[ODE] RE: Torque: Bug in 0.025, apparently fixed in CVS

[Ruud van Gaal]

> Hello Ruud and the list -
>
> I am convinced. There WAS a bug in ode-0.025. It is fixed in the current CVS
> version. Russ, maybe you could look at this and make sure that the bug has
> been
> intentionally and completely fixed, and that this is not just a lucky
> conicidence?

Phew, thanks Norman, for confirming and testing this. I'll sleep better 
tonight. ;-)
A little note than to Russ; I do think that multiplying with R is the right 
thing, as is in the code. But probably somewhere else, where the torque is 
applied, a bad tranposed matrix order is used instead of the right one. Perhaps 
in one of the many dot product macros/inlines.

> Now that I think about it, I did have some strange torque reversal behavior,
> but I attributed it to an orientation error in my level editor. It looks like
> this was exactly the problem that I had seen and written off. Good that you
> pointed this out, Ruud.

Well, it isn't that I was so brilliant to find this out; the code displayed the 
problem by itself (and it worked before introducing ODE) and cars are quite 
sensitive so you'll quickly see the odd behavior (it had the tendency to roll 
jitteringly by itself a little, barely damped by the suspension).

I hope that Russ can find out what has changed, so this bug won't creep up 
again when some more code changes.

Ruud