<xstd/ints/cstdlib/div.hpp>

xstd::div

template<integer I>
  constexpr auto div(I numer, I denom) noexcept(nothrow_const_operators<I>) -> div_result<I>;

The quotient and remainder of truncated integer division, at the operands' own width.

Parameters

Parameter	Meaning
numer	the numerator
denom	the denominator

Preconditions: denom != 0, and for a signed I, not both numer == xstd::numeric_limits<I>::min() and denom == -1. Each is undefined behaviour that / and % already have.

The second precondition does not apply to an unsigned I, where min() is 0 and -1 converts to max(): div(0, max()) is {0, 0}.

Return value

Returns: {.quotient = numer / denom, .remainder = numer % denom}, the truncated division of [expr.mul]: the quotient rounds toward zero, numer == denom * quotient + remainder, |remainder| < |denom|, and a nonzero remainder has the sign of numer.

Exceptions

The exception specification is nothrow_const_operators<I>.

Remarks

It generalizes div, ldiv, lldiv and imaxdiv ([cstdlib.syn], constexpr since P0533R9) to every integer, xstd::int128 among them, which div does not take, and returns div_result<I> at the operands' own width rather than at a promoted one.

Notes

• A call is a constant expression wherever I's own / and % are.
• Both operands have the one type I: a literal divisor beside an integer-class numerator deduces int for it, and the call is ill-formed rather than converted.

See also

Entity	Relation
div_result	the result type
div_euclid, div_floor, div_ceil	the other roundings
std::div	the standard functions it generalizes