Question

#include <iostream>
#include <vector>
using namespace std;

template <typename nameOfTheVariableTypeA, 
      typename nameOfTheVariableTypeB> nameOfTheVariableTypeB functionX 
                                               (nameOfTheVariableTypeA argA,
                                               (nameOfTheVariableTypeB argB)
{
    nameOfTheVariableTypeA tempArgA;
    nameOfTheVariableTypeB tempArgB;

    argA.push_back(22);

    tempArgA = argA;
    cout << "\ntempArgA: " << tempArgA[0] << "\n";

    tempArgB = argB;
    cout << "\ntempArgB: " << tempArgB << "\n";

    return tempArgB;
}

int main ()
{
    functionX (12, 12.4567);

    vector <int> f;
    functionX (f, 12.4567);

    return 0;
}

From a template book:
An attempt to instantiate a template for a type that doesn't support all the operations used within it will result in a compile-time error.

The error I received for the above code is:

1. error: request for member ‘push_back’ in ‘argA’, which is of non-class type ‘int’

2. error: subscripted value is neither array nor pointer

What’s the point that I am missing?

Answer 1

First off, there’s a superfluous parenthesis:

template <typename nameOfTheVariableTypeA,  
      typename nameOfTheVariableTypeB> nameOfTheVariableTypeB functionX  
                                               (nameOfTheVariableTypeA argA, 
                                              /*(*/nameOfTheVariableTypeB argB) 
// Note: extra parenthesis not needed here -----^ 
{

Assuming that’s just a typo, let’s take a look the first call to functionX():

functionX(12, 12.4567);  

Now, template functions require that all template parameters be specified before you can call them. But in certain cases the compiler can deduce the required type for nameOfTheVariableTypeA and nameOfTheVariableTypeB for the function call to work.

In this case, 12 is an integer literal, so it has type int. 12.4567 is a floating-point literal, so it has type double. Thus, within functionX(), nameOfTheVariableTypeA is of type int and nameOfTheVariableTypeB is of type double.

Now that all the template parameters have been specified (in this case, deduced), the compiler can instantiate the template function. That is, the compiler creates a function that looks like this:

// Hypothetical function generated by the compiler
double functionX_int_double(int argA, double argB)  
{  
    int tempArgA;  
    double tempArgB;  
    // ...

It’s as though the compiler simply substituted nameOfTheVariableTypeA and nameOfTheVariableTypeB with the deduced types. Clearly, the argument argA and variable tempArgA has type int. You get the first error because int doesn’t have a class member function called push_back(). It’s the same reason why this won’t work:

int i = 20;
i.push_back(22);

You also get the second error because the subscript operator [] is not defined for ints. Again, it’s the same reason why this won’t work:

int j = 21;
cout << j[0];

Note that this kind of information can be gleaned from the compiler errors themselves. Make sure you read them!