Combining Functions with different return types

How in the world do I combine these functions into one? Do I make them a template? I can't find anything similar out there. Please forward any help to <removed>

Code:

---

// Converts a String to an Integer
int GetIntVal(string strConvert)
{
    int intReturn;
    intReturn = atoi(strConvert.c_str());

    return intReturn;
}

// Converts a String to a Double (Float)
double GetDoubleVal(string strConvert)
{
    double doubleReturn;
    doubleReturn = atof(strConvert.c_str());

    return doubleReturn;
}

---

Since you asked on a forum, the answers should be given on the forum, not via email. You can enable email notification if you wish to be notified via email (I believe it is called thread subscription).

As for the problem: why do you want to do that? I would suggest function overloading, but that is not possible since the function signatures are the same except for the name. You could turn them into a function template and then specialise the function template, but to call the correct version would mean having to explicitly provide the template argument.

A template version may look like this:

Code:

---

#include <string>
#include <sstream>
#include <iostream>

using namespace std;

template <typename T> T getvalue(const string & s)
{
    T v;
    stringstream str;

    str << s;
    str >> v;

    return v;
}

int main()
{
    string pi("3.141592654");

    cout << getvalue<int>(pi) << "\n"
        << getvalue<double>(pi) << endl;

    return 0;
}

---

But there are some problems with this function. It does not check for conversion errors, and I do not know how the strstream operations affect the variable's precision.

If you really want to do that, then you might as well use boost::lexical_cast.

It didn't even occur to me to use a strstream! That will work great! The precision is only there to 13 decimal places, but I don't need that kind of precision; THANKS!

Code:

---

#include <string>
#include <sstream>
#include <iostream>

using namespace std;

template <typename T> T getvalue(const string & s)
{
    T v;
    stringstream str;
   
    str << s;
   
   
   
    str >> v;

    return v;
}

int main()
{
    string pi("3.141592653589793238462643383279502884197169399375");
   
    cout.precision(40);
    cout << getvalue<int>(pi) << "\n"
        << getvalue<double>(pi) << endl;

    system("PAUSE");
    return 0;
}

---

FYI, it's not streams that have trouble with precision. It's the floating point types themselves, which represent approximations of real values. A double is not large enough to store pi to the 40th digit exactly.