A fast, easy to use, free, and community supported 2D game engine
Greg
Member
I was curious was the common approach to designing Actors (the hero, monsters, npcs, projectiles) in a game. Right now I am sub-classing CocosNode and encapsulating cpShape. My initialization method takes in a Sprite and I add it to my cpBody as well as the cpShape. From my point of view, the key bits for each actor are the Sprite, the cpBody and cpShape ... then you add behaviours and superfluous variables to track health or armor or whatever.
Any opinions on this?
Greg
Member
Here is my code example:
Hero.h:
#define kHeroCollisionType 0
@interface Hero : CocosNode {
cpShape *shape;
}
@property cpShape *shape;
// behavior methods to be added
@end
Init method in Hero.m:
- (id) initWithSprite:(Sprite *)sprite {
self = [super init];
if(self != nil) {
cpBody *body = cpBodyNew(250.0f, INFINITY);
body->p = cpv(50, 40);
body->data = sprite;
[sprite setPosition:cpv(body->p.x, body->p.y)];
shape = cpCircleShapeNew(body, 29.0f, cpvzero);
shape->e = 0.0; // elasticity
shape->u = 1.0; // friction
shape->data = sprite;
shape->collision_type = kHeroCollisionType;
}
return self;
}
Greg
Member
@jd: What do you mean automatically use the Sprite dimensions for my collision area? Do you mean calculate the vertices of the Sprite and use them to create a polygon shape?
Also, I am setting the variable "sprite" to both the cpBody and cpShape. Is that necessary? Or should I only assign it to cpBody, which then can be accessed via shape->body?
I do this. The collision area the rectangle around the sprite:
-(void) addBodyToHero: (AtlasSprite *)theSprite
{
int num = 4;
CGPoint verts[] = {
ccp(-(theSprite.contentSize.width/2),-(theSprite.contentSize.height/2)),
ccp(-(theSprite.contentSize.width/2), (theSprite.contentSize.height/2)),
ccp( (theSprite.contentSize.width/2), (theSprite.contentSize.height/2)),
ccp( (theSprite.contentSize.width/2),-(theSprite.contentSize.height/2)),
};
cpBody *body = cpBodyNew(1.0f, cpMomentForPoly(1.0f, num, verts, CGPointZero));
body->p = theSprite.position;
// You can comment the following line if you are using Actions for movements rather than chipmunk forces
//cpSpaceAddBody(space, body);
cpShape* shape = cpPolyShapeNew(body, num, verts, CGPointZero);
shape->e = 0.0f; shape->u = 0.0f;
shape->data = theSprite;
shape->collision_type = 3;
cpSpaceAddShape(space, shape);
cpSpaceAddCollisionPairFunc(space, 1, 3, &heroCollision, self);
}As sunsu mentioned, I'm fairly certain it's one or the other for each sprite. In my game, I use actions for some things and chipmunk for others. I find that actions are much easier for me to deal with for simple movements with durations, but for projectiles I use chipmunk. For collisions, it won't matter if one is handled by chipmunk and another by actions. I have to mention though, that my "eachShape" function does a check for the sprite type to handle properly. I posted some code in this thread:
http://www.cocos2d-iphone.org/forum/topic/1142
You must log in to post.
