Integers and overflow

In most languages, integers are stored in a fixed number of bits. A 32-bit signed int holds roughly minus 2.1 billion to plus 2.1 billion, and a 64-bit one about 9.2 × 10^18. Negative numbers use two's complement, a scheme where the top bit acts as the sign.

Go past the limit and you get integer overflow. Java and C# wrap silently to a large negative number by default, and in C++ signed overflow is undefined behaviour. Python integers grow as large as memory allows.

JavaScript numbers stop being exact above 2^53, which is why BigInt exists. Overflow bugs show up in money totals, timestamps and counters, so know the limits of the types you use.