[ODE] new ODE 0.8 heightfield seems working wrong

[tuan kuranes]

I'm patch author.

The patch was mainly fixing a problem with object being stuck in height 
field at first (due to use of wrong normals for collision) and then 
ended up in a rewrite + speed.
> The author of the patch Tuan Kuranes might be a better person to ask
There is surely a way to debug that very easily and fast, but as my time 
is rather limited, the very best option is really to send/give me access 
to a sample application that demonstrates the bug.
Bug could be dependent on many things, including height field data, and 
reproducing the bug will surely take a lot more time than fixing it.
For instance, reproducing it just modifying the ode_height field test 
application would be great, as well as useful to serve as test for 
future improvements.

>  - or, alternatively there is always the option of debugging what the cause of the problem is and reporting back?
>   
That would indeed be the best way. Placing some breakpoint in height 
field.cpp should get you fixed very fast about what's happening.

Tuan Kuranes.