path: root/docs/TYPES.md

# Typing in Puck

Puck has a comprehensive static type system, inspired by the likes of Nim, Rust, and Swift.

## Basic types

Basic types can be one-of:
- `bool`: internally an enum.
- `int`: integer number. x bits of precision by default.
  - `uint`: same as `int`, but unsigned for more precision.
  - `i8`, `i16`, `i32`, `i64`, `i128`: specified integer size
  - `u8`, `u16`, `u32`, `u64`, `u128`: specified integer size
- `float`: floating-point number.
  - `f32`, `f64`: specified float sizes
- `byte`: an alias to `u8`.
- `char`: a *four-byte* value. For working with Unicode. An alias to `u32`.
  - these are *packed* when part of a string: and so indexing directly into a string is a no-op. string access is O(n), swift-style.
- `str`: a string type. mutable. internally a byte-array: externally a char-array.
- `void`: an internal type designating the absence of a value.
  - possibly, the empty tuple. then would `empty` be better? or `unit`?
- `never`: a type that denotes functions that do not return.
  - distinct from returning nothing.
  - the bottom type.

`bool`, `int`/`uint` and siblings, `float` and siblings, `byte`, and `char` are all considered **primitive types** and are _always_ copied (unless passed as `var`). More on when parameters are passed by value (copied) vs. passed by reference can be found in the [memory management document](MEMORY_MANAGEMENT.md).

Primitive types combine with `str`, `void`, and `never` to form the **basic types**. `void` and `never` will rarely be referenced by name: instead, the absence of a type typically implicitly denotes one or the other. Still, having a name is helpful in some situations.

### integers

todo

### strings

Strings are:
- mutable
- internally a byte array
- externally a char (four bytes) array
- prefixed with their length and capacity
- automatically resize like a list

They are also quite complicated. Puck has full support for Unicode and wishes to be intuitive, performant, and safe, as all languages wish to be. Strings present a problem that much effort has been spent on in Swift and Rust (primarily) to solve.
## Container types

Container types, broadly speaking, are types that contain other types. These exclude the types in [advanced types](#advanced-types).

### Container types

Container types can be one-of:
- `array[S, T]`: Fixed-size arrays. Can only contain one type `T`. Of size `S` and cannot grow/shrink.
  - Initialized in-place with `[a, b, c]`.
- `list[T]`: Dynamic arrays. Can only contain one type `T`. May grow/shrink dynamically.
  - Initialized in-place with `[a, b, c]`. (this is the same as arrays!) <!-- Disambiguated from arrays in much the same way uints are disambiguated from ints. -->
- `slice[T]`: Slices. Used to represent a "view" into some sequence of elements of type `T`.
  - Cannot be directly constructed. They are **unsized**. <!-- May be initialized from an array, list, or string, or may be used as a generic parameter on functions (more on that later). -->
  - Slices cannot grow/shrink. Their elements may be accessed and mutated. As they are underlyingly a reference to an array or list, they **must not** outlive the data they reference.
- `str`: Strings. Complicated, they are alternatively `list[byte]` or `list[char]` depending on who's asking. In the context of iterable types, they are treated as `list[char]`.

All of these above types are some sort of sequence: and so have a length, and so can be _iterated_.
For convenience, a special `iterable` generic type is defined for use in parameters: that abstracts over all of the container types. This `iterable` type is also extended to any collection of a single type with a length.
- Is this redundant with `slice[T]`? todo

Elements of container types can be accessed by the `container[index]` syntax. Slices of container types can be accessed by the `container[lowerbound..upperbound]` syntax. Slices of non-consecutive elements can be accessed by the `container[a,b,c..d]` syntax, and indeed, the previous example expands to these. They can also be combined: `container[a,b,c..d]`.

### Abstract data types

There are an additional suite of related types: abstract data types. While falling under container types, these do not have a necessarily straightforward or best implementation, and so multiple implementations are provided.

Abstract data types can be one-of:
- `BitSet[T]`: high-performance sets implemented as a bit array.
  - These have a maximum data size, at which point the compiler will suggest using a `HashSet[T]` instead.
- `AssocTable[T, U]`: simple symbol tables implemented as an association list.
  - These do not have a maximum size. However, at some point the compiler will suggest using a `HashTable[T, U]` instead.
- `HashSet[T]`: standard hash sets.
- `HashTable[T, U]`: standard hash tables.

