Module kaydle_primitives::number

source · [−]
Expand description

Parsers and utility types for parsing KDL numbers. While KDL doesn’t distinguish between floats and integers, or signed and unsigned numbers, this module uses some simple logic to detect different integer types, to aid with deserialization.

Number type logic

Currently, we use a simple set of rules to decide if a KDL number should be parsed as an i64, u64, or f64:

  • If the number contains a fractional part or exponent, it’s unconditionally parsed as an f64.
  • Otherwise, if it’s negative, it’s parsed as an i64.
  • Otherwise, it’s parsed as a u64.

These rules may change in the future. Possible improvements:

  • If it includes an exponent, it might be an integer, even if it includes a fractional parts.
  • If it’s an integer but it overflows an i64 or u64, it could be stored in an f64 instead.
  • It’s also possible we could use type hints (from serde or KDL annotations) to guide the parse as well.

Structs

BoundsError

A Bounds error occurred during number parsing. This type is incomplete and will grow in the future.

Enums

KdlInt

A KDL integer, which is either signed or unsigned. In practice, we always parse positive integers as unsigned and negative integers as signed.

KdlNumber

A KDL Number. The KDL spec doesn’t distinguish between integers and floats, or between signed an unsigned numbers, but kaydle uses a heuristic to pick a type for deserialization purposes.

Sign

The parsed sign of a number

Traits

IntBuilder

Helper trait for building or recognizing integers.

NumberBuilder

Trait for building KDL numbers

Functions

parse_number

Parse a KDL number