[Fluxus] libode internal error 2

[evan.raskob [lists]]

no, scratch that - this would require major changes to fluxus,  
changing basically every datatype to double from float.

not sure how to fix this, except to create a variant using single- 
precision maths on OS X.  otherwise, any attempt to use libode will  
result in garbage numbers being passed in.

what do you guys think?


-evan


On Jan 6, 2010, at 12:56 PM, evan.raskob [lists] wrote:

> no, I see - by default, libode is double-precision on OS X.  you  
> can't just use the compiler flag to switch it to use single-precision.
>
> this is actually a pain, but fixable - all bounding boxes and  
> physics functions will need to use doubles instead of floats.  I  
> started doing this, but I'm not exactly sure of all implications,  
> because the physics bounding box relies on the Primitive- 
> >GetBoundingBox() function.
>
> I might try it out and see...
>
> cheers
> evan
>
>
>
> On Jan 6, 2010, at 12:52 PM, evan.raskob [lists] wrote:
>
>> The problem is single vs. double precision - by default it is set  
>> to double on OS X.
>> I'm working on a fix, think it's just a matter of adding compiler  
>> flag -dSINGLE
>>
>> let you know in a sec if it works
>>
>> On Jan 6, 2010, at 12:01 PM, gabor papp wrote:
>>
>>>> I can't reproduce this with ode 0.11.1 - is there a particular  
>>>> command
>>>> which crashes? I've ran self-test, collide-seq.scm and
>>>> physics-joints.scm ok.
>>> on osx i tried ode 0.11.1 with this script:
>>>
>>> (define b (build-cube))
>>> (active-box b)
>>>
>>> got the following crash:
>>>
>>> ODE Message 2: mass must be > 0 in dMassCheck() File mass.cpp  
>>> Line 49
>>>
>>> ODE Message 2: mass must be > 0 in dMassCheck() File mass.cpp  
>>> Line 49
>>>
>>> ODE Message 2: mass must be > 0 in dMassCheck() File mass.cpp  
>>> Line 49
>>>
>>> ODE INTERNAL ERROR 1: assertion "dMassCheck(mass)" failed in  
>>> dBodySetMass() [ode.cpp]
>>> Abort trap
>>>
>>> best,
>>> gabor
>>
>> Evan Raskob
>> ML Studio
>> 4-8 Arcola Street
>> London E8 2DJ
>> United Kingdom
>>
>> http://mlstudio.co.uk
>> http://pixelist.info
>>
>
> Evan Raskob
> ML Studio
> 4-8 Arcola Street
> London E8 2DJ
> United Kingdom
>
> http://mlstudio.co.uk
> http://pixelist.info
>

Evan Raskob
ML Studio
4-8 Arcola Street
London E8 2DJ
United Kingdom

http://mlstudio.co.uk
http://pixelist.info