harmony 鸿蒙TypeScript and JavaScript Coding Style Guide

  • 2023-06-24
  • 浏览 (702)

TypeScript and JavaScript Coding Style Guide

Overview

Purpose

This guide is applicable to the scenario where TypeScript and JavaScript are used during OpenHarmony application development.

Based on the technical features of language engines and OpenHarmony, as well as industry standards and practices, this guide provides coding guide for improving code standardization, security, and performance.

Source

This guide is developed based on the JavaScript Coding Style Guide [1]. It integrates the technical features of OpenHarmony without violating industry rules, including ESLint and TSC configuration.

The correct and incorrect examples, marked with @typescript-eslint, in ESLint rules come from ESLint Rules [2].

Document Structure

All the restrictions provided in this section are raised for security purposes and therefore are mandatory.

Conventions

Rule: a convention that must be complied with

Recommendation: a convention that must be taken into consideration

It is necessary to understand the reason for each rule or recommendation and try to comply with them.

OpenHarmony Application Environment Restrictions

Use Strict Mode

[Category] Rule

[Description]

Strict mode is a way to opt in to a restricted variant of JavaScript, thereby implicitly opting out of sloppy mode. Strict mode is introduced in ECMAScript 5 and enabled by using the use strict syntax.)

  1. Strict mode eliminates some JavaScript silent errors by throwing errors.
  2. Strict mode fixes defects that make it difficult for JavaScript engines to perform optimizations. Sometimes, the same code can run faster in strict mode than in non-strict mode.
  3. Strict mode prohibits some syntax that may be defined in future versions of ECMAScript.

Note: Currently, only TypeScript or JavaScript code in strict mode is supported on OpenHarmony ArkCompiler.

Do Not Use eval()

[Category] Rule

[Description]

Using eval() causes code disorder and poor readability.

[Incorrect Example]

console.log (eval ({ a:2 })); // Output [object Object].
console.log(eval('"a" + 2')); // Output 'a2'.
console.log(eval('{a: 2}')); // Output 2.
console.log(eval('let value = 1 + 1;'); // Output undefined.

Do Not Use with() {}

[Category] Rule

[Description]

Using with() makes the code semantically unclear because the object of with() may conflict with local variables, changing the original semantics.

[Incorrect Example]

const foo = { x: 5 };
with (foo) {
  let x = 3;
  console.log(x);  // x = 3
}
console.log(foo.x);  // x = 3

Do Not Dynamically Create Functions

[Category] Rule

[Description]

A function can be defined in three ways: declaration, constructor, and expression. No matter which means is used, the function created is an instance of the Function object and inherits all default or custom methods and attributes of the Function object. Defining a function by using a constructor is similar to the using of eval(). The constructor accepts any string as the function body, causing security vulnerabilities.

[Incorrect Example]

let add = new Function('a','b','return a + b');
// The function constructor can have only one parameter, which can be any string.
let dd = new Function('alert("hello")');

[Correct Example]

// Function declaration.
function add(a,b){
  return a+b;
}
// Function expression.
let add = function(a,b){
  return a+b;
}

Declaration and Initialization

Use const or let Instead of var to Declare Variables

[Category] Rule

[Description]

ECMAScript 6 allows the use of the let and const keywords to declare variables in the block scope instead of the function scope. The block scope is commonly used in many programming languages. It is helpful in preventing errors. Use const to define read-only variables, and use let to define other variables.

[Incorrect Example]

var number = 1;
var count = 1;
if (isOK) {
  count += 1;
}

[Correct Example]

const number = 1;
let count = 1;
if (isOK) {
  count += 1;
}

Data Types

Do Not Omit 0s Before and After the Decimal Point of a Floating-Point Number

[Category] Rule

[Description]

In JavaScript, a floating-point number must contain a decimal point, but no digit is required before or after the decimal point. This makes it difficult to distinguish between the decimal point and the dot operator. For this reason, you must use a digit before and after the decimal point for a floating-point number in OpenHarmony code.

[Incorrect Example]

const num = .5;
const num = 2.;
const num = -.7;

[Correct Example]

const num = 0.5;
const num = 2.0;
const num = -0.7;

Use isNaN() to Check Whether a Variable Is NaN

[Category] Rule

[Description]

In JavaScript, NaN is a particular value of a numeric data type. It represents a non-numeric value in the double-precision 64-bit format, as defined in the IEEE floating-point standard. NaN is unique in JavaScript because it is not equal to any value, including itself. Therefore, the result of comparison with NaN is confusing, as the values of NaN !== NaN and NaN != NaN are both true. Therefore, you must use Number.isNaN() or the global isNaN() function to check whether a variable is NaN.

