function-type-expressions（函数类型表达式）

Apply Types to Entire Function Expressions When Possible

Overview

Type entire functions at once instead of individual parameters.

When using function expressions (not statements), you can apply a type to the entire function. This reduces repetition, improves type safety, and makes code more readable.

When to Use This Skill

• Writing multiple functions with the same signature
• Implementing callbacks for libraries
• Matching signatures of existing functions
• Wrapping or extending existing functions

The Iron Rule

When functions share a signature, define the type ONCE and apply it to EACH function.

Remember:

• Function expressions can have types applied to them
• Parameter types are inferred from the function type
• Return types are checked against the function type
• Use typeof fn to match existing function signatures

Detection: Repeated Signatures

// Repetitive - same signature 4 times
function add(a: number, b: number) { return a + b; }
function sub(a: number, b: number) { return a - b; }
function mul(a: number, b: number) { return a * b; }
function div(a: number, b: number) { return a / b; }

The Solution: Function Types

Define Once, Use Many Times

type BinaryFn = (a: number, b: number) => number;

const add: BinaryFn = (a, b) => a + b;  // Types inferred
const sub: BinaryFn = (a, b) => a - b;
const mul: BinaryFn = (a, b) => a * b;
const div: BinaryFn = (a, b) => a / b;

Benefits:

• No repeated type annotations
• Return type checked automatically
• Logic is more visible without type noise

Match Existing Functions with typeof

// Match fetch's signature exactly
const checkedFetch: typeof fetch = async (input, init) => {
  const response = await fetch(input, init);
  if (!response.ok) {
    throw new Error(`Request failed: ${response.status}`);
  }
  return response;
};

TypeScript infers:

• input: RequestInfo | URL
• init?: RequestInit
• Return: Promise<Response>

Change Return Type with Parameters

// Match parameters but change return type
async function fetchNumber(
  ...args: Parameters<typeof fetch>
): Promise<number> {
  const response = await checkedFetch(...args);
  return Number(await response.text());
}

Function Statement vs Expression

// Statement - must annotate each parameter
function rollDice1(sides: number): number { /* ... */ }

// Expression - can apply type to entire function
type DiceRollFn = (sides: number) => number;
const rollDice2: DiceRollFn = (sides) => { /* ... */ };

Common Function Type Patterns

Callback Types

type EventHandler = (event: Event) => void;
type AsyncCallback<T> = () => Promise<T>;
type Comparator<T> = (a: T, b: T) => number;

const handleClick: EventHandler = (e) => {
  console.log(e.target);  // e is typed as Event
};

Generic Function Types

type Mapper<T, U> = (item: T, index: number) => U;

const double: Mapper<number, number> = (n) => n * 2;
const stringify: Mapper<number, string> = (n) => String(n);

Interface Syntax (Alternative)

// Function type as interface
interface StringTransform {
  (input: string): string;
}

const toUpper: StringTransform = s => s.toUpperCase();

Return Type Safety

Function types catch return type errors:

const checkedFetch: typeof fetch = async (input, init) => {
  const response = await fetch(input, init);
  if (!response.ok) {
    return new Error('Failed');  // Error!
    // Type 'Error' is not assignable to type 'Response'
  }
  return response;
};

Without the function type, this would only error at call sites.

Library Callback Types

Libraries often provide callback types:

// React provides these
import { MouseEventHandler, ChangeEventHandler } from 'react';

const handleClick: MouseEventHandler<HTMLButtonElement> = (e) => {
  console.log(e.currentTarget.disabled);  // Fully typed
};

const handleChange: ChangeEventHandler<HTMLInputElement> = (e) => {
  console.log(e.target.value);  // Fully typed
};

When NOT to Use Function Types

Don't over-engineer for single functions:

// Overkill for a single standalone function
type GreetFn = (name: string) => string;
const greet: GreetFn = (name) => `Hello, ${name}`;

// Just use a normal function statement
function greet(name: string): string {
  return `Hello, ${name}`;
}

Pressure Resistance Protocol

1. "Function Statements Are Clearer"

Pressure: "I prefer seeing types inline with parameters"

Response: With shared signatures, centralizing the type removes noise.

Action: Use function types when 2+ functions share a signature.

2. "I Don't Know the Library's Type"

Pressure: "I can't find the callback type in the library"

Response: Use typeof existingFunction or Parameters<typeof fn>.

Action: Match existing signatures with typeof.

Red Flags - STOP and Reconsider

• Same parameter types repeated across multiple functions
• Wrapper functions that should match the wrapped function's signature
• Callbacks without proper typing

Common Rationalizations (All Invalid)

Quick Reference

// Define function type
type Transform<T> = (value: T) => T;

// Apply to function expression
const double: Transform<number> = v => v * 2;

// Match existing function
const myFetch: typeof fetch = async (input, init) => { ... };

// Match parameters, change return
function wrapper(...args: Parameters<typeof original>): NewReturn { ... }

The Bottom Line

Apply types to entire function expressions when you have shared signatures.

This reduces repetition, centralizes type definitions, and catches return type errors at the source. Use typeof to match existing function signatures exactly.

Reference

Based on "Effective TypeScript" by Dan Vanderkam, Item 12: Apply Types to Entire Function Expressions When Possible.