i/float_text

The text form of a Float (§11.8): the shortest digits that read back as the same Float, correctly rounded. core.inspect and every Float in an inspected value use this function, so it uses no core.inspect itself.

The digit loop is the Dragon interval construction (Steele & White, “How to Print Floating-Point Numbers Accurately”, 1990; Burger & Dybvig, “Printing Floating-Point Numbers Quickly and Accurately”, 1996): the exact value and the two rounding midpoints are exact integers, so every decision is an integer comparison, also for subnormals and the largest finite value. No floating-point logarithm is used.

to_stringCopy Anchor Link

The shortest text that reads back as x.

float_text.to_string(0.1)      // "0.1"
float_text.to_string(1.0e20)   // "1.0e20"
float_text.to_string(-0.0025)  // "-0.0025"

to_string = (x : Float) -> String {