{

using boost::decimal::decimal64_t; // Type decimal64_t

constexpr decimal64_t a {"-5.123456891234567"}; // Constructs -5.123456 from string

constexpr decimal64_t b {"3.123456891234567"}; // Constructs 3.123456 from string

constexpr decimal64_t c {a + b};

// Here we can see that the result is exact

constexpr decimal64_t neg_two {-2};

static_assert(c == neg_two, "Result should be exact");

// We can use std::setprecision and std::numeric_limits in their usual ways

std::cout << std::setprecision(std::numeric_limits<decimal64_t>::digits10)

<< "A: " << a << '\n'

<< "B: " << b << '\n'

<< "A + B: " << c << '\n';

// The decimal library provides comprehensive implementations of the <cmath> functions

// that you would expect to have with the builtin floating point types

//

// They are all located in namespace boost::decimal::,

// as overloading namespace std is not allowed

constexpr decimal64_t abs_c {boost::decimal::abs(c)};

std::cout << "abs(A + B): " << abs_c << '\n';

// All cmath functions are constexpr even if their std:: counterparts are not

constexpr decimal64_t sqrt_two {boost::decimal::sqrt(abs_c)};

// Value compute by N[Sqrt[2], 50] using Wolfram Alpha or Mathematica

constexpr decimal64_t wa_sqrt_two {"1.4142135623730950488016887242096980785696718753769"};

std::cout << "sqrt(abs(A + B)): " << sqrt_two << '\n'

<< "Wolfram Alpha sqrt(2): " << wa_sqrt_two << '\n';