Yesterday I had my first Advenced Programming lab session. Part of the instructions get you to investigate how small a number has to be in order to throw a division by zero exception. I discovered, however, that if you use a double, it does not, in fact, throw any such error. I did some investigation, and thought I might share what I learnt here.

According to Google, C++ only throws a division by zero error if it is doing integer math. It doesn't seem to care if it is doing floating point math. This is because floating point numbers can hold the value Infinity, and integers can't. This means that none of the following statements will throw an error:

double x = 10.0 / 0.0;
float y = 371 / 0.0;
int z = 89.0 / 0.0;

The last line of the above looks like it should throw an error, but it doesn't. The reason for this is that it calculates 89.0 / 0.0 using floating point maths to be Inf, and then converts that into an integer. Since Inf doesn't convert directly, it picks the next best thing I'd assume. The following, however, will all throw an error:

int x = 10 / 0;
float y = 1000 / 0;
double z = 4560 / 0;