mathtruncxx

mathtruncxx is an open-source C++ library that provides robust truncation utilities for converting floating-point and fixed-point values to integral types. It emphasizes deterministic, well-defined behavior across platforms and seeks to replace ad hoc casts with explicit truncation routines that are predictable and easy to audit. The library targets developers who need precise control over how values are reduced to integers, including safe handling of edge cases such as overflow and NaN inputs.

Core features include templated truncation functions that produce an integral result by discarding fractional parts toward zero, together with overloads that perform overflow checks and return error indicators or throw exceptions when the conversion would overflow the destination type. It supports a wide range of integral types, including built-in 8-, 16-, 32- and 64-bit types, and, where the compiler supports it, extended 128-bit types. The project also provides a safe_cast family that yields std::optional<T> or a similar error-reporting mechanism instead of failing silently.

Design and usage: mathtruncxx is distributed as a header-only library and emphasizes constexpr compatibility, enabling compile-time truncations where possible. It aims to be dependency-free beyond the C++ standard library and to integrate smoothly with existing code by providing familiar naming and semantics alongside explicit error handling.

History and licensing: the project began as a community effort in the early 2020s and has since evolved through multiple releases. It is released under a permissive open-source license and hosted on a public repository with documentation and examples.

See also: std::trunc, Boost safe_cast, saturating_cast utilities.