CS6 - Actionscript 3.0 Collisions, points and the lack of a bump - actionscript-3

Question

I've been toying with flash for a bit, I usually solve my problems fairly quick, however I have been stuck on this one for a few days, google hasn't been useful, and I do not have friends that know much about flash.
So, I've been working on making a basic 2D platform. I'm having problems figuring out how to implement the points on the character and bg.hitTestPoint. As it stands, before I worked the points and everything in, it was working fine. I have a trace in, to let me know when it considers it a collision with the object (In this case a wall).
The character's dimensions are:
Top Center: (42,0)
Bot Center: (42,112)
R. Center: (84,56)
L. Center: (0,56)
Total Image: 84 x 112 (px)
The character's location on the screen:
Top Center: (46, 460)
Bot Center: (46, 379)
L. Center: (20, 440)
R. Center: (80,420)
Here are the blocks:
The bump/collision variables:
var leftB:Boolean = false;
var rightB:Boolean = false;
var topB:Boolean = false;
var botB:Boolean = false;
var topBPt:Point = new Point(42, 0);
var botBPt:Point = new Point(42, 112);
var leftBPt:Point = new Point(0, 56);
var rightBPt:Point = new Point(84, 56);
The hitTestPoint block: (To save space, I'll just do the top one.)
if(bg.hitTestPoint(fag.x + topBPt.x, fag.y + topBPt.y, true))
{
trace("Top Bump works.");
topB = true;
}
else
{
topB = false;
}
if(topB)
{
if(ySpeed < 0)
{
ySpeed *= -0.5;
}
}
I tried to use different dimensions, there is a small chance I may just not have understood the full explanation on the site I was reading.
Also, I have a feeling I can work around it by just making it a hitTestObject, and manually making each part of the wall different...But I don't see that as efficient or reasonable. Thanks for any help.

Answer 1

The math for finding out the collision for rectangular areas is generally the same.
First of all after glancing your code, define your screen objects as objects... I mean sublass Sprite class and add properties to it, do not do it via globals...
Like:
public class myScreenObjs extends Sprite {
// Add the tricky cool specialities and properties here
}
Normally a sprite exposes its x and y coordinates. If dynamically created x and y are at top left of the sprite, if via cs6 you better take the o with cross from center to top left for having a common ground of reference. Centered x y references are good for 3d specially when rotation needed.
A sprite has width and height properties also.
Lets assume you have sprites c (character) and b (block to bump)
Given:
left is left x of an object.
right is right x of an object.
top is top y of an object.
bottom is bottom y of an object.
I know this is a not very optimal function but I am writing it for demonstration purpose.
function bumpTest(c: Sprite, b: Sprite):Object {
var
r: * = {},
cLeft : int = c.x, // doing this for demonstrating you can directly use c.x
cRight : int = c.x + c.width,
cTop : int = c.y, // same
cBottom : int = c.y + c.height,
bLeft : int = b.x, // same
bRight : int = b.x + b.width,
bTop : int = b.y, // same
bBottom : int = b.y + b.height;
r.left = Boolean ((cLeft >= bLeft) && (cLeft <= bRight));
r.right = Boolean ((cRight >= bLeft) && (cRight <= bRight));
r.top = Boolean ((cTop >= bTop) && (cTop <= bBottom));
r.bottom = Boolean ((cBottom >= bTop) && (cBottom <= bBottom));
r.bump = Boolean ((r.left || r.right)&&(r.top || r.bottom));
return(r);
}
Now r has the complete answer, If r.bump is false the other values in r are irrelevant.
If r.bump turns to be true, then you can check via r.left etc from where c bumps into b. since:
r.left true means c left is into or at boundary of b.
r.right true means c right is into or at boundary of b.
r.top true means c top is into or at boundary of b.
r.bottom true means c bottom is into or at boundary of b.