[ODE] float/double error in common.h

[Russ Smith]

fixed.

thanks,
  russ.


> In common.h, the results of the new dFMod() and dCopySign() 
> macros are getting cast to float even in the dDOUBLE case.
> This makes my double-precision test suite very angry. :)


-- 
Russell Smith
http://www.q12.org