rfcs/docs/index-type-operator.md

index type operators

Summary

This RFC proposes the addition of one type operator, index, which can be used to look up a specific property of another type (like TypeScript's Indexed Access Type).

Motivation

The primary motivation of this proposal is to allow Luau to define a type by accessing a property of another existing type. For instance, consider the following example code:

type Person = {
  age: number,
  name: string,
  alive: boolean
}

local bob: Person = {
  age = 22,
  name = "Bob",
  alive = true
}

local function doSmt(param: typeof(bob["age"])) -- param = string
  -- rest of code
end

type unionType = typeof(bob["age"]) | typeof(bob["name"]) | typeof(bob["alive"]) -- unionType = number | string | boolean

This is a valid Luau program; however, in order to define the type of Person["age"] we had to first declare a variable bob and utilize the typeof type operator. This is quite cumbersome when developers want to typecheck using the type of Person["age"] without having to declare a variable first. Additionally, in order to define the union type of all the properties of Person, current Luau requires an explicit list of each property using typeof.

The expected outcome of the index type operator is that it will enhance developer experience and allow Luau developers to more easily develop well-typed programs.

Design

The solution to this problem is a type operator, index, that can compute the type based on the static properties of Person. Formally, the index type operator will take in two arguments: the type to index and the type to index it with. This would allow us to instead write the following code:

type Person = {
  age: number,
  name: string,
  alive: boolean
}

local function doSmt(param: index<Person, "age">) -- param = number
  -- rest of code
end

type idxType = index<Person, keyof<Person>> -- idxType = number | string | boolean

type idxType2 = index<Person, "age" | "name"> -- idxType2 = number | string

Now, the type of doSmt()'s parameter can be defined without declaring a variable bob. Additionally, regardless of how the type Person grows, idxType will always be defined as the union of all the properties.

Error messages will be displayed for incorrect syntax. If the given type used to access a property is invalid,

type age = index<Person, "ager"> -- Error message: Property 'ager' does not exist on type 'Person'.

If the given value used to access a property is not a type,

local key = "age"
type age = index<Person, key> -- Error message: Type 'key' cannot be used as an index type.

One edge case for this feature is in the case of:

type Person2 = {
  age: number,
  name: string,
  alive: boolean,
  job: string
}

local function edgeCase(p: Person)
  type unknownType = index<typeof(p), "job"> -- unknownType = unknown
end

Although the literal type job is not a property in Person, the program should not error here, but instead return the type unknown. This is due to the fact that at run time, parameter p can be of type Person or Person2 (since Person2 is a subtype of Person). As a result, the index type of job on Person throws an error, but on Person2 returns a type of string. As a result, at compile time, the type of unknownType should be unknown.

Implementation is straight forward: the type of the indexee will be determined (table, array, etc) -> search through the properties of the indexee and return the corresponding type of the indexer if it exists; otherwise, return an error or unknown type depending on the scope.

Drawbacks

The only drawback to implementing this feature is that supporting type operators will require a lot of work in the underlying infrastructure. However, the new type inference engine for Luau already contains the infrastructure needed to support type operators. In fact, the implementation of this new feature can be mirrored by the implementation of the keyof type operator as they share the same syntax and a small difference in semantics. As such, the author of this RFC hopes that this effort is a win for OOP in Luau, supported by the technical credit built up from the new type inference engine.

Alternatives

An alternative to the proposed implementation is utilizing inexact table types and exact table types (to be implemented) to handle the edge in a simpler manner. If we know that the indexee is an inexact table type, the index type operator can return the unknown type. If we know that the indexee is an exact table type, we can return an error.