Unlike iterable types, abstract data types are not iterable by default: as they are not ordered, and thus, it is not clear how they should be iterated. Despite this: for utility purposes, an `elems()` iterator based on a normalization of the elements is provided for `set` and `HashSet`, and `keys()`, `values()`, and `pairs()` iterators are provided for `table` and `HashTable` based on a normalization of the keys. This is deterministic to prevent user reliance on shoddy randomization, see Golang.

## Parameter types

Some types are only valid when being passed to a function, or in similar contexts.
No variables may be assigned these types, nor may any function return them.
These are monomorphized into more specific functions at compile-time if needed.

Parameter types can be one-of:
- mutable: `func foo(a: mut str)`: Denotes the mutability of a parameter. Parameters are immutable by default.
  - Passed as a `ref` if not one already, and marked mutable.
- static: `func foo(a: static str)`: Denotes a parameter whose value must be known at compile-time. Useful with `when` for writing generic code.
- generic: `func foo[T](a: list[T], b: T)`: The standard implementation of generics, where a parameter's exact type is not listed, and instead statically dispatched based on usage.
- constrained: `func foo(a: str | int | float)`: A basic implementation of generics, where a parameter can be one-of several listed types. Makes for particularly straightforward monomorphization.
  <!-- - Separated with the bitwise or operator `|` rather than the symbolic or `||` or a raw `or` to give the impression that there isn't a corresponding "and" operation (the `&` operator is preoccupied with strings). -->
- functions: `func foo(a: func (x, y: int): int)`: First-class functions.
  - Functions may be prefixed with modifiers: one-of `pure`, `yeet`, `async`.
  - Syntactic sugar is available: `func foo(a: (int, int) -> int)`. This is not usable with modifiers.
- interfaces: `func foo(a: Stack[int])`: User-defined duck typing. Equivalent to and more general than Rust's traits or Haskell's typeclasses.
  - ex. for above: the `Stack[T]` interface specifies generic `push` and `pop` operations
  - More on interfaces in the [interfaces section](#interfaces).
- built-in interface: `func foo(a: enum)`: Included, special interfaces for being generic over [advanced types](#advanced-types).
  - Of note is how `slice[T]` functions: it is generic over `lists` and `arrays` of any length.

These parameter types (except `static`) share a common trait: they are not *sized*. The exact type is not generally known until compilation - and in the case of interfaces, sometimes not even during compilation! As the size is not always rigorously known, problems arise when attempting to construct parameter types or compose them with other types: and so this is disallowed. Not all is lost, however, as they may still be used with *indirection* - detailed in the [section on reference types](#reference-types).

### generic types

Functions can take a _generic_ type, that is, be defined for a number of types at once:

```puck
func add[T](a: list[T], b: T) =
  return a.add(b)

func length[T](a: T) =
  return a.len # monomorphizes based on usage.
  # lots of things use .len, but only a few called by this do.
  # throws a warning if exported for lack of specitivity.

func length(a: str | list) =
  return a.len
```

The syntax for generics is `func`, `ident`, followed by the names of the generic parameters in brackets `[T, U, V]`, followed by the function's parameters (which may then refer to the generic types).
Generics are replaced with concrete types at compile time (monomorphization) based on their usage in function calls within the main function body.

Constrained generics have two syntaxes: the constraint can be defined directly on a parameter, leaving off the `[T]` box, or it may be defined within the box as `[T: int | float]` for easy reuse in the parameters.

Other constructions like modules and type declarations themselves may also be generic.

## Reference Types

Types are typically constructed by value on the stack. That is, without any level of indirection: and so type declarations that recursively refer to one another, or involve unsized types (notably including parameter types), would not be allowed. However, Puck provides two avenues for indirection.

