比特币源码阅读（0.16）（八）

uint256和arith_uint256

class arith_uint256 : public base_uint<256> {
public:
    arith_uint256() {}
    arith_uint256(const base_uint<256>& b) : base_uint<256>(b) {}
    arith_uint256(uint64_t b) : base_uint<256>(b) {}
    explicit arith_uint256(const std::string& str) : base_uint<256>(str) {}

    /**
     * The "compact" format is a representation of a whole
     * number N using an unsigned 32bit number similar to a
     * floating point format.
     * The most significant 8 bits are the unsigned exponent of base 256.
     * This exponent can be thought of as "number of bytes of N".
     * The lower 23 bits are the mantissa.
     * Bit number 24 (0x800000) represents the sign of N.
     * N = (-1^sign) * mantissa * 256^(exponent-3)
     *
     * Satoshi's original implementation used BN_bn2mpi() and BN_mpi2bn().
     * MPI uses the most significant bit of the first byte as sign.
     * Thus 0x1234560000 is compact (0x05123456)
     * and  0xc0de000000 is compact (0x0600c0de)
     *
     * Bitcoin only uses this "compact" format for encoding difficulty
     * targets, which are unsigned 256bit quantities.  Thus, all the
     * complexities of the sign bit and using base 256 are probably an
     * implementation accident.
     */
    arith_uint256& SetCompact(uint32_t nCompact, bool *pfNegative = nullptr, bool *pfOverflow = nullptr);
    uint32_t GetCompact(bool fNegative = false) const;

    friend uint256 ArithToUint256(const arith_uint256 &);
    friend arith_uint256 UintToArith256(const uint256 &);
};