftLimit = imageHalf;
float rightLimit = screeSize.width - imageHalf -20;
float bottomLimit = imageHalf;
float topLimit = screeSize.height - imageHalf - 20;
if (pos.y < bottomLimit) {
pos.y = bottomLimit;
}
else if (pos.y > topLimit){
pos.y = topLimit;
}
if (pos.x < leftLimit) {
pos.x = leftLimit;
}
else if (pos.x > rightLimit){
pos.x = rightLimit;
}
player.position = pos;
}
In landscape orientation, the accelerometer x and y values are flipped. Furthermore, in interface orientation landscape right, x is negative. In landscape left, y is negative:
// landscape right:
position.x = -acceleration.y
position.y = acceleration.x
// landscape left:
position.x = acceleration.y
position.y = -acceleration.x
Look at answer in this question
Best way for accelerometer usage in cocos2d game.