Reference types can be one-of:
- `ref T`: An automatically-managed reference to type `T`. This is a pointer of size `uint` (native).
- `ptr T`: A manually-managed pointer to type `T`. (very) unsafe. The compiler will yell at you.

```puck
type BinaryTree = ref struct
  left: BinaryTree
  right: BinaryTree

type AbstractTree[T] = interface
  func left(self: Self): Option[AbstractTree[T]]
  func right(self: Self): Option[AbstractTree[T]]
  func data(self: Self): T

type AbstractRoot[T] = struct
  left: ref AbstractTree[T]
  right: ref AbstractTree[T]

# allowed, but unsafe & strongly discouraged
type UnsafeTree = struct
  left: ptr UnsafeTree
  right: ptr UnsafeTree
```

The `ref` prefix may be placed at the top level of type declarations, or inside on a field of a structural type. `ref` types may often be more efficient when dealing with large data structures. They also provide for the usage of unsized types (functions, interfaces, slices) within type declarations.

The compiler abstracts over `ref` types to provide optimization for reference counts: and so a distinction between `Rc`/`Arc`/`Box` is not needed. Furthermore, access implicitly dereferences (with address access available via `.addr`), and so a `*` dereference operator is also not needed. Much care has been given to make references efficient and safe, and so `ptr` should be avoided if at all possible. The compiler will yell at you if you use it (or any other unsafe features).

The implementation of `ref` is delved into in further detail in the [memory management document](MEMORY_MANAGEMENT.md).

## Advanced Types

The `type` keyword is used to declare aliases to custom data types. These types are *algebraic*: they function by composition. Algebraic data types can be one-of:
- `struct`: An unordered, named collection of types. May have default values.
- `tuple`: An ordered collection of types. Optionally named.
- `enum`: Ordinal labels, that may hold values. Their default values are their ordinality.
- `union`: Powerful matchable tagged unions a la Rust. Sum types.
- `interface`: Implicit typeclasses. User-defined duck typing.

There also exist `distinct` types: while `type` declarations define an alias to an existing or new type, `distinct` types define a type that must be explicitly converted to/from. This is useful for having some level of separation from the implicit interfaces that abound.

### structs

Structs are an *unordered* collection of named types.

They are declared with `struct[identifier: Type, ...]` and initialized with brackets: `{field: "value", another: 500}`.

```puck
type LinkedNode[T] = struct
  previous, next: Option[ref LinkedNode[T]]
  data: T

let node = {
  previous, next = None
  data = 413
}

func pretty_print(node: LinkedNode[int]) =
  print node.data
  if node.next of Some(node):
    node.pretty_print()

# structural typing!
prints_data(node)
```

Structs are *structural* and so structs composed entirely of fields with the same signature (identical in name and type) are considered *equivalent*.
This is part of a broader structural trend in the type system, and is discussed in detail in [subtyping](#subtyping).

### tuples

Tuples are an *ordered* collection of either named and/or unnamed types.

They are declared with `tuple[Type, identifier: Type, ...]` and initialized with parentheses: `(413, "hello", value: 40000)`. Syntax sugar allows for them to be declared with `()` as well.

They are exclusively ordered - named types within tuples are just syntax sugar for positional access. Passing a fully unnamed tuple into a context that expects a tuple with a named parameter is allowed so long as the types line up in order.

```puck
let grouping = (1, 2, 3)

func foo: tuple[string, string] = ("hello", "world")
```

Tuples are particularly useful for "on-the-fly" types. Creating type aliases to tuples is discouraged - structs are generally a better choice for custom type declarations.

### enums

Enums are *ordinal labels* that may have *associated values*.

They are declared with `enum[Label, AnotherLabel = 4, ...]` and are never initialized (their values are known statically).
Enums may be accessed directly by their label, and are ordinal and iterable regardless of their associated value. They are useful in collecting large numbers of "magic values", that would otherwise be constants.

```puck
type Keys = enum
  Left, Right, Up, Down
  A = "a"
  B = "b"
```

In the case of an identifier conflict (with other enum labels, or types, or...) they must be prefixed with the name of their associated type (separated by a dot). This is standard for identifier conflicts: and is discussed in more detail in the [modules document](MODULES.md).

### unions

Unions are *tagged* type unions. They provide a high-level wrapper over an inner type that must be safely accessed via pattern matching.

