Types

Types

Thagore is statically typed. Every binding has a known type by the end of type checking, either from an explicit annotation or from inference.

Primitive types

Type	Size	Range / semantics	Default value
i32	32-bit signed integer	-2147483648 to 2147483647	none; bindings must be initialized
i64	64-bit signed integer	64-bit signed integer range	none; bindings must be initialized
f64	64-bit IEEE floating point	double-precision floating point	none; bindings must be initialized
bool	logical boolean	true or false	none; bindings must be initialized
str	runtime-managed string handle	immutable string value	none; bindings must be initialized
void	no value	used for functions that do not return a value	implicit at function end
ptr	raw pointer-like runtime value	low-level interop and runtime APIs	none; bindings must be initialized

Thagore does not currently auto-initialize local variables. A declaration must either include a value or receive one before use.

Simple annotations

func main() -> i32:
  let count: i32 = 42
  let ratio: f64 = 3.5
  let ok: bool = true
  let name: str = "thagore"
  return 0

Inference

You can omit the type when the initializer is enough:

func main() -> i32:
  let count = 42
  let ratio = 3.5
  let label = "compiler"
  return 0

Arrays

The type checker supports runtime arrays as Array[T] or Array<T>. In current code, arrays mostly enter the program through runtime-backed APIs such as std.io.read_ints.

func sum_items(items: Array[i32]) -> i32:
  let total: i32 = 0
  for item in items:
    total = total + item
  return total

Struct types

Structs are nominal types:

struct Point:
  x: i32
  y: i32

func sum(point: Point) -> i32:
  return point.x + point.y

Generic type syntax

The AST and parser support generic type syntax with either angle brackets or legacy brackets:

func takes_vec(items: Vec<i32>) -> i32:
  return 0

func takes_array(items: Array[i32]) -> i32:
  return 0

Array[i32] is the currently exercised form in compiler tests and fixtures.

Conversion patterns

The language does not have a general cast operator today. Use stdlib helpers instead:

import std.string as string

func main():
  let text = string.from_int(42)
  println(text)

import std.string as string

func main() -> i32:
  let value: i32 = string.to_int("42")
  return value - 42

import std.string as string

func main():
  println(string.from_f64(3.5))
  println(string.from_bool(true))