NewtonUserJointAddLinearRow

[SilentAssassin21]

suppose i have 2 pivots p1 and p2.
if i set the dir vector: dir = (p2-p1).normalise();

if I supply this vector to the NewtonUserJointAddLinearRow, what happens?? i have 2 options according to http://newtondynamics.com/wiki/index.ph ... dLinearRow .. the p1 and p2 stick together... or they don't

[JoeJ]

bove points will move along the 'dir' vector, until they meet at a plane which have the normal 'dir', and a position somewhere between p1 and p2.
That's why you need 3 Rows with orthonormal directions to do a point to point constraint.

[SilentAssassin21]

bove points will move along the 'dir' vector, until they meet at a plane which have the normal 'dir', and a position somewhere between p1 and p2.
That's why you need 3 Rows with orthonormal directions to do a point to point constraint.

thats why i need 3 rows ? i mean , this is what i want , for these 2 points to meet

[JoeJ]

Yes, if you have 3 rows with 3 orthonormal directions, p1 and p2 will meet at the intersection point of those 3 planes.
If you have only one row, p1 and p2 will not meet at a point, only at a single plane - guite unpredictable - like a ball that you can move along the floor plane, but you can't lift it up.
If you have only two rows with directions like x and y axis of parent body, the child body is free to move at the infinite parent z axis line.