[ODE] More news on the crash problem

[Alex Hetu]

I previously found another way of crashing the tri-collider. It implied
colliding a sphere with a mesh containing only one triangle in it. The crash
would actually occur in opcode, but only because primitive tests were not
enabled for the sphere collider in the dxTriMesh constructor. I replaced the
line "_SphereCollide.SetPrimitive( false )" with
"_SphereCollider.SetPrimitive( true )" and get no more crashes.

I'm not getting any crashes, but the boxes are indeed messed up. They
collide with the trimesh, but often get glued to them, as if the collider
doesn't try to push them away from each other. I don't know how ODE solves
this, i'm really not an expert, but to me it looks like it's having big
problems with boxes VS mesh inter-penetration. Spheres work 100% fine
though.

Alex Hetu

-----Original Message-----
From: ode-admin@q12.org [mailto:ode-admin@q12.org]On Behalf Of Erwin de
Vries
Sent: Thursday, August 28, 2003 12:07 PM
To: ode@q12.org
Subject: Re: [ODE] More news on the crash problem


> And here.. you can see two contact points with..
> guess what? an invalid NaN normal.
>
> I have honnestly no idea how the collision functions can
> generate a contact point with invalid normal...?

This looks interesting. If i see it correctly there is absolutely no
geometry near the generated contacts, right?

> Does anybody have an idea? I don't want to loose my time
> debugging code that isn't even mine.. is it a known bug,
> or should i investigate further?
>
> Note that the bug happens "rarely", but since i logged the
> exact position, velocities, orientations when this happens,
> i'm now able to replicate it 100% of the time with a single
> debug key.

I'm not sure what the problem might be. If you're able to replicate it
instantly it should be quite easy to find out what line in the collision
function causes the wrong contacts.

> Any suggestion is welcome.

Well. i started working on a replacement for the current box-collision
function today, but i dont know when i will release it. Might be soon, might
not be so soon.

Erwin

_______________________________________________
ODE mailing list
ODE@q12.org
http://q12.org/mailman/listinfo/ode