Variadic Attributes

Introduction

This proposal introduces variadic attributes to Zirric, enabling attributes and functions to accept a variable number of arguments. This feature addresses the need for multi-instance constraints (e.g., @Requires(Countable, Iterable)) while maintaining a clean and expressive syntax.

Variadic attributes are particularly useful for:

• Representing requirements or constraints on fields, parameters, or types.
• Improving readability and reducing verbosity compared to alternatives like @HasAll([Countable, Iterable]).

Motivation

Zirric currently lacks a concise way to specify multiple constraints or requirements in attributes. For example, to enforce that a parameter must satisfy both Countable and Iterable, users must use workarounds like @HasAll([Countable, Iterable]), which is less readable and feels unnatural.

This could be useful in scenarios such as:

1. Type Constraints: Annotating parameters or fields to require multiple traits or interfaces.

   fn process(@Requires(Countable, Iterable) data) {}
   

2. Tooling and Compile-Time Checks: Enabling tools (e.g., LSP, compilers) to validate that values meet multiple constraints.
3. Library Design: Allowing library authors to define flexible APIs that accept variadic constraints.

Benefits

• Readability: @Requires(Countable, Iterable) is more intuitive than @HasAll([Countable, Iterable]).
• Flexibility: Supports empty lists, optional constraints, and future extensibility.
• Consistency: Aligns with variadic functions in other languages, making it familiar to developers.

---

Proposed Solution

Syntax

Introduce a @Variadic attribute to mark attribute fields as accepting a variable number of arguments. When used, the field’s type is treated as an Array of the specified element type.

Example: Defining a Variadic Attribute

attr Requires {
  @Variadic()
  @ArgType(Type)  // Accepts types or type-like attributes
  constraints
}

Example: Using a Variadic Attribute

// Function with variadic constraints
fn process(@Requires(Countable, Iterable) data) {
  // `data` must satisfy both `Countable` and `Iterable`
}

// Valid calls
process([1, 2, 3])  // Array of Int (assuming Int is Countable and Iterable)
process([])         // Empty array

// Invalid call (if String doesn't satisfy Iterable)
process("hello")    // Compiler error: String is not Iterable

Behavior

• Type Handling: Variadic arguments are always treated as an Array of the specified element type (e.g., @ArgType(Type)).
• Empty Lists: Supported for optional constraints (e.g., @Requires()).
• Type Safety: The compiler enforces that all provided arguments match the element type (e.g., Type).

---

Detailed Design

Attribute Definition

• The @Variadic attribute marks a field within another attribute as variadic.
• The field’s type is implicitly an Array of the type specified by @ArgType.

  attr VariadicExample {
    @Variadic()
    @ArgType(Int)  // Accepts variadic Int arguments
    numbers
  }
  

Usage in Functions

• Variadic attributes can be applied to function parameters, fields, or types.

  fn example(@VariadicExample(1, 2, 3) nums) {}
  

• Functions themselves can also have a variadic parameter syntax similar to attributes.

  fn sum(@Variadic() @ArgType(Int) numbers) {
    // `numbers` is an Array of Int
  }
  sum(1, 2, 3, 4)  // Valid call
  

Compiler and Tooling Support

• Parsing: The compiler treats variadic attribute fields as arrays during parsing.
• Type Checking: Ensures all arguments match the element type specified by @ArgType.
• Error Messages: Provides clear errors for type mismatches or invalid usage.

Edge Cases

• Empty Lists: @Requires() is valid and results in an empty array of constraints.
• Single Argument: @Requires(Countable) is treated as an array with one element.
• Type Mismatches: The compiler rejects arguments that don’t match @ArgType.

Changes to the Standard Library

New Attributes

1. @Variadic:

   • Marks a field as accepting a variable number of arguments.
   • Only valid within attribute definitions.

2. @ArgType:

   • Specifies the type of elements in a variadic field (e.g., @ArgType(Type)).
   • Required for variadic fields to enforce type safety.

Example Additions

// Standard library addition
attr Variadic {}
attr ArgType {
  type
}

Alternatives Considered

1. @HasAll([...]) Syntax:

   • Pros: No new syntax required.
   • Cons: Less readable and feels like a workaround.

2. Dedicated Grouping Syntax (e.g., @Requires{Countable, Iterable}):

   • Pros: Clean and intuitive.
   • Cons: Doesn’t translate to function arguments and requires new grammar rules.

3. Multiple Attributes (e.g., @Requires(Countable) @Requires(Iterable)):

   • Pros: Simple and explicit.
   • Cons: Verbose and ambiguous for single-instance attributes.

Why @Variadic Was Chosen:

• Balances flexibility and readability.
• Works seamlessly with function arguments and fields.
• Aligns with Zirric’s existing attribute system.

Acknowledgements

• Prior Art: Inspired by variadic functions in languages like C, JavaScript, and Rust.
• Contributors: @vknabel for designing the proposal.

Open Questions

1. Should there be spreads to pass arrays as variadic arguments (e.g., @Requires(...myArray))?
2. Should @Variadic require an optional min or max parameter or a range to enforce bounds on the number of arguments?