[ODE] ODE Message 2: Vector has zero size and NaN values.

Gonçalo Lopes goncaloclopes at gmail.com
Fri Jun 22 11:50:33 MST 2007


Hi all,

Thanks for your answers but in this case it was my own lame fault. A
division by zero was creeping up in some obscure multiplications, due
to rounding errors.

I'm sorry for bothering you, but thanks for the suggestions anyway,
I'll be on the lookout for them as well.

Sorry for the trouble,

Gonçalo


On 6/22/07, Remi Ricard <remi.ricard at simlog.com> wrote:
> Hi,
>
> >
> > I'm currently working in an interactive application with physics and
> > I've been getting this strange warning message:
> >
> > ODE Message 2: Vector has zero size
> >
> > coming from dNormalize4() and shortly after this, ODE starts spitting
> > out invalid floating point values (NaNs) for body positions inside the
> > simulation.
> >
> > Any clue has to what can usually cause this error? I've checked
> > collision functions but everything's quite simple and no strange
> > values pop up.
> >
> > I'm hammering in positions for dynamic collision bodies every few
> > frames so I thought this might influence the contacts, but I really
> > have no clue has to what I should be looking for.
>
>
> I was getting this problem when 2 collisions were occuring for the same
> bodies at the "same" position.
>
> Ex:
> Geom_1_0 of Body_1 collides with Geom2 of Body_2 at pos= 1,2,3
> Geom_1_1 of Body_1 collides with Geom2 of Body_2 at pos=1,2,3.
>
> This is possible if Geom_1_0 and Geom_1_1 share a common vertex and this
> vertex collides with Geom2
>
>
> Remi
>



More information about the ODE mailing list