Closed IAmJulianAcosta closed 9 years ago
Realise this is a very old issue, lost to depths of github notification hell.
For reference for anyone that might read this, the issue is that 'changing' shapes at runtime will very likely lead to breaking the behaviour of the physics since the new shapes, whenever they occupy a superset of space then their predecessor (they get bigger, or move such that the new shape is not completely inside the old shape) ultimately leads to the objects that were colliding with the old shape now being penetrated inside of the new shape which could cause the object to then be resolved badly by being pushed out in the 'wrong' direction. If in this kinect example if you were moving your arm very fast with thin sections since you're removing shapes and adding new ones instead of having shapes 'moving' then you completely negate any continuous collision detection since the 'arm' is not really moving at all, it's just teleporting from one place to another (ignoring that it's also changing it's shape) which can certainly lead to objects completely 'falling through' the arm.
Hi, I'm using nape width AIRKinect (http://www.as3nui.com/air-kinect/), and I'm trying to generate a Body from the image generated from kinect (which I've done succesfully).
I'm trying to make a game when people have to collect items over their arms, but I'm having tunneling problems, when the person raises an arm very fast, some of the objects fall.
I tried the following:
I used this example: BodyFromGraphic
Here is a sample code (without using kinect, just using a generated shape to illustrate the problem):
Am I doing anything wrong? Thanks.