New collision, just for fun.

[Fuzzy]

DST and I sat up all night and talked about a new way of doing collisions. Here is the function I came up with. Hes probably using a modified version of this. I hope he will post his too.

Note that this does exactly the same thing as a wireframe or the other actors. its just a different way of doing it. Make of it what you will

Code: Select all

int BoxCollide(float ActX, float ActY, int ULX, int ULY, int LRX, int LRY){
int tmp = 0;
if((ActX > ULX) && (ActX < LRX) && (ActY > ULY) && (ActY < LRY))
{tmp = 1; }
return tmp;
};


float ActX and float ActY are the actor x and y. ULX and ULY are x and y coordinates for an invisible box(the upper left corner) LRX and LRY are the lower right corner values. tmp is a temp value to store a return number that indicates true or false.

What it does is check that player x is more than ULX, and less than LRX. then it checks the same with y values. if any is false, then the function returns zero. Otherwise, it returns 1, which means the actor is inside the invisible box.

use it like this:

Code: Select all

if (BoxCollide(player.x, player.y, -50, -50, 50, 50) == 1)
{
    do something here;
}


[Bee-Ant]

[Fuzzy]

Here is one to check if an actor is inside a circle.

Code: Select all

// A function to find whether a given point is inside a circular area
int CircleCollide(float ActX,float ActY,int r)
{
  int tmp = 0;
  if( pow(x,2)+pow(y,2)-pow(r,2) < 0){ tmp=1;}
  return tmp;
};


The variables are the same, except for r, which is a radius.

[Bee-Ant]

I was frustrated about NON-SOLID physical response...
I thought about the solution...and this is the outcome...
I dont really advanced about array thing, so I made this simple solution

[Bee-Ant]

"Just For Fun" I made "Fake Physical Response" where I dont use real physical response

[DST]

I'm still experimenting with implementing the boxcollide, the largest problem being that the only way i can think to program the ULX etc. values for a collide object is (collide.x-collide.width*.5)....and that's just for ONE UL value! happening in a repeating collision is a lot of work for the cpu. But I am sure there is a simpler way.

I think the best way to use the returned tmp variable is basically to write your own physical response; one that only affects player's y. The existing physical response in GE was designed as a general purpose response for any shape or use; and it takes angle into account and shoves the player back at whatever angle/energy. What we are trying to do is create a very simple, solid one for collisions with tops of platforms only. This response needs to do two things:

1. Counter the effect of gravity
2. Shove the player back to the top of the platform if he happened to fall partially thru it.

I am sure we can make this work!

[Bee-Ant]

1. Counter the effect of gravity
2. Shove the player back to the top of the platform if he happened to fall partially thru it.

That's what I'd tried with that Fake Physical Response, but I failed...
I hope makslane could fix the collision...
esp foor physical response to be more more more and more solid

[Fuzzy]

I gave it some thought:

how much force does it take to stop a falling box? Exactly as much force as it is exerting but in an opposite direction. Physical response handles this well by allowing us to multiply by 0.0. But it affects both x and y axis.

This only affects the event actor. For that reason I also left out mass. Platforms dont move anyway.

I was going to call it WhimsicalResponse, but thats a little too whimsical. Perhaps EgotisticalResponse would work too. Please, give it a home and a decent name.

Its going to be problematic with a directional velocity too, but we can convert that.

Code: Select all

void FuzzicalResponse(float event_mult_x, float event_mult_y, float f_velocity_x, float f_velocity_y)
{
    xvelocity = event_mult_x*f_velocity_x;
    yvelocity = event_mult_y*f_velocity_y;
}


now if you use this and jump on a platform, you will possibly continue sliding, at least until you press the opposite direction key.