cstddef.h

Go to the documentation of this file.

#pragma once

// FastLED Standard Definition Types (stddef.h equivalents)
//
// Provides size_t, ptrdiff_t, max_align_t, and offsetof macro.
 
namespace fl {
 
// FastLED equivalent of std::nullptr_t
typedef decltype(nullptr) nullptr_t;
 
// FastLED equivalent of std::size_t and std::ptrdiff_t
typedef __SIZE_TYPE__ size_t;
 
#ifdef __PTRDIFF_TYPE__
typedef __PTRDIFF_TYPE__ ptrdiff_t;
#else
// Fallback for compilers that don't define __PTRDIFF_TYPE__ (like older AVR-GCC)
typedef long ptrdiff_t;
#endif
 
// FastLED equivalent of std::max_align_t
// Use union approach (like standard library implementations) for fast compilation
// Template-based selection to handle platforms where long double == double
namespace detail {
    // Union variant with long double (for platforms with distinct long double like x86/x64)
    typedef union {
        long long ll;
        long double ld;
        void* p;
    } max_align_with_ld;
 
    // Union variant without long double (for ARM/AVR where long double == double)
    typedef union {
        long long ll;
        double d;
        void* p;
    } max_align_without_ld;
}
 
// Template selector - automatically chooses correct variant at compile time
template <bool UseLD>
struct max_align_selector {
    typedef detail::max_align_with_ld type;
};
 
template <>
struct max_align_selector<false> {
    typedef detail::max_align_without_ld type;
};
 
// Select appropriate union based on whether long double is distinct from double
typedef typename max_align_selector<(sizeof(long double) > sizeof(double))>::type max_align_t;
 
} // namespace fl
 
// Declare size_t and ptrdiff_t in global namespace for third-party C++ code
// (avoid max_align_t and nullptr_t which conflict with system headers)
typedef fl::size_t size_t;
typedef fl::ptrdiff_t ptrdiff_t;
 
 
// FastLED equivalent of offsetof macro (stddef.h)
// Computes byte offset of a member within a struct/class at compile time
// Uses compiler builtin to avoid including <stddef.h>
// Note: This is C++ only and depends on __cplusplus guard above
#define FL_OFFSETOF(type, member) __builtin_offsetof(type, member)