[ODE] Box v. Trimesh, Fishy or Insane?

Flavien Brebion f.brebion at vrcontext.com
Sun Aug 24 11:39:02 2003


No. I was under the impression faststep was a version
optimize for speed. I have no performance problem right
now, just instabilities and collision problems, so i'm
just using dWorldStep. I didn't modify the stack size
either; if it crashes due to the stack size, the problem
is likely elsewhere.

I have found the cause of the crash though. As i suspected
it's related to the fact i have a car getting stuck into
the ground. This at some time probably generates some
kind of "infinite" impulse response, and on the next
frame, the position of my object is "not-a-number". I
haven't investigated more yet.

F. Brebion



-----Original Message-----
From: Shaul Kedem [mailto:shaul_kedem@yahoo.com]
Sent: dimanche 24 aout 2003 09:47
To: Flavien Brebion; ode@q12.org
Subject: RE: [ODE] Box v. Trimesh, Fishy or Insane?


Hi,
 I wrote about a problem with box vs mesh in the other
mesh test program some time ago... 
 It looks like a more singular environment to test the
bug in. simply start the program (I don't remember the
name) and drop a box (no spheres before) - the box
will drop through the mesh and off to oblivion... 
 I tested it again with a cvs snapshot two weeks ago
or so and it wasn't fixed (btw, I changed my machine
in between so I can safely say this happens in AMD 1.1
and PIV 2.4...)

 Flavien, no need to ask if you changed the stack size
to max and tried it with faststep, right ?

Shaul
--- Flavien Brebion <f.brebion@vrcontext.com> wrote:
> Just a small reply to say that i have seen the same
> behavior in
> my racing game, with my cars colliding with the
> terrain (in ODE
> terms, translated to an OBB/tri-mesh test). I'm not
> sure if it's
> a collision problem (could it be caused by the
> recent contact
> bug that was fixed?), or a response problem? In any
> case i'd love
> to know what's going on, why, and how to get it
> fixed.
> 
> I also sometimes gets some crashes in ODE, with a
> "full stack"
> error. The timestep is non-null and everything seems
> to be working
> during many minutes, but rarely it crashes and i
> have no idea why.
> However when it happens, invariably it was when one
> of my OBBs was
> stuck into a tri-mesh. Maybe both problems are
> linked. Did anybody
> notice such behavior or should i more likely be
> looking for a bug
> in my code?
> 
> F. Brebion
> 
> 
> -----Original Message-----
> From: ode-admin@q12.org [mailto:ode-admin@q12.org]On
> Behalf Of Panayoti
> Sent: jeudi 21 aout 2003 23:40
> To: ode@q12.org
> Subject: [ODE] Box v. Trimesh, Fishy or Insane?
> 
> 
> Is it me or is the box v. trimesh results sometimes
> a little fishy?
> 
> I ran the moving trimesh test, and I notice that the
> sphere v. trimesh
> is rock-solid, but when I start throwing boxes in
> there they start to
> float, get stuck in the bunny and all kinds of weird
> things. Have I done
> something wrong or is this currently a "problem"?
> 
> Thanks for the reply,
> 
> Panayoti
> 
> 
> _______________________________________________
> ODE mailing list
> ODE@q12.org
> http://q12.org/mailman/listinfo/ode
> _______________________________________________
> ODE mailing list
> ODE@q12.org
> http://q12.org/mailman/listinfo/ode


__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com