They are declared with `union[Variant(Type), ...]` and initialized with the name of a variant followed by its inner type constructor in brackets: `Square(side: 5)`. Tuples and structs are special-cased to eliminate extraneous parentheses.

```puck
type Value = u64
type Ident = str
type Expr = ref union
  Literal(Value)
  Variable(Ident)
  Abstraction(param: Ident, body: Expr)
  Application(body: Expr, arg: Expr)
  Conditional(
    condition: Expr
    then_case: Expr
    else_case: Expr
  )
```

They take up as much space in memory as the largest variant, plus the size of the tag (one byte).

#### pattern matching

Unions abstract over differing types. In order to *safely* be used, their inner types must be accessed via *pattern matching*: leaving no room for type confusion. Pattern matching in Puck relies on two syntactic constructs: the `match` statement, forcing qualification and handling of all possible types of a variable, and the `of` statement, querying type equality while simultaneously binding new identifiers to underspecified portions of variables.

```puck
use std.tables

func eval(context: mut HashTable[Ident, Value], expr: Expr): Result[Value]
  match expr
  of Literal(value): Okay(value)
  of Variable(ident):
    context.get(ident).err("Variable not in context")
  of Application(body, arg):
    if body of Abstraction(param, body as inner_body):
      context.set(param, context.eval(arg)?) # from std.tables
      context.eval(inner_body)
    else:
      Error("Expected Abstraction, found {}".fmt(body))
  of Conditional(condition, then_case, else_case):
    if context.eval(condition)? == "true":
      context.eval(then_case)
    else:
      context.eval(else_case)
  of expr:
    Error("Invalid expression {}".fmt(expr))
```

The match statement takes exclusively a list of `of` sub-expressions, and checks for exhaustivity. The `expr of Type(binding)` syntax can be reused as a conditional, in `if` statements and elsewhere.

The `of` *operator* is similar to the `is` operator in that it queries type equality, returning a boolean. However, unbound identifiers within `of` expressions are bound to appropriate values (if matched) and injected into the scope. This allows for succinct handling of `union` types in situations where `match` is overkill.

Each branch of a match expression can also have a *guard*: an arbitrary conditional that must be met in order for it to match. Guards are written as `where cond` and immediately follow the last pattern in an `of` branch, preceding the colon.

### interfaces

Interfaces can be thought of as analogous to Rust's traits, without explicit `impl` blocks and without need for the `derive` macro. Types that have functions fulfilling the interface requirements implicitly implement the associated interface.

The `interface` type is composed of a list of function signatures that refer to the special type `Self` that must exist for a type to be valid. The special type `Self` is replaced with the concrete type at compile time in order to typecheck. They are declared with `interface[signature, ...]`.

```puck
type Stack[T] = interface
  push(self: mut Self, val: T)
  pop(self: mut Self): T
  peek(self: Self): T

func takes_any_stack(stack: Stack[int]) =
  # only stack.push, stack.pop, and stack.peek are available methods
```

Differing from Rust, Haskell, and many others, there is no explicit `impl` block. If there exist functions for a type that satisfy all of an interface's signatures, it is considered to match and the interface typechecks. This may seem strange and ambiguous - but again, static typing and uniform function call syntax help make this a more reasonable design. The purpose of explicit `impl` blocks in ex. Rust is three-fold: to provide a limited form of uniform function call syntax; to explicitly group together associated code; and to disambiguate. UFCS provides for the first, the module system provides for the second, and the third is proposed to not matter.

Interfaces cannot be constructed because they are **unsized**. They serve purely as a list of valid operations on a type within a context: no information about their memory layout is relevant. The concrete type fulfilling an interface is known at compile time, however, and so there are no issues surrounding interfaces as parameters, just when attempted to be used as (part of) a concrete type. They can be used as part of a concrete type with *indirection*, however: `type Foo = struct[a: int, b: ref interface[...]]` is perfectly valid.

Interfaces also *cannot* extend or rely upon other interfaces in any way. There is no concept of an interface extending an interface. There is no concept of a parameter satisfying two interfaces. In the author's experience, while such constructions are powerful, they are also an immense source of complexity, leading to less-than-useful interface hierarchies seen in languages like Java, and yes, Rust.

Instead, if one wishes to form an interface that *also* satisfies another interface, they must include all of the other interface's associated functions within the new interface. Given that interfaces overwhelmingly only have a handful of associated functions, and if you're using more than one interface you *really* should be using a concrete type, the hope is that this will provide explicitness.