[Incorrect Example]

if (foo == NaN) {
  // ...
}
if (foo != NaN) {
  // ...
}

[Correct Example]

if (isNaN(foo)) {
  // ...
}
if (!isNaN(foo)) {
  // ...
}

Do Not Use == or === to Check Whether Floating-Point Numbers Are Equal

[Category] Rule

[Description]

Due to the precision problem, mathematically equal numbers may not be equal in floating-point arithmetic. Therefore, the equality operators == and === cannot be used to check whether floating-point numbers are equal.

[Incorrect Example]

0.1 + 0.2 == 0.3; // false
0.1 + 0.2 === 0.3; // false

[Correct Example]

const EPSILON = 1e-6;
const num1 = 0.1;
const num2 = 0.2;
const sum = 0.3;
if(Math.abs(num1 + num2 - sum) < EPSILON) {
  ...
}

Do Not Define or Use Non-numeric Attributes (Except Length) for Arrays

[Category] Rule

[Description]

In JavaScript, an array is an object, which can be added with attributes. To facilitate processing and avoid errors, use an array to store only a group of ordered data, that is, data with continuous indexes. If attributes must be added, use Map or Object instead.

[Incorrect Example]

const myHash = [];
myHash['key1'] = 'val1';
myHash['key2'] = 'val2';
myHash[0] = '222';
for (const key in myHash) {
  // The value of key is 0, key1, and key2.
  console.log(key);
}
console.log(myHash.length); // The array length is 1.

[Correct Example]

Use Map and Object for non-numeric attributes.

const map = new Map();
map.set('key1', 'val1');
map.set('key2', 'val2');
for(const [key, value] of map) {
  console.log('Attribute:' + key +', value: '+ value);
}

Preferentially Use Array Object Methods for Array Traversal

[Category] Rule

[Description]

To traverse an array, preferentially use the methods provided by Array, such as forEach(), map(), every(), filter(), find(), findIndex(), reduce(), and some(). Do not use for in to traverse an array.

[Incorrect Example]

const numbers = [1, 2, 3, 4, 5];
let sum = 0;
// Use for in to traverse the array.
for (const num in numbers) {
  console.log(num);
  sum += num;
}
// Use for to traverse an existing array to generate a new array.
const increasedByOne = [];
for (let i = 0; i < numbers.length; i++) {
  increasedByOne.push(numbers[i] + 1);
}

[Correct Example]

const numbers = [1, 2, 3, 4, 5];
// Use for of to traverse the array and obtain the sum.
let sum = 0;
for (const num of numbers) {
  sum += num;
}
// Use forEach to traverse the array and obtain the sum.
let sum = 0;
numbers.forEach(num => sum += num);
// Better: Use the reduce method to obtain the sum.
const sum = numbers.reduce((total, num) => total + num, 0);
// Use forEach to traverse an existing array to generate a new array.
const increasedByOne = [];
numbers.forEach(num => increasedByOne.push(num + 1));
// Better: Use the map method.
const increasedByOne = numbers.map(num => num + 1);

Operation and Expressions

Use === and !==, Instead of == or !=, to Check Whether the Operands Are Equal

[Category] Rule

[Description]

In JavaScript, type conversion is automatically performed when the == operator is used for equality judgment. For example, the values of [] == false, [] == ![], and 3 == ‘03’ are true. For higher efficiency, use === in comparison when the type is determined.

[Incorrect Example]

age == bee
foo == true
bananas != 1
value == undefined
typeof foo == 'undefined'
'hello' != 'world'
0 == 0
true == true

[Correct Example]

age === bee
foo === true
bananas !== 1
value === undefined
typeof foo === 'undefined'
'hello' !== 'world'
0 === 0
true === true

[Exceptions]

// Use the following to determine whether an object is null:
obj == null
obj != null

Do Not Assign Values in Control Conditional Expressions

[Category] Rule

[Description]

Control conditional expressions are usually used in conditional statements such as if, while, for, and ?:. Assigning values in this type of expression often leads to unexpected behavior and poor code readability.

[Incorrect Example]

// It is difficult to understand the value assignment in the control conditional expression. 
if (isFoo = false) {
  ...
}

[Correct Example]

const isFoo = someBoolean; // Assign a value above and directly use it in the if statement.
if (isFoo) {
  ...
}

Functions

Use the Same Return Statement

[Category] Rule

[Description]

Unlike statically typed languages that force a function to return a value of a specified type, JavaScript allows different code paths in a function to return different types of values.

JavaScript returns undefined in the following cases:

  1. The return statement is not executed before the exit.
  2. The return statement is executed, but no value is explicitly assigned.
  3. The return undefined statement is executed.
  4. The return void statement is executed, followed by an expression (for example, a function call).
  5. The return statement is executed, followed by another expression equivalent to undefined.

