cocos2d for iPhone

A fast, easy to use, free, and community supported 2D game engine

1. 

   kudely
   Member

   Hey guys. I'm using cocos2d and box2d to make a platformer. I'm currently messing around making enemies and one thing that I wanted to do was wait until they are shown on screen for them to start updating their position.

   Is there any easy way to go about detecting whether the sprite is actually shown on the screen? The camera moves around based on the player's position in game. I thought about calculating a relative position to the player, but the camera isn't always fixed exactly on the player at one specific point so that won't work.

   Are there any methods that I can call to get the current view rect based on the layer or something similar?

   Thanks.

   Posted 3 months ago #
2. 

   HiRez
   Member

   If you're not zooming, the dimensions of the view rect should be fixed, right? So just calculate what those are (you'd have to do this based on the current device) and then do your check based on offsets from the camera position. You might want to use this (or a somewhat larger rect) to decide what enemies are active anyway, to avoid running a bunch of offscreen actions.

   Or a slightly less efficient but much easier way would be to just do a simple radius check from the camera position (the "activation" radius would have to be at least 1/2 the diagonal length of the screen plus 1/2 the diagonal length of your largest sprite).

   Posted 3 months ago #
3. 

   kudely
   Member

   No, I'm not using any kind of zooms. I was messing around with the CCLayer's bounding box, trying to figure out how to solve the problem using that. I have a HALF of a solution. I'm not exactly sure why it only works half way. The way I currently have it set up, it detects correctly if the enemy is on the screen, but if I move my character past the enemy (my x coordinate is greater than his), it says the enemy is off the screen.

   I'll show you some code. The following is called within the layer's update method:

   
// Update each of the enemies
for (Enemy *enemy in enemies)
{
if ([enemy isOnScreen:self.boundingBox])
{
[enemy update];
NSLog(@"enemy on screen");
}
else
{
[enemy stopMoving];
NSLog(@"enemy off screen");
}
}


   And here's the method within the CCSprite that is being called to check whether the enemy is on the screen:
   
-(BOOL) isOnScreen:(CGRect) screenBox
{
CGPoint location = [[CCDirector sharedDirector] convertToGL: self.position];
screenBox.origin.x = (-screenBox.origin.x) + windowWidth/2;
screenBox.origin.y = (-screenBox.origin.y) + windowHeight/2;

   if (CGRectContainsPoint(screenBox, location))
return YES;
else
return NO;
}


   As you can see, I had to mess with the origin of the rect in order to get it to work at all. Like I said, it works for half of the screen, but for some reason, when the enemy enters the left half of the screen, it stops working.

   Any ideas as to what is going on? I'm truly confused by the CCLayer's bounding box coordinates. They don't really make sense to me.

   Posted 3 months ago #
4. 

   grillaface
   Member

   I think the answer is in how your camera code works.

   I've seen several posts by Riq that advise against messing with the camera position and to instead use CCFollow. Your layer will run the follow action and will target the player, and then it's the layer itself that moves.

   This way checking for the enemy's position relative to the layer's offset (now layer.position) you'll know if it's onscreen.

   Posted 3 months ago #
5. 

   grillaface
   Member

   And come to think of it - if you are moving the camera and not the layer... If your enemy is at (2000,0) and you've moved the camera +2000 in the x direction... See what I mean?

   Posted 3 months ago #
6. 

   HiRez
   Member

   The only problem with CCFollow that I see is he said "the camera isn't always fixed exactly on the player". You could maybe get around this by making a camera target dummy sprite (optionally attached to the player) that would allow some variance and using the CCFollow on that.

   A CCLayer's bounding box doesn't really have a meaning, since it's essentially an infinite plane and is only defined by the content. You can write a function that does CGRectUnion() on all the individual sprites that will give you a dynamic bounding box.

   But honestly I think it's just easier to use the radius method (compare camera location to each sprite distance) as an activation check. Sure, you'll get a few sprites active above and below the screen (or on the sides if you're in portrait mode), but it probably won't affect performance much and is very easy to implement, and fast to process per-frame.

   Posted 3 months ago #
7. 

   kudely
   Member

   @grillaface, My layer actually is moving around currently. I move it based on the character's position. I haven't messed with the actual camera at all. I guess the way I worded it in my post made that a little confusing. Sorry about that.

   @HiRez, I think I'm going to do what you mentioned. Since my character will usually be moving to the right in a sidescroller fashion, I may just check the x coordinate of each enemy, and if it's within the layer's current bounds then I'll "activate" that enemy, which will make it to where I no longer need to check if it has entered, because it will be permanently activated.

   Posted 3 months ago #
8. 

   tbone28
   Member

   Just get the nodes position converted to world

   CGPoint p = [sprite.parent convertToWorld:sprite.position];
   // compare p to [[CCDirector sharedDirector] winSize]

   I have an isOnScreen method for all my entities that I can share if you want but I am not at my computer for the next couple days.

   And it doesn't matter where I move my layer.

   Posted 3 months ago #

RSS feed for this topic

Reply

You must log in to post.