Safe Numbers

A safe and fully featured replacement for C++ builtin numeric types. NOTE: This is an experimental library and is subject to change without warning.

Supported Platforms

This library requires C++20 and is tested on the following platforms:

• GCC 11 and later
• Clang 13 and later
• Visual Studio 2022 and later
• Intel OneAPI DPC++

This library also supports compilation with NVCC (tested with 12.8 and later).

Synopsis

The Safe Numbers library provides a number of fixed width integers and floating point types:

#include <boost/safe_numbers.hpp>

namespace boost::safe_numbers {

// Unsigned integer types with specified bit width

class u8;
class u16;
class u32;
class u64;
class u128;

// Signed integer types with specified bit width

class i8;
class i16;
class i32;
class i64;
class i128;

// Binary floating point types with specified bit width

class f32;
class f64;

// Every width is also available with a saturating or strict error policy
// in place of the default throwing one, e.g.:

class sat_u8;    // clamps to the numeric limits instead of throwing
class strict_i32; // terminates on any error
class sat_f64;   // raw IEEE 754 arithmetic with no checks

} // namespace boost::safe_numbers

These types operate much like the built-in numeric types but with far stricter behavior to enforce correctness. This includes no implicit conversions, no mixed type operations, and (by default) throwing an exception on underflow, overflow, or other incorrect operation. A fully featured implementation analogous to the STL is included with <cmath>, <charconv>, <format>, etc. support.

Using these types is straightforward and can be learned by example. Their usage is designed to closely resemble that of the builtin-numeric types.

Full Documentation:

The complete documentation can be found at: https://develop.safe-numbers.cpp.al/