<!-- While functions are the primary way of performing operations on types, they are not the only way, and listing all explicitly can be painful - instead, it can be desired to be able to *associate a type* and any field access or existing functions on that type with the interface. todo: i have not decided on the syntax for this yet. -->

Interfaces compose with [modules](MODULES.md) to offer fine grained access control.

<!-- todo: I have not decided whether the names of parameters is / should be relevant, or enforcable, or present. I'm leaning towards them not being present. But if they are enforcable, it makes it harder to implicitly implement the wrong interface. Design notes to consider: https://blog.rust-lang.org/2015/05/11/traits.html -->

### type aliases and distinct types

Any type can be declared as an *alias* to a type simply by assigning it to such. All functions defined on the original type carry over, and functions expecting one type may receive the other with no issues.

```puck
type Float = float
```

It is no more than an alias. When explicit conversion between types is desired and functions carrying over is undesired, `distinct` types may be used.

```puck
type MyFloat = distinct float
let foo: MyFloat = MyFloat(192.68)
```

Types then must be explicitly converted via constructors.

## Errata

### default values

Puck does not have any concept of `null`: all values *must* be initialized.
But always explicitly initializing types is syntactically verbose, and so most types have an associated "default value".

**Default values**:
- `bool`: `false`
- `int`, `uint`, etc: `0`
- `float`, etc: `0.0`
- `char`: `'\0'`
- `str`: `""`
- `void`, `never`: unconstructable
- `array[T]`, `list[T]`: `[]`
- `set[T]`, `table[T, U]`: `{}`
- `tuple[T, U, ...]`: `(default values of its fields)`
- `struct[T, U, ...]`: `{default values of its fields}`
- `enum[One, Two, ...]`: `<first label>`
- `union[T, U, ...]`: **disallowed**
- `slice[T]`, `func`: **disallowed**
- `ref`, `ptr`: **disallowed**

For unions, slices, references, and pointers, this is a bit trickier. They all have no reasonable "default" for these types *aside from* null.
Instead of giving in, the compiler instead disallows any non-initializations or other cases in which a default value would be inserted.

todo: consider user-defined defaults (ex. structs)

### subtyping

Mention of subtyping has been on occasion in contexts surrounding structural type systems, particularly the section on distinct types, but no explicit description of what the subtyping rules are have been given.

Subtyping is the implicit conversion of compatible types, usually in a one-way direction. The following types are implicitly convertible:
- `uint` ==> `int`
- `int` ==> `float`
- `uint` ==> `float`
- `string` ==> `list[char]` (the opposite no, use `pack`)
- `array[T; n]` ==> `list[T]`
- `struct[a: T, b: U, ...]` ==> `struct[a: T, b: U]`
- `union[A: T, B: U]` ==> `union[A: T, B: U, ...]`

### inheritance

Puck is not an object-oriented language. Idiomatic design patterns in object-oriented languages are harder to accomplish and not idiomatic here.

But, Puck has a number of features that somewhat support the object-oriented paradigm, including:
- uniform function call syntax
- structural typing / subtyping
- interfaces

```puck
type Building = struct
  size: struct[length, width: uint]
  color: enum[Red, Blue, Green]
  location: tuple[longitude, latitude: float]

type House = struct
  size: struct[length, width: uint]
  color: enum[Red, Blue, Green]
  location: tuple[longitude, latitude: float]
  occupant: str

func init(_: type[House]): House =
  { size: {length, width: 500}, color: Red
    location: (0.0, 0.0), occupant: "Barry" }

func address(building: Building): str =
  let number = int(building.location.0 / building.location.1).abs
  let street = "Logan Lane"
  return number.str & " " & street

# subtyping! methods!
print House.init().address()

func address(house: House): str =
  let number = int(house.location.0 - house.location.1).abs
  let street = "Logan Lane"
  return number.str & " " & street

# overriding! (will warn)
print address(House.init())

# abstract types! inheritance!
type Addressable = interface for Building
  func address(self: Self)
```

These features may *compose* into code that closely resembles its object-oriented counterpart. But make no mistake! Puck is static first and functional somewhere in there: dynamic dispatch and the like are not accessible (currently).