[ODE] Is it time to accept dcyl2 contribution?

Bram Stolk bram at sara.nl
Fri Feb 25 15:39:59 MST 2005


On 02/25/05 16:24:19, Megan Fox wrote:
> Doesn't the dCylinder2 have trouble with falling through solid
> surfaces when it meets them plane-on with its top or bottom?
> 
> (that is - when the flat top or bottom of the dCylinder2 is co-planar
> with the colliding surface, doesn't it have issues with falling
> through the colliding surface)

If it has, I havent noticed it, and it must be 100% co-planar and
highly sensitive because:

I use cyls as car wheels on a mesh. This works just fine, even
if the car is lying on its side (on the wheels), and those wheels
are coplanar with the mesh under it.


> 
> I also thought it was missing a collider for one or more of the other
> geoms?


It has box, sphere and mesh.

Most non-toy car-applications are likely to model the wheels as
cyls on a generic triangle mesh. The code does this flawlessly.

I rewrote test_buggy in the following way:
(1) Buggy has 4 wheels, not 3.
(2) Buggy is RWD (rear-wheel-drive).
(3) Wheels have cylinder geom, not sphere geom for true WYSIWYI
    (What you see is what you intersect :-)
(4) Wheels now show slip.
(5) World is generic triangle mesh, not a simple box/plane combo.
(6) Steering is improved.
It would be a useful ode demo if ode adopted cyl2. As it is now,
newbies have a rough time making their own car demo, especially
if they are not on windows, and cannot try carworld.

  bram


> -Megan Fox
> 
> > Hello,
> >
> > In the cvs contrib dir, there is a contribution by Nguyen Binh
> > for non-capped regular cylinders, called dCylinder2
> >
> > I've been using them with ODE, and I've found them to be stable
> > and usable.
> >
> > Currently, adding dCyl2 to ODE is an akward procedure: you
> > have to follow a lot of steps to get it into ODE. The later
> > supplied patch file from Adam is not complete, as it fails to
> > do addCollider for meshes, and does not patch the Makefile.
> >
> > Instead of fixing this patchfile in the contrib dir, I propose to
> > adopt dCyl2 into ODE proper. (to the CVS tree at least), so
> > that it can be further tested, and then added to releases.
> >
> > So, what do the maintainers think of this?
> >
> > Yours Sincerely,
> >
> >   Bram Stolk
> >
> > _______________________________________________
> > ODE mailing list
> > ODE at q12.org
> > http://q12.org/mailman/listinfo/ode
> >
> 





More information about the ODE mailing list