Sublcass의 constexpr 클래스는 다른 사용자 정의 변환 동작을 구하는

Question

내가 가지고있는 constexpr 만드는 데 사용되는 라이브러리에서 클래스와 다음 부동 소수점 숫자의 고정 소수점 표현 관리 : 나는 '

template< unsigned Integer, unsigned Fraction > 
class UFixed : FIXED_POINTS_DETAILS::UFixedBase< Integer, Fraction > 
{ 
public: 
    static_assert((Integer + Fraction) <= FIXED_POINTS_DETAILS::BitSize<uintmax_t>::Value, "Platform does not have a native type large enough for UFixed."); 

public: 
    using IntegerType = FIXED_POINTS_DETAILS::LeastUInt<Integer>; 
    using FractionType = FIXED_POINTS_DETAILS::LeastUInt<Fraction>; 
    using InternalType = FIXED_POINTS_DETAILS::LeastUInt<Integer + Fraction>; 

    using ShiftType = FIXED_POINTS_DETAILS::LeastUInt<Integer + Fraction>; 
    using MaskType = FIXED_POINTS_DETAILS::LeastUInt<Integer + Fraction>; 

    constexpr const static uintmax_t IntegerSize = Integer; 
    constexpr const static uintmax_t FractionSize = Fraction; 
    constexpr const static uintmax_t LogicalSize = IntegerSize + FractionSize; 
    constexpr const static uintmax_t InternalSize = FIXED_POINTS_DETAILS::BitSize<InternalType>::Value; 

    constexpr const static uintmax_t Scale = 1ULL << FractionSize; 

public: 
    constexpr const static ShiftType IntegerShift = FractionSize; 
    constexpr const static ShiftType FractionShift = 0; 

    constexpr const static MaskType IntegerMask = FIXED_POINTS_DETAILS::IdentityMask<IntegerSize>::Value; 
    constexpr const static MaskType FractionMask = FIXED_POINTS_DETAILS::IdentityMask<FractionSize>::Value; 

    constexpr const static MaskType IdentityMask = (IntegerMask << IntegerShift) | (FractionMask << FractionShift); 

    constexpr const static MaskType MidpointMask = FIXED_POINTS_DETAILS::MsbMask<FractionSize>::Value; 
    constexpr const static MaskType LesserMidpointMask = MidpointMask - 1; 

private: 
    using Base = FIXED_POINTS_DETAILS::UFixedBase<Integer, Fraction>; 
    using RawType = typename Base::RawType; 

public: 
    using Base::Base; 

    constexpr UFixed(void); 
    constexpr UFixed(const IntegerType & integer); 
    constexpr UFixed(const IntegerType & integer, const FractionType & fraction); 

    constexpr InternalType getInternal(void) const; 
    constexpr IntegerType getInteger(void) const; 
    constexpr FractionType getFraction(void) const; 

    constexpr explicit operator IntegerType(void) const; 
    constexpr explicit operator float(void) const; 
    constexpr explicit operator double(void) const; 

    template< unsigned IntegerOut, unsigned FractionOut > 
    constexpr explicit operator UFixed<IntegerOut, FractionOut>(void) const; 

    constexpr static UFixed fromInternal(const InternalType & value); 

    UFixed & operator ++(void); 
    UFixed & operator --(void); 
    UFixed & operator +=(const UFixed & other); 
    UFixed & operator -=(const UFixed & other); 
    UFixed & operator *=(const UFixed & other); 
    UFixed & operator /=(const UFixed & other); 

public: 
    constexpr const static UFixed Epsilon = UFixed::fromInternal(1); 
    constexpr const static UFixed MinValue = UFixed::fromInternal(0); 
    constexpr const static UFixed MaxValue = UFixed::fromInternal(~0); 

    // 40 digits is probably enough for these 
    constexpr const static UFixed Pi = 3.1415926535897932384626433832795028841971; 
    constexpr const static UFixed E = 2.718281828459045235360287471352662497757; 
    constexpr const static UFixed Phi = 1.6180339887498948482045868343656381177203; 
    constexpr const static UFixed Tau = 6.2831853071795864769252867665590057683943; 
}; 

을 두 번째 - 마지막 블록에서 IntegerType, float 및 double에 대한 사용자 정의 변환 연산자를 제외하고 실질적으로 동등하게 동작하는 UFixed의 하위 클래스를 상속을 사용하여 정의하는 방법이 있는지 궁금합니다. explicit.

이 내 요구 사항에 작동하는 것 같다 또한 기본 클래스로 "constexpr"클래스

Answer 1

: 왜 당신은 할 수 없었다

template <unsigned Integer, unsigned Fraction> 
struct MyUFixed : UFixed<Integer, Fraction> 
{ 
    typedef UFixed<Integer, Fraction> fixed_point_t; 
    using fixed_point_t::fixed_point_t; //pull in base constructors 
    using typename fixed_point_t::IntegerType; //pull in "IntegerType" 

    using fixed_point_t::operator double; 
    using fixed_point_t::operator float; 
    using fixed_point_t::operator IntegerType; 

    constexpr MyUFixed(const fixed_point_t& other) : 
    fixed_point_t(other) //constructor from base type 
    {} 

    //implicit conversions 

    constexpr operator float() const 
    { 
     return static_cast<float>(*this); 
    } 

    constexpr operator double() const 
    { 
     return static_cast<double>(*this); 
    } 

    constexpr operator IntegerType() const 
    { 
     return static_cast<IntegerType>(*this); 
    } 
};