[ODE] NOT A NUMBER / Divide by zero

DjArcas djarcas at hotmail.com
Sun May 25 04:06:02 2003


This is still happening on occasions. All the numbers related to an objects
(position, anuglar and linear velocity, are going to NOT A NUMBER)

What other causes are there for this? And is there any way I can trap them
happening with an assert?

----- Original Message -----
From: "DjArcas" <djarcas@hotmail.com>
To: "DjArcas" <djarcas@hotmail.com>; <ode@q12.org>
Sent: Saturday, May 24, 2003 7:13 PM
Subject: Re: Not a Number (was [ODE] 2d physics, and possibly unrelated
confusedness.)


> Solution:
>
> I forgot to do
>
> dBodySetMass(mBodyID,&mMass);
>
>
>
> Thus my objects had a mass of zero (?)
>
>
>
> Sounds like a missing ASSERT to me...
>
> ----- Original Message -----
> From: "DjArcas" <djarcas@hotmail.com>
> To: "DjA" <djarcas@hotmail.com>; <ode@q12.org>
> Sent: Thursday, May 22, 2003 8:08 PM
> Subject: Not a Number (was [ODE] 2d physics, and possibly unrelated
> confusedness.)
>
>
> > Watching the position and velocity of the objects - they all go to 'NOT
A
> > NUMBER' - what could be causing that? (ie is it a bug inside ODE, or am
I
> > doing something bad... I don't think I am tho!)
> >
> >
> >
> > ----- Original Message -----
> > From: "DjA" <djarcas@hotmail.com>
> > To: <ode@q12.org>
> > Sent: Wednesday, May 21, 2003 10:27 AM
> > Subject: Re: [ODE] 2d physics, and possibly unrelated confusedness.
> >
> >
> > > Ah, that worked (pretty much) excellently, thank you very much.
> > >
> > > However, it has thrown light on another problem - not sure if it's
> > related.
> > > It's certainly something to do with my creation of contact points and
> the
> > > callback procedure, I'm sure. Essentially, if I have softCFM and
> softERP,
> > > then the objects can intersect by quite some margin - one object is a
> > > powered 'spaceship', and if I try, it can completely enter any other
> > > object - if it gets far enough in, then the physics go crazy, with
> objects
> > > moving at silly speeds around (and off) the screen. (My 'world' is
> > > surrounded by planes at the screen edges, and objects just seem to get
> > > wedged in the corners there)
> > >
> > > I've tried making softer, bouncier contacts, but then a plane seems to
> act
> > > as water - an object sinks waaaay in, then accelerates and bounces
out.
> If
> > I
> > > make much harder contacts, then it seems ok - but if a small box
gently
> > > lands on top of a larger box, then the big box will wobble slightly,
> then
> > > flick the little box off at great speeds. Also, if I try hard enough,
I
> > can
> > > get it to go crazy and move at great speeds to get wedged in the
> > corners...
> > > again. :)
> > >
> > > All my objects have a mass of roughly 1 unit per unit cubed. My time
> step
> > is
> > > 1/60 (As an aside - this isn't documented what it's supposd to be - I
> just
> > > assumed that 1.0 = 1.0 seconds), and my gravity is 98.10 (otherwise
> > objects
> > > just seem to soar)
> > >
> > > Does anyone have any ideas where I might start looking to solve these
> > > issues?
> > >
> > > ----- Original Message -----
> > > From: <psero@gmx.de>
> > > To: <ode@q12.org>
> > > Sent: Saturday, May 17, 2003 7:36 PM
> > > Subject: [ODE] 2d physics
> > >
> > >
> > > > Hi--
> > > >
> > > > to simulate 2d physics I made a 'Plane2D' joint type for ODE:
> > > >
> > > > http://www.mathematik.uni-bielefeld.de/~pserocka/ODE-Plane2D/
> > > >
> > > > For usage and installation see Plane2D.README,
> > > > the code can be found in Plane2D.tar.
> > > >
> > > > Peter
> > > >
> > > > --
> > > > +++ GMX - Mail, Messaging & more  http://www.gmx.net +++
> > > > Bitte lächeln! Fotogalerie online mit GMX ohne eigene Homepage!
> > > >
> > > > _______________________________________________
> > > > ODE mailing list
> > > > ODE@q12.org
> > > > http://q12.org/mailman/listinfo/ode
> > > >
> > > _______________________________________________
> > > ODE mailing list
> > > ODE@q12.org
> > > http://q12.org/mailman/listinfo/ode
> > >
> >
>