Float32

Float32 stores a 32-bit IEEE 754 floating-point value. An annotation makes a decimal literal use this format:

local ratio: Float32 = 0.5
local scientific: Float32 = 2.5e3
local doubled = ratio * 2.0

Arithmetic rounds to Float32. Division by zero follows IEEE behavior and can produce infinity. Ordering with NaN is always false.

local count: Int = 7
local ratio = Float32(count)
local wide = Float64(ratio)

Integer-to-float conversion uses IEEE round-to-nearest, ties-to-even. Widening Float32 to Float64 is exact.