[ODE] ODE world stepping + TOI based collision detection problem

[Roland Plüss]

> It is fairly simple to integrate time-of-impact with ODE that brute-force 
> steps to smallest time of impact. I did this a few years ago, and I will 
> have to search for the small change. I will find the patch and give you an 
> update.
> A more suble approach for continuous physics would require changes in island 
> management. This is one of the reasons why I started the Bullet Physics 
> Library. Its CCD is work in progress.
>   
hehe, so you are the guy behind this. just dropped over the project page
a couple of days ago trying to get an overview of physics libs around at
the time beeing.
> With a very minor change in code, you can let ODE integrate the position 
> based on a fraction of the timestep (deltatime),
> where the fraction is in interval [0..1]. Default is 1, which means it steps 
> the full deltatime. This might be enough for you.
>   
that's kinda what i had in mind though it requires back-stepping ( as
ODE doesn't allow to 'think ahead' or i could not manage to force it to
do so ). hence i would need to step by delta 1 and if things hit i need
to step back. doing so though results in a dead-loop that never ends.
could not figure out so far what's the problem but not using ODE ( but
TOI ) results in no dead-loop. :/

-- 

Leader and Head Programmer
- Game: Epsylon ( http://epsylon.rptd.dnsalias.net/ )
- Game Engine: Drag(en)gine (
http://epsylon.rptd.dnsalias.net/dragengine.php )
- Normal Map Generator: DENormGen (
http://epsylon.rptd.dnsalias.net/denormgen.php )


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
Url : http://q12.org/pipermail/ode/attachments/20060824/5029f008/signature-0001.bin