[ODE] dCylinder with new collision detection system

[Olivier Michel]

slipch wrote:

>OM> Thanks for your fast reply. Do you mean that the zip attachment you sent 
>OM> should work with the new version of the collision detection ? If not 
>OM> what should changed so that it works ? What are those problems with Cyl 
>OM> Cyl collision ? May I help getting them fixed ?
>
>Yes it should work with new version of collision detection. At least
>it works in my project.
> As to problems with cyl-cyl collision they are contact points not
> always are in the cylinders intersection volume. There is no visible
> effect on simulation (as seems to me) except maybe dancing in stack
> (there is other reasons  for the dancing indeed) but it is an IMO
> error.
> This problem appears when edge circles touches. To solve it I need to
> find a "good" axis for separation. The best one mast be line which
> connect the nearest circle points - but it needs too much
> computation. I have another solution - find mean points on the part
> of circles in cylinders volumes it also seems to be a "good" axis.
> Currently I use a trick which allows to set contact points near the
> place it should be.
>
Ok, I am just facing this problem. Two cylinders are interpenetrating 
each other, collision is detected too late...
It's a bit annoying. Do you have some schedule to fix this issue ? If 
not, could you give me more details about this mean points technique you 
are thinking about ? I am not sure I could help doing this, but I may try...

Thanks,

-Olivier