I am currently trying to build a simple console calculator in C++ and was thinking about what to do when the user types 0/0. In my calculator I convert each number from the users input to a double and therefore get the division 0.0/0.0 which should result in NaN.
However when I try to output this to the console I am noticing some weird behaviors with a changed locale. With large results it is easier to read the number when there is a separator between every 3 digits so I originally changed the local to en_US, which would format me a number like 1234.56 as 1,234.56 which is indeed the wanted behavior. But with NaN as output it didn't quite work as I thought it would.
In the following sample code I use std::sqrt(-1) to get NaN as result since my compiler (MSVC) doesn't allow me to divide 0/0.0. At first I don't specify a locale (which should give me the "C" locale if I am not mistaken, at least I got the same output with nothing vs "C"). After that I changed to locale to "" and "en_US" which both gave me different output with some seemingly random seperators in between.
Code:
#include <iostream>
#include <locale>
#include <cmath>
int main()
{
//std::cout.imbue(std::locale("C"));
std::cout << std::sqrt(-1) << std::endl;
std::cout.imbue(std::locale(""));
std::cout << std::sqrt(-1) << std::endl;
std::cout.imbue(std::locale("en_US"));
std::cout << std::sqrt(-1) << std::endl;
std::cin.get();
return 0;
}
Output:
-nan(ind)
-na.n(i.nd)
-na,n(i,nd)
For my calculator not to print some weird output I could just check if the value is NaN before outputting it, but I still wonder why NaN gets represented different (with some weird characters in between) depending on my locale.
