Setting type restrictions/Error handling

[trikker]

So I am finishing the chapter on errors in my book and it's taking me back to a program I wrote for another exercise and telling me to modify the program to write out an error if the result cannot be represented as an int.

Currently the code is this:

// Celsius to Fahrenheit and Fahrenheit to Celsius
#include "std_lib_facilities.h"

double ctof (double orig_c)
{
       double f = ((orig_c*9.0)/5.0)+32.0;
       return f;
}

double ftoc (double orig_f)
{
       double c = ((orig_f-32.0)/9.0)*5.0;
       return c;
}

int main()
{
    double orig_c = 0;
    double orig_f = 0;
    cout << "Input Celsius, and then fahrenheit.\n";
    cin >> orig_c >> orig_f;
    double f = ctof(orig_c);
    double c = ftoc(orig_f);
    cout << orig_c << " Celsius equals " << f << " Fahrenheit.\n";
    cout << orig_f << " Fahrenheit equals " << c << " Celsius.\n";
    keep_window_open();
}

Obviously the first step I need to take is to set "f" and "c" as type int, but after that when dealing with errors, I'm not sure how to put into code what I want to happen. I mean I want to say something along the lines of

// Celsius to Fahrenheit and Fahrenheit to Celsius
#include "std_lib_facilities.h"

class Bad_type{}        // type specifically for reporting errors from ctof and ftoc

double ctof (double orig_c)
{
       int f = ((orig_c*9)/5)+32;   // took out the .0's (they were there in first place because
                                                 // without them for some reason the answer was construed
       if (f ! int) throw Bad_type();    // obviously the if (f ! int) won't work but that's the question
       else
       return f;
}

double ftoc (double orig_f)
{
       int c = ((orig_f-32)/9)*5;             // same comments as above
       if (c ! int) throw Bad_type();
       else
       return c;
}

int main()
try{
    double orig_c = 0;
    double orig_f = 0;
    cout << "Input Celsius, and then fahrenheit.\n";
    cin >> orig_c >> orig_f;
    double f = ctof(orig_c);
    double c = ftoc(orig_f);
    cout << orig_c << " Celsius equals " << f << " Fahrenheit.\n";
    cout << orig_f << " Fahrenheit equals " << c << " Celsius.\n";
    keep_window_open();
}
catch (Bad_type){     // catches Bad_type and reports error
    cout << "Result cannot be represented as an integer."
}

Any ideas how I can specify the type restriction? Thanks in advance. And sorry if the code is messy I actually edited the second set of code in this post window.

[JohnW@Wessex]

You could maybe try comparing the double result to double(INT_MAX)

double f = ((orig_c*9.0)/5.0)+32.0;
                             
if (f > double(INT_MAX)) throw Bad_type();
else
    return int(f);

Alternatively, you could equate the result to a type that is guaranteed to be larger than an int and compare with INT_MAX

_int64 f = ((orig_c*9.0)/5.0)+32.0;
                             
if (f > INT_MAX) throw Bad_type();
else
    return int(f);