In a function (especially in a larger function), if any code path explicitly returns a value but the other code paths do not explicitly return a value, this may be a coding error. Therefore, use the same return statement in a function.

[Incorrect Example]

function doSomething(condition) {
  if (condition) {
    ...
    return true;
  } else {
    ...
    return;
  }
}
function doSomething(condition) {
  if (condition) {
    ...
    return true;
  }
}

[Correct Example]

// Ensure that all paths return values in the same way.
function doSomething(condition) {
  if (condition) {
    ...
    return true;
  } else {
    ...
    return false;
  }
}

function doSomething(condition) {
  if (condition) {
    ...
    return true;
  }
  ...
  return false;
}

Use rest Instead of arguments

[Category] Rule

[Description]

The rest parameter is an array, and all array methods, such as sort, map, forEach, and pop, can be directly used on it. arguments is a class array. Therefore, use the rest syntax instead of arguments. In addition, place the rest parameter as the last parameter in the list.

[Incorrect Example]

function concatenateAll() {
  // arguments is a class array, and the join method can be used only after arguments is converted to a real array.
  const args = Array.prototype.slice.call(arguments);   
  return args.join('');
}

[Correct Example]

function concatenateAll(...args) {
  return args.join('');
}

Do Not Assign this to a Variable; Use this in a Scope

[Category] Rule

[Description]

Arrow functions provide a simpler syntax. this in an arrow function points to the object when it is defined. Assigning this to a variable is confusing.

[Incorrect Example]

function foo() {
  const self = this;
  return function() {
    console.log(self);
  };
}

[Correct Example]

function foo() {
  return () => {
    console.log(this);
  };
}

For details, see @typescript-eslint/no-this-alias.

The description of ESLint is stricter. Therefore, do not assign this to a variable in any case.

Classes and Objects

Use Dots to Access Object Attributes and [] to Access Computed Attributes

[Category] Rule

[Description]

In JavaScript, you can use dots (foo.bar) or square brackets (foo[‘bar’]) to access attributes. The use of dots is preferred because it is more readable, concise, and more suitable for JavaScript compression.

[Correct Example]

const name = obj.name;
const key = getKeyFromDB();
const prop = obj[key]; // Use [] only when the attribute name is a variable.

Do Not Modify the Prototype of a Built-in Object or Add Methods to the Prototype

[Category] Rule

[Description]

As a set of public interfaces, built-in objects have conventional behavior modes. Any modification to the prototype of a built-in object may damage the semantics. Therefore, never modify the prototype of a built-in object, or add methods to the prototype.

[Incorrect Example]

Array.prototype.indexOf = function () {
  return -1;
}
// Used in other places.
const arr = [1, 1, 1, 1, 1, 2, 1, 1, 1];
console.log (arr.indexOf(2)); // Output -1.

Do Not Delete the Computed Attributes of an Object

[Category] Rule

[Description]

The delete operation changes the layout of an object. Deleting the computed attributes of an object is a dangerous behavior. It greatly restricts the runtime optimization and affects the execution performance.

You are advised not to delete any attribute of an object. If necessary, use map and set.

[Incorrect Example]

// Can be replaced with the constant equivalents, such as container.aaa
delete container['aaa'];

// Dynamic, difficult-to-reason-about lookups
const name = 'name';
delete container[name];
delete container[name.toUpperCase()];

[Correct Example]

// The code affects the optimization performance to some extent, but it is better than deleting the computed attributes.
delete container.aaa;

delete container[7];

For details, see @typescript-eslint/no-dynamic-delete.

Exceptions

Do Not Use return, break, continue, or throw to End the Finally Block Abnormally

[Category] Rule

[Description]

In the finally block, if the return, break, continue, or throw statement is directly used or the exception that occurs during the method calling is not handled, the finally block ends abnormally. Such a block affects the throwing of exceptions in the try or catch block, or even the return value of the method. Therefore, you must ensure that the finally block ends normally.

[Incorrect Example]

function foo() {
  try {
    ...
    return 1;
  } catch(err) {
    ...
    return 2;
  } finally {
    return 3;
 }
}

[Correct Example]

function foo() {
  try {
    ...
    return 1;
  } catch(err) {
    ...
    return 2;
  } finally {
    console.log('XXX!');
  }
}

Async Functions

Use return await Only When Necessary

[Category] Rule

[Description]

The return value of an async function is always encapsulated in Promise.resolve. The return await statement does not actually do anything, but adds time before Promise.resolve or reject. Therefore, use return await only in the try or catch statement to capture errors of another promise-based function.

