MyBit

mybit.h — tiny header-only byte-swap utilities

Usage:

• Include the header in your C or C++ project:

  #include "mybit.h"

• Example: compile and run the test:

  • GCC: gcc -std=c11 -O2 test.c -o test_mybit ; ./test_mybit g++ -std=c++11 -O2 testcpp.cpp -o testcpp_mybit ; ./testcpp_mybit

  ---

  • Clang: clang -std=c11 -O2 test.c -o test_mybit ; ./test_mybit clang++ -std=c++11 -O2 testcpp.cpp -o testcpp_mybit ; ./testcpp_mybit

  ---

  • MSVC: cl /std:c11 /O2 test.c /Fe:test_mybit.exe && test_mybit.exe cl /std:c++14 /O2 testcpp.cpp /Fe:testcpp_mybit.exe && testcpp_mybit.exe

• Compatible versions:

  C: Compatible from C89/C90 onwards. C++ (using ONLY C functions): Compatible from C++98 onwards. C++ (using C++'s own features): Compatible from C++14 or C++11 onwards (with caution).

• Freestanding:

  Since this library does not depend on an operating system (there is no stdlib.h or stdio.h, i.e., freestanding), it can be used anywhere, on any operating system, in a driver, in a kernel, and even in the firmware of a microchip without an OS (bare-metal).

Features

• Header-only: Drop mybit.h into your project and you are ready to go.
• Portability: Supports C (pre-C11 fallback, C11 _Generic) and C++ (C++11 up to C++20 constexpr templates).
• Hardware Acceleration: Automatically detects and uses compiler builtins...
• Safe Macro Evaluation: Fully protected against macro double-evaluation side effects (e.g., safely pass expressions like mybit_bswap(*ptr++) even in older pre-C11 compilers).
• Signed Integer Support: Robust type casting rules seamlessly handle signed integers (int, short, long long) by forcing logical bitwise operations, eliminating compiler-specific undefined behaviors like arithmetic right-shift sign extensions.

Notes:

• Defines mybit_bswap16, mybit_bswap32, mybit_bswap64, plus a generic mybit_bswap(x) wrapper for C and C++.
• Defines portable bit utilities:
  • mybit_has_single_bit32, mybit_bit_floor32, mybit_bit_ceil32
  • mybit_has_single_bit64, mybit_bit_floor64, mybit_bit_ceil64
  • mybit_clz32, mybit_ctz32, mybit_popcount32, mybit_rotl32, mybit_rotr32, mybit_reverse32
  • mybit_clz64, mybit_ctz64, mybit_popcount64, mybit_rotl64, mybit_rotr64, mybit_reverse64
• Endian load/store helpers:
  • mybit_load_le16, mybit_store_le16, mybit_load_be16, mybit_store_be16
  • mybit_load_le32, mybit_store_le32, mybit_load_be32, mybit_store_be32
  • mybit_load_le64, mybit_store_le64, mybit_load_be64, mybit_store_be64
• This library intentionally does not expose legacy byteswap* symbols to avoid collisions; add your own compatibility aliases if you need them.
• Additional helpers:
  • mybit_countl_one32, mybit_countr_one32, mybit_countl_one64, mybit_countr_one64
  • C++ constexpr templates in namespace mybit for countl_one, countr_one, bit_floor, bit_ceil, has_single_bit, etc.
• Define MYBIT_ASSUME_LITTLE_ENDIAN to 1 or 0 to override endianness detection for cross-compilation.

Added in this update

This release adds full 8-bit and 16-bit counterparts for the existing 32/64-bit helpers, generic C wrappers, and C++ overloads/templates.

• 8-bit helpers: mybit_clz8, mybit_ctz8, mybit_popcount8, mybit_rotl8, mybit_rotr8, mybit_reverse8, mybit_has_single_bit8, mybit_bit_floor8, mybit_bit_ceil8, plus mybit_load_le8/mybit_load_be8 and mybit_store_le8/mybit_store_be8.
• 16-bit helpers: mybit_clz16, mybit_ctz16, mybit_popcount16, mybit_rotl16, mybit_rotr16, mybit_reverse16, mybit_has_single_bit16, mybit_bit_floor16, mybit_bit_ceil16, plus endian load/store mybit_load_le16/mybit_load_be16 and mybit_store_le16/mybit_store_be16.
• C generic wrappers: When compiling as C11 the header exposes _Generic wrappers such as mybit_clz(x), mybit_ctz(x), mybit_popcount(x), mybit_rotl(x, r), mybit_rotr(x, r), mybit_reverse(x), mybit_bit_floor(x), mybit_bit_ceil(x), and mybit_has_single_bit(x) that dispatch to the correct width (8/16/32/64). Pre-C11 builds use sizeof-based macro fallbacks to provide the same convenience.
• C++ wrappers and templates: C++ builds provide overloads and constexpr templates in the mybit namespace (e.g., mybit::popcount, mybit::countl_zero, mybit::countr_zero, mybit::rotl, mybit::reverse, mybit::bit_floor, mybit::bit_ceil, mybit::load_le, mybit::store_le, mybit::pext, mybit::pdep). These are enabled for unsigned integer types and provide type-safe, constexpr-capable alternatives to the C macros.

Example (C):

#include "mybit.h"
uint32_t x = 0x0F00;
int leading = mybit_clz(x); // dispatches to mybit_clz32

Example (C++):

#include "mybit.h"
using namespace mybit;
constexpr auto p = popcount(uint32_t{0xFF00});
constexpr auto f = bit_floor(uint32_t{15});