Object turns itself into another type?

Hello!

I have a contairer class Number which is going to be able to contain a few different number types, including an integer class and a fraction class. The thing is that, if a fraction suddenly turns out to have an integer value, it is going to convert itself from a private function, to an integer object. What is the simplest method to achieve that? The number class uses a boost::variant to store the actual number.

Why not have a variable describing the "representation", say an enum with the values integer, fraction, etc.

--
Mats

Why not use boost::variant? So, what I really need to know, how can the data type turn the value of the container (which would be itself) into another number type?

I would recommend not transforming it into an integer, but instead Having the fraction class print like an integer when it simplifies to an integer. If you have two classes that change into each other so easily then they should probably just be one class. After all, the interface to fractions and integers should be the same, otherwise you may get surprises when you apply a function to two values and expect a fraction result, but get an integer instead.

There is no need to morph between integers and fractions (more precisely, "rational numbers": values that can be represented as a ratio of two integers). An integer is simply a fraction that has denominator equal to one.

If you must, simply support operations that allow conversion between integer and fraction (converting a integer to a fraction will produce a fraction with denominator equal to one, converting a fraction to an integer will convert with rounding as required).

Quote:

---

Originally Posted by King Mir

I would recommend not transforming it into an integer, but instead Having the fraction class print like an integer when it simplifies to an integer. If you have two classes that change into each other so easily then they should probably just be one class. After all, the interface to fractions and integers should be the same, otherwise you may get surprises when you apply a function to two values and expect a fraction result, but get an integer instead.

---

Yeah exactly. Like:

Code:

---

fraction foo(1, 2) // one-half
fraction bar = (foo+foo) / 3;
// Now bar = 0, not 1/3

---