Rounding and remainders v0.1.0

Ceil

pub func Ceil(x: float64) -> float64
pub func Ceil(x: float32) -> float32

Rounds x toward positive infinity to an integral value. A negative value between zero and negative one becomes -0.0. A NaN or infinity passes through unchanged.

Source

Parameters

Returns: the smallest integral value greater than or equal to x

Floor

pub func Floor(x: float64) -> float64
pub func Floor(x: float32) -> float32

Rounds x toward negative infinity to an integral value. The sign of zero is preserved. A NaN or infinity passes through unchanged.

Source

Parameters

Returns: the largest integral value less than or equal to x

Mod

pub func Mod(x: float64, y: float64) -> float64
pub func Mod(x: float32, y: float32) -> float32

Computes the floating-point remainder of x / y, with the sign of x. The result is x - n * y, where n is x / y truncated toward zero. Its magnitude is less than |y|. A zero divisor, an infinite dividend, or a NaN operand returns a NaN; an infinite divisor returns x. The significands are reduced with integer long division so the intermediate product cannot discard the remainder.

Source

Parameters

Returns: the exact remainder, or a NaN when no remainder is defined

Round

pub func Round(x: float64) -> float64
pub func Round(x: float32) -> float32

Rounds x to the nearest integral value, with halfway cases away from zero. Thus Round(2.5) is 3.0 and Round(-2.5) is -3.0. The sign of zero is preserved, and a NaN or infinity passes through unchanged. The fractional part is compared directly instead of adding one half, which avoids prematurely rounding the largest representable value below 0.5 up to 1.0.

Source

Parameters

Returns: the nearest integral value

RoundingMode

pub enum RoundingMode: int32 {
    TowardZero = 0,
    TowardNegative = 1,
    TowardPositive = 2,
    ToNearestTiesAway = 3,
    ToNearestTiesEven = 4
}

Which way a value that falls between two integers should go. Copy.

Source

Cases

RoundTiesToEven

pub func RoundTiesToEven(value: float64) -> float64
pub func RoundTiesToEven(value: float32) -> float32

The nearest integer to value, with a halfway value going to the even neighbour. RoundTiesToEven(2.5) is 2 and RoundTiesToEven(3.5) is 4, where Round gives 3 and 4. The sign of a zero is preserved, and a NaN or infinity passes through unchanged.

Source

Parameters

Returns: the nearest integral value, resolving halfway cases toward the even neighbour

RoundTo

pub func RoundTo(
    value: float64,
    mode: RoundingMode
) -> float64

pub func RoundTo(
    value: float32,
    mode: RoundingMode
) -> float32

Rounds value in the requested direction at float32 precision.

Source

Parameters

Returns: the integral value selected by mode

ToIntegerChecked<T>

pub func ToIntegerChecked<T>(
    value: float64,
    sample: T,
    mode: RoundingMode,
    result: *var T
) -> bool

value rounded in the named direction and converted to an integer, reporting whether it fits. sample is read by value and only its type matters, exactly as Core's MaximumOf reads one. Returns true when the rounded value is not representable in the integer type — because it is a NaN, an infinity, or simply out of range — and leaves result zero in that case. This is the only place a float becomes an integer, and it refuses rather than saturating, because a caller who has not thought about the out-of-range case is better served by a flag than by a plausible wrong number. result must be non-null and writable.

Source

Type parameters

Parameters

Returns: true when the rounded value is not representable as T

Trunc

pub func Trunc(x: float64) -> float64
pub func Trunc(x: float32) -> float32

Rounds x toward zero to an integral value at float32 precision. The wide result is exactly representable at the input width, so narrowing it back to float32 loses nothing.

Source

Parameters

Returns: the integral part of x