Weird template usage - any ideas?

Hi everyone,

I am having problems understanding the following piece of code. In particular, I don't understand the lines,

template<class T, T t = T()>

and,

template<bool b>

Any help would be highly appreciated.

Thanks,
Zeeshan

Code:

---

#include <iostream>

template<class T, T t = T()>
class A
{
private:
    template<bool b>
    class B
    {
    public:
        static const int m_n = b ? 1 : 0;
    };

public:
    static const int m_value = B<(t > T())>::m_n - B<(t < T())>::m_n;
};

int main()
{
    std::cout << A<int, -9>::m_value
              << A<bool, true>::m_value
              << A<char>::m_value << std::endl;

    return 0;
}

---

Basically what's happening is that T t = T() will resolve into an object of type T. This value can either be a default constructed object, or it can be supplied by you. When this is compiled it will compute the m_value.

Code:

---

template<class T, T t = T()>

---

Notice that this expression in this context is a default value.

We can figure it out by drilling down the examples.

Code:

---

A<int, -9>::m_value
//We have to substitute this: static const int m_value = B<(t > T())>::m_n - B<(t < T())>::m_n;
B<(-9 > 0)>::m_n - B<(-9 < 0)>::m_n

---

So what is B, and m_n, and what is m_value? Well, if we look:

Code:

---

template<bool b>
    class B
    {
    public:
        static const int m_n = b ? 1 : 0;
    };

---

The template parameter is just a value we use. So we just substitute again.

Code:

---

B<(-9 > 0)>::m_n - B<(-9 < 0)>::m_n
B<false>::m_n - B<true>::m_n;
0 - 1
-1

---

m_value should be -1.

The others will be left for you to figure out.

Template parameters can be types, or integral values known at compile time. The template takes a type parameter T, and some integral parameter t of type T. The parameter t will default, if needed, to the default value of type T, which is 0 for integer types and false for bool type.

Proper instantiations of the type might be A<int, 2> or A<int>. In the latter case, the second template parameter defaults to 0, the default constructed value of type int.

T can't be any type here, it must be an integral type.

Template parameters can be types, or integral values known at compile time. The template takes a type parameter T, and some integral parameter t of type T. The parameter t will default, if needed, to the default value of type T, which is 0 for integer types and false for bool type.

Proper instantiations of the type might be A<int, 2> or A<int>. In the latter case, the second template parameter defaults to 0, the default constructed value of type int.

T can't be any type here, it must be an integral type.