[ODE] On the meaning of ODE stability (was some quick notes on 0; 6) ...

[Jon Watte (ODE)]

Geoff Carlton wrote:
> Jon Watte (ODE) wrote:
>   
>> with the app, not in the system). FWIW: I, too, have some modifications 
>> to the ODE library I use. Some of which have been submitted as patches, 
>> and some of which have been rejected.
>>
>>   
>>     
> I'm curious.  What modifications were rejected?
>   
For example, I fixed a bug where some geoms could push themselves 
straight through a box with a bit of persistence, because of a denormal 
division issue. My fix was rejected because it hard-coded epsilons 
(based on dSINGLE or dDOUBLE) instead of using some systemwide define 
(which nobody could ever actually show existed, because I needed "square 
root of denormal epsilon").

Cheers,

          / h+