bug with 2d box collision

Physics
#1

Hello,

have you ever seen a bug when a dynamic ball object collide with the extremity edge of a kinematic rectangle.

I speak about a brick breaker game, the ball touch the horizontal bar on the corner, and sometimes there are classical collision, sometimes there are a bug (  and statics objects like walls disappear ).

And the log error print no error !

Is someone have an idea about that ?

The ball has 8 radius and the rectangle bar has 100 width + 14 height, so the tickness is not the problem.

thanks

#2

When I print the x coordonate of the hidden wall, I see that the wall is not removed but print 

wall.x = 1.#QNAN

log error show nothing.

I think it is an inside bug on corona fondation. 

edit : here is 2 videos

https://youtu.be/97O25OB2_fU

https://youtu.be/rTjPW0eYQ8k

#3

ok guys, I understand.

It’s because I used the x coordonnate of the object involved during the collision that there are bugs.

#4

Ok I finally find what was erroneous in the code.

I also thinking that the docs speaks about changing the value of ball.x during the collision, but reading the value is allowed and does no disturbances.

In fact during my calculation I done one moment a wrong calculation that replace vx by an invalid number, something like √(-230) !

Thanks and Ciao 

Ok I finally find what was erroneous in the code.

I also thinking that the docs speaks about changing the value of ball.x during the collision, but reading the value is allowed and does no disturbances.

In fact during my calculation I done one moment a wrong calculation that replace vx by an invalid number, something like √(-230) !

Thanks and Ciao 

#5

When I print the x coordonate of the hidden wall, I see that the wall is not removed but print 

wall.x = 1.#QNAN

log error show nothing.

I think it is an inside bug on corona fondation. 

edit : here is 2 videos

https://youtu.be/97O25OB2_fU

https://youtu.be/rTjPW0eYQ8k

#6

ok guys, I understand.

It’s because I used the x coordonnate of the object involved during the collision that there are bugs.

#7

Ok I finally find what was erroneous in the code.

I also thinking that the docs speaks about changing the value of ball.x during the collision, but reading the value is allowed and does no disturbances.

In fact during my calculation I done one moment a wrong calculation that replace vx by an invalid number, something like √(-230) !

Thanks and Ciao 

Ok I finally find what was erroneous in the code.

I also thinking that the docs speaks about changing the value of ball.x during the collision, but reading the value is allowed and does no disturbances.

In fact during my calculation I done one moment a wrong calculation that replace vx by an invalid number, something like √(-230) !

Thanks and Ciao