[Fluxus] physics problems again

[evan.raskob [lists]]

I've noticed crashes as well on OS X.
I think double-precision is bad on OS X with single-precision fluxus.
Not sure what to do about this, currently, except recompile ode as  
single-precision?



On Mar 2, 2010, at 9:42 AM, gabor papp wrote:

> hi all,
>
> i'm having problems with the physics system again.
>
> in fedora 12 even the simplest physics script with collisions works  
> strange. if collision occurs the objects disappear. two colliding  
> cubes or a cube with a ground plane show the same behaviour.
>
> (clear)
> (ground-plane #(0 1 0) 0)
> (active-box (build-cube))
>
> another problem which seems to appear when ode is compiled with  
> double precision (tested in osx). i have this simple script which  
> explodes an object:
>
> (clear)
> (hint-none)
> (hint-wire)
> (backfacecull 0)
> (collisions 1)
>
> (gravity #(0 0 0))
> (let ([s (build-sphere 10 10)])
>        (with-primitive s
>            (for ([i (in-range (quotient (pdata-size) 3))])
>                (let* ([t (build-polygons 3 'triangle-list)]
>                        [p0 (pdata-ref "p" (* i 3))]
>                        [p1 (pdata-ref "p" (+ (* i 3) 1))]
>                        [p2 (pdata-ref "p" (+ (* i 3) 2))])
>                    (with-primitive t
>                        (pdata-set "p" 0 p0)
>                        (pdata-set "p" 1 p1)
>                        (pdata-set "p" 2 p2))
>                    (active-box t))))
>        (destroy s))
>
> with float precision it works, but with double precision it crashes  
> fluxus. could someone test if it works please? do you have any  
> suggestions how to track down these problems?
>
> best,
> gabor
>
>