// For high precision arithmetic, we have to be
// more relaxed then using ==
// Because 2/sqrt( (0,0)<->(2,2) ) == 1/sqrt( (0,0)<->(1,1) )
- // is not always true (at least, it is not for ttmath)
+ // is not always true (at least, not for some user defined types)
return math::equals_with_epsilon(dx, other.dx)
&& math::equals_with_epsilon(dy, other.dy);
}