Currently the library uses the delaunator::Point struct to represent points in the algorithm. This is not ideal, as some people might want to use their own custom types, and would prefer to not have to convert the input to one of the three currently accepted types. One way to do this might be to implement Point as trait that requires implementation of two methods: x(&self) and y(&self), as other methods are already internally implemented.
Currently the library uses the
delaunator::Point
struct to represent points in the algorithm. This is not ideal, as some people might want to use their own custom types, and would prefer to not have to convert the input to one of the three currently accepted types. One way to do this might be to implement Point as trait that requires implementation of two methods:x(&self)
andy(&self)
, as other methods are already internally implemented.