Re[2]: [ODE] dCylinder bugs?

[slipch]

Hello All,

Tuesday, February 18, 2003, 12:07:06 PM, you wrote:
I have already tried it.
Seems it works right.

It is a pity mistake. However it lead to wrong result in rare cases.
Because it results only in separation axis.

One may use any separation axis indeed but some of them may be useless
and it always needs to have enough exes to separate the geometries.

 Anyway it is a mistake in my code

Russ, can you fix it in CVS.
I am sorry to trouble you again.

file: dCylinder.cpp
line: about 1023
function dCollideCylS

change

sign = (dDOT14(normal,R+1) > 0) ? REAL(1.0) : REAL(-1.0);

for

sign = (dDOT14(Ax,R+1) > 0) ? REAL(1.0) : REAL(-1.0);



-- 
Best regards,
 Konstantin Slipchenko                            mailto:slipch@gsc-game.kiev.ua