Numeric conversions

Convert an existing numeric value by calling the target type:

local count: Int = 42
local compact = Int8(count)
local unsigned = UInt32(count)
local measurement = Float64(count)

This is compiler-known syntax, not an ordinary overload or runtime lookup. It accepts exactly one numeric operand. Pop does not perform implicit widening between existing values.

Integer to integer

The result must fit the target’s exact range:

local wide: Int = 1_000
local compact = UInt8(wide) -- traps: greater than 255

Changing width or signedness is always explicit. Out-of-range conversion traps with NumericConversion; it never wraps.

Integer to floating point

The conversion uses IEEE round-to-nearest, ties-to-even in the target format:

local count: UInt64 = 9_007_199_254_740_993
local approximate = Float64(count)

Large integers may round because a floating-point format cannot represent every integer in its range.

Floating point to integer

Conversion truncates toward zero:

local positive = Int32(12.75)  -- 12
local negative = Int32(-12.75) -- -12

NaN, infinity, and out-of-range results trap with NumericConversion.

Floating point width

Float32 to Float64 is exact. Float64 to Float32 rounds to the target representation:

local compact: Float32 = 1.25
local wide = Float64(compact)
local compactAgain = Float32(wide)

Converting a value to its own canonical numeric type preserves it.