3 Bugs! Thx VolatileZapper!

Hi all,

found 3 new bugs, 2 of them are due to boundingVolumes, and another due to the particle system.

In BoundingSphere() sometimes collision doesn’t occur when obviously the bullet is going straight through the alien

In BoundingBox(), during its initialisation, new Vector3f(0, 0, 0) is setup making the box at 0, 0, 0, making a collision between the alien and ship because they too are at 0,0,0.

And when you move the particle system, the whole thing speeds up!

need more info on how you are setting up bounds. Bounding boxes are setup at 0,0,0 which is supposed to be relative to the geometry’s rotation, translation, etc.

Move the particle system? explain.


One of my tasks this week was to double check the accuracy of picking and collision vs. bounding spheres and boxes. I’ve done so, using TestPick and TestCollision and turning the drawing of Bounds on and (in the case of collision) slowing things done to watch the action frame by frame from multiple angles.

The two tests checked out mathematically correct. As long as you have the latest from cvs and are correctly using bounds, all should be well.

If you have the latest from cvs, and are definitely not using both BoundingBox and BoundingSphere, try drawing the bounds with drawBounds(rootnode). Let me know if the problem remains.

PS: I’ll be working on allowing the merging of box and sphere bounds soon… Any ideas on which should win out in a merge? ie, should bsphere + bbox = bbox or bsphere? Or perhaps use the hierarchy structure to determine this.

See here: