rfcs/docs/user-defined-type-functions.md

User-Defined Type Functions

Summary

The new type solver for Luau supports a limited number of built-in type functions that enable developers to be expressive with their code. To further enhance this expressivity and the flexibility of the type system, this RFC proposes the design and implementation of user-defined type functions, a feature that will allow developers to define custom type functions.

Motivation

The primary motivation for introducing user-defined type functions is to extend the flexibility of the type system and empower developers to create more expressive and type-safe libraries. Current limitations in the type system restrict custom type manipulations, which are crucial for creating advanced patterns and abstractions. By supporting user-defined type functions, we aim to:

• Enable more precise type definitions that can capture complex relationships and constraints
• Facilitate the creation of reusable and composable libraries
• Enhance code safety and maintainability with better type-level programming

The expected outcome is a more versatile type system that can adapt to a wider range of programming patterns, ultimately improving the Luau developer experience.

Design

Type functions wil be defined with the following syntax:

type function f(...)
    -- implementation of the type function
end

For instance, the rawget type function can be written as:

type function rawget(tbl, prop)
    print("First parameter should be a table") -- produces a warning
    if (~tbl.istable()) then
        error("First parameter of rawget is not a table!") -- fails to reduce
    end

    return tbl.getProps()[prop.getValue()]
end

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

type ty = rawget<Person, "name"> -- ty = string

Type functions operate on two stages: type analysis and runtime. When calling type functions at the type level (e.g. annotating type of type function to a variable), angle brackets must be used, but when calling the at the runtime level (e.g. calling other type functions), parenthesis must be used. Declarations of type functions use parenthesis because it defines the runtime operations on the runtime representation of types.

For the current being (and until we find a better solution), type functions will not be able to refer other type functions, aliases, anything else that is not in the local scope, mainly due to the fact that recursive calls can run for arbitrary amount of time that can halt analysis from making further progress. For example, reducing this type function will recurse until stack overflow and cause the whole analysis to crash:

type function neverending(t)
    return neverending(t) -- note: we use parenthesis here because we are passing in the runtime value of types, not static values
end

We have considered adding an user-configured execution limit based on time or instruction count for reducing type functions where if a type function does not finish executing under the constraint, it fails to reduce. However, time-based timeouts are dependent on CPU; programs that type check on fast CPUs may not type check on slower CPUs. Similarly, instruction-based timeouts are dependent on the compiler version; programs that type check on versions of compilers where they produce less instructions may not type check on compilers that do not carry the same optimizations. We will not (and probably never) allow type functions to call regular functions for the sake of sandboxing the runtime and analysis.

To give warnings, developers can use print() with custom warning messages. To fail reductions, developers can use error() with custom error messages. If nothing is returned by the type function, it will fail to reduce with the default message: "Failed to reduce <Name> type function with no return values".

To allow Luau developers to modify the runtime values of types in type functions, this RFC proposes introducing a new userdata called typelib. An typelib object is a runtime representation of all types within the program and provides a basic set of library methods that can be used to modify types. As such, under the hood, each argument of a type function is serialized into a userdata called typelib. Most importantly, they are only accessible within type functions and are not a runtime type for other use cases than type functions.

typelib library methods (dropdown)

Note: methods under a different type heading (ex: `Singleton`) imply that the methods are only available for those types. At the implementation level, there is a check to make sure that the type-specific methods are being called on the correct types (e.g, for `getIndexer()`, assert that `isTable()` is true).

Any

Function Declaration	Return Type	Description
isstring()	boolean	returns true if self is of type string
isnumber()	boolean	returns true if self is of type number
isboolean()	boolean	returns true if self is of type boolean (e.g. true or false)
istable()	boolean	returns true if self is of type table
isthread()	boolean	returns true if self is of type thread
isfunction()	boolean	returns true if self is of type function
isbuffer()	boolean	returns true if self is of type buffer
isnil()	boolean	returns true if self is of type nil
isclass()	boolean	returns true if self is of type class (do we need this?)
isbooleansingleton()	boolean	returns true if self is a boolean singleton
isstringsingleton()	boolean	returns true if self is a string singleton
isa(arg: typelib)	boolean	returns true if arg is the same type as self

Primitive

Function Declaration	Return Type	Description
gettype()	string	returns either "nil", "boolean", "string", "thread", "function", "table", or "buffer"

Singleton

Function Declaration	Return Type	Description
getvalue()	string	returns either "true", "false", or a string singleton

Table

Function Declaration	Return Type	Description
getprops()	table	returns a type representation of tables (e.g. {name = "John"} will return {[string] = "string"})
getindexer()	table	returns a type representation of arrays (e.g. {1, "hi", 3} will return {[number] = "number" | "string"})

Class

Function Declaration	Return Type	Description
getparent()	typelib	returns typelib userdata of self's parent

Implementation

The implementation of user-defined type functions can be broken down into the following parts:

AST: Introduce a new AST node for type functions called AstStatTypeFunction.

Parser: When parsing TypeAliases, if the next lexeme is "function", parse it into the AstStatTypeFunction node.

Constraint Generator: Generate a TypeAliasExpansionConstraint when visiting the AstStatTypeFunction node.

Constraint Solver: Generate a new ReduceConstraint and append it to the list of unsolved constraints when reducing TypeAliasExpansionConstraint.

Reduction Constraint: Serialize the function arguments into typelib and execute the body of the function. Reduce to the deserialized version of the function return value.

typelib: Using the Lua API, create a library that interfaces between C++ and Luau and supports all of the function calls, serialization, and deserialization.

Drawback

Type functions are handled at the analysis time, while typelib is an implementation in the runtime. As a result, the proposed design causes the analysis time to be dependent on Luau's runtime to reduce user-defined type functions. This is generally discouraged as it is best to isolate the compile time, analysis time, and runtime from each other for the purpose of maintaining a clean separation of concerns, which helps minimize side effects and dependencies across different phases of the program execution and improves the modularity of the compiler. Overlaps between the analysis time and runtime can lead to code that is more complex and harder to manage, potentially increasing the risk of bugs and making the outcomes less predictable.

The build / analysis times will also be negatively impacted by this feature as reducing type functions takes variable amount of time based on the program. The larger the type function, the longer it will take to reduce it in the constraint solver.

Alternatives

table Runtime Representation

Currently, the runtime representation of types is a userdata called typelib. Another representation is to use the already-existing type in Luau table; instead of serializing types into typelib, we can serialize them into tables with predefined properties. For instance, the representation for a string singleton "abc" could be {type = "stringSingleton", value = "abc"}. So instead of writing:

type function isSingleton(t)
    if t:isStringSingleton() then
        return t
    end
end

developers could write:

type function isSingleton(t)
    if t.type == "stringSingleton" then
        return t
    end
end

In some sense, this design could be considered "cleaner" than introducing an entirely new userdata, but it requires developers to have a deeper understanding of the type runtime representation. For example, under the proposed design, a new table type can be created using typelib.new("table"), while under the alternative design, tables must be declared with attributes: {type = "table", props = {}, indexer = {}}. This adds complexity to the developer experience and increases the chance of making syntax errors in their program.

More Builtin Type Functions

Another alternative is to simply let go of this idea and create more built-in type functions that cover a wider range of common use cases. This approach clearly lacks the flexibility of fully user-defined type functions because type manipulation would be limited to the predefined set of type functions designed by the Luau team. Furthermore, continuously expanding the set of built-in type functions could lead to bloat and complexity within the core language, making it harder to maintain.

Type Function Type Checking

In the future, we could investigate adding type checking to user defined type functions. In Haskell's type families, this is done with kinds (aka types of types). Similarly, we could look into introducing kinds to Luau, but for the purpose of this RFC, this feature will be left for future works.