The old OrientedBB Problem

Hi,



Comrade works at the Time with the OrientedBoundingBox. In 9 of 10 Situation it worked well, but someTimes not.



Now we changed



 158     rotate.mult(xAxis, toReturn.xAxis);
 159     rotate.mult(yAxis, toReturn.yAxis);
 160     rotate.mult(zAxis, toReturn.zAxis);



to


 158     rotate.mult(Vector3f.UNIT_X, toReturn.xAxis);
 159     rotate.mult(Vector3f.UNIT_Y, toReturn.yAxis);
 160     rotate.mult(Vector3f.UNIT_Z, toReturn.zAxis);



and all seems fine, can somebody have a look on this part?
is this the Solution?

Greetings, HellG (and th guy who found it out)

It is interesting. I have some problems with bounding volumes too, especially BoundingBox and OrientedBoundingBox. Please post a patch in the section "Contribution Depot". We need to write some unit tests to check that was not working and what is working now. Thank you comrade  :smiley:

Ok, I'm Sorry ,



We have seen the TransformMethod as absolut calculation, but its build for transforming LocalBounds to worldBounds  .



Perhabs OBB with OBB CollisionDetection works better than with AABB. The most Problem we had, came from all static temporary Values in BoundingVolume an its Extentions. But we used it Multithreaded. When all Detections run in the GL Thread, it should be no problem

You can remove the static identifier of these variables as it is useless and dangerous.