Classification and special values v0.1.0

Telling NaNs, infinities, zeros, subnormals and normal values apart, and making the special ones.

`FloatClass`

pub enum FloatClass: int32 {
    NotANumber = 0,
    Infinite = 1,
    Zero = 2,
    Subnormal = 3,
    Normal = 4
}

The five kinds IEEE 754 sorts every value into. Copy.

Source

Cases

Name	Value	Description
`NotANumber`	`0`	Not a number: the result of an operation with no meaningful answer.
`Infinite`	`1`	One of the two infinities.
`Zero`	`2`	Positive or negative zero.
`Subnormal`	`3`	A value too small to carry a full significand, whose precision degrades toward zero.
`Normal`	`4`	An ordinary value at full precision.

`IsNan`

pub func IsNan(value: float64) -> bool
pub func IsNan(value: float32) -> bool

Whether value is a NaN. A NaN is the only value not equal to itself, which is why this needs no bit pattern.

Overload	Summary
`pub func IsNan(value: float64) -> bool`	Whether `value` is a NaN.
`pub func IsNan(value: float32) -> bool`	Determines whether `value` is a NaN at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value to test

Returns: true when value is a NaN

`IsInfinite`

pub func IsInfinite(value: float64) -> bool
pub func IsInfinite(value: float32) -> bool

Determines whether value is an infinity at float32 precision.

Overload	Summary
`pub func IsInfinite(value: float64) -> bool`	Whether `value` is one of the two infinities.
`pub func IsInfinite(value: float32) -> bool`	Determines whether `value` is an infinity at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value to test

Returns: true when value is positive or negative infinity

`IsFinite`

pub func IsFinite(value: float64) -> bool
pub func IsFinite(value: float32) -> bool

Whether value is an ordinary number: neither a NaN nor an infinity.

Overload	Summary
`pub func IsFinite(value: float64) -> bool`	Whether `value` is an ordinary number: neither a NaN nor an infinity.
`pub func IsFinite(value: float32) -> bool`	Determines whether `value` is finite at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value to test

Returns: true when value is neither a NaN nor an infinity

`IsZero`

pub func IsZero(value: float64) -> bool
pub func IsZero(value: float32) -> bool

Whether value is either zero. Both zeros answer true; SignBit is what tells them apart.

Overload	Summary
`pub func IsZero(value: float64) -> bool`	Whether `value` is either zero.
`pub func IsZero(value: float32) -> bool`	Determines whether `value` is either signed zero at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value to test

Returns: true when value is positive or negative zero

`IsSubnormal`

pub func IsSubnormal(value: float64) -> bool
pub func IsSubnormal(value: float32) -> bool

Whether value is subnormal: nonzero, but too small to carry a full significand.

Overload	Summary
`pub func IsSubnormal(value: float64) -> bool`	Whether `value` is subnormal: nonzero, but too small to carry a full significand.
`pub func IsSubnormal(value: float32) -> bool`	Determines whether `value` is subnormal at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value to test

Returns: true when value is subnormal

`IsNormal`

pub func IsNormal(value: float64) -> bool
pub func IsNormal(value: float32) -> bool

Whether value is an ordinary value at full precision. Zero is not normal, which surprises people and is the definition every IEEE 754 implementation uses.

Overload	Summary
`pub func IsNormal(value: float64) -> bool`	Whether `value` is an ordinary value at full precision.
`pub func IsNormal(value: float32) -> bool`	Determines whether `value` is normal at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value to test

Returns: true when value is normal

`Classify`

pub func Classify(value: float64) -> FloatClass
pub func Classify(value: float32) -> FloatClass

Which of the five kinds value belongs to.

Overload	Summary
`pub func Classify(value: float64) -> FloatClass`	Which of the five kinds `value` belongs to.
`pub func Classify(value: float32) -> FloatClass`	Classifies `value` at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value to classify

Returns: the IEEE-754 class of value

`SignBit`

pub func SignBit(value: float64) -> bool
pub func SignBit(value: float32) -> bool

Whether value's sign bit is set. This is the sign bit, not a comparison: SignBit(-0.0) is true where -0.0 < 0.0 is false, and a negative NaN answers true where every comparison against it answers false. That is the whole reason to ask by bit.

Overload	Summary
`pub func SignBit(value: float64) -> bool`	Whether `value`'s sign bit is set.
`pub func SignBit(value: float32) -> bool`	Determines whether the sign bit of `value` is set at `float32` precision.

Source

Parameters

Name	Type	Description
`value`	`float64` / `float32`	value whose sign is inspected

Returns: true when the sign bit is set

`PositiveInfinity64`

pub func PositiveInfinity64() -> float64

Positive infinity.

Source

Returns: positive infinity at float64 precision

`PositiveInfinity32`

pub func PositiveInfinity32() -> float32

Positive infinity at the narrow width.

Source

Returns: positive infinity at float32 precision

`QuietNan64`

pub func QuietNan64() -> float64

A quiet NaN.

Source

Returns: A quiet NaN at float64 precision

`QuietNan32`

pub func QuietNan32() -> float32

A quiet NaN at the narrow width.

Source

Returns: A quiet NaN at float32 precision

Absolute value, minimum and maximum

The magnitude of a floating-point value and the smaller or larger of two.

Exponentials and logarithms

Powers of e and two, and logarithms in base e, two and ten.

Edit this page on GitHub