[Incorrect Example]

async function foo() {
  return await bar();
}

[Correct Example]

async function foo() {
  return bar();
}
async function foo() {
  await bar();
  return;
}
async function foo() {
  const baz = await bar();
  return baz;
}
async function foo() {
  try {
    return await bar();
  } catch (error) {
    // here can be executed, go on
  }
}

Do Not Wait for a Non-thenable Value

[Category] Rule

[Description]

await converts a non-thenable value to a promise that has been processed normally and waits for the processing result. In this case, the use of await affects the code performance.

[Incorrect Example]

async function f3() {
  const y = await 20;
  console.log(y); // 20
}

f3();
console.log(30);

// output
// 30
// 20

[Correct Example]

async function f3() {
  const y = 20;
  console.log(y); // 20
}

f3();
console.log(30);

// output
// 20
// 30

For details, see @typescript-eslint/await-thenable.

Types

Forcibly Mark null and undefined as Independent Types

[Category] Rule

[Description]

Marking null and undefined as independent types improves code security and avoids null pointer exceptions.

[Incorrect Example]

let userName: string;
userName = 'hello';
userName = undefined;

[Correct Example]

let userName: string|undefined;
userName = 'hello';
userName = undefined;

Explicitly Declare the Return Value Types of Functions and Class Methods

[Category] Rule

[Description]

Explicitly declare the return value type to ensure that the return value is assigned to the variable of the correct type. In addition, the explicit declaration ensures that undefined is not assigned to a variable when there is no return value.

[Incorrect Example]

// When there is no return value, the return value type is not declared as void.
function test() {
  return;
}
// The return value type is not declared as number.
function fn() {
  return 1;
};
// The return value type is not declared as string.
let arrowFn = () => 'test';
class Test {
  // When there is no return value, the return value type is not declared as void.
  method() {
    return;
  }
}

[Correct Example]

// When there is no return value, explicitly declare the return value type as void.
function test(): void {
  return;
}
// Explicitly declare the return value type as number.
function fn(): number {
  return 1;
};
// Explicitly declare the return value type as string.
let arrowFn = (): string => 'test';
class Test {
  // When there is no return value, explicitly declare the return value type as void.
  method(): void {
    return;
  }
}

For details, see @typescript-eslint/explicit-function-return-type.

Use Consistent Export Types

[Category] Rule

[Description]

To export a type, write the type to export and the objects to export separately.

[Incorrect Example]

interface ButtonProps {
  onClick: () => void;
}
class Button implements ButtonProps {
  onClick() {
    console.log('button!');
  }
}
export { Button, ButtonProps };

[Correct Example]

interface ButtonProps {
  onClick: () => void;
}
class Button implements ButtonProps {
  onClick() {
    console.log('button!');
  }
}
export { Button };
export type { ButtonProps };

For details, see @typescript-eslint/consistent-type-exports.

Use Consistent Import Types

[Category] Rule

[Description]

To import a type, write the type to import and the objects to import separately.

[Incorrect Example]

import { Foo } from 'Foo';
import Bar from 'Bar';
type T = Foo;
const x: Bar = 1;

[Correct Example]

import type { Foo } from 'Foo';
import type Bar from 'Bar';
type T = Foo;
const x: Bar = 1;

For details, see @typescript-eslint/consistent-type-imports.

Do Not Use any

[Category] Rule

[Description]

If the any type is used, all type checks during compilation are ignored. This behavior is not necessary and does not meet expectations. If the type is unknown, use unknown. If the introduced third-party component does not use TypeScript or does not provide the TypeScript type declaration, you can use any to declare the related third-party component objects.

Do Not Define the any Type

[Category] Rule

[Description]

Do not define the any type. This restriction makes types as clear as possible in TypeScript and helps the runtime optimization.

[Incorrect Example]

const age: any = 'seventeen';
function greet(): any {}
function greet(param: Array<any>): string {}

[Correct Example]

const age: number = 17;
function greet(): string {}
function greet(param: Array<string>): string {}

For details, see @typescript-eslint/no-explicit-any.

Do Not Pass In any as a Parameter

[Category] Rule

[Incorrect Example]

declare function foo(arg1: string, arg2: number, arg3: string): void;

const anyTyped = 1 as any;

foo(...anyTyped);
foo(anyTyped, 1, 'a');

const tuple1 = ['a', anyTyped, 'b'] as const;
foo(...tuple1);

[Correct Example]

declare function foo(arg1: string, arg2: number, arg3: string): void;

foo('a', 1, 'b');

const tuple1 = ['a', 1, 'b'] as const;
foo(...tuple1);

For details, see @typescript-eslint/no-unsafe-argument.

Do Not Use any in Value Assignment

[Category] Rule

[Incorrect Example]

const x = 1 as any,

const x: Set<string> = new Set<any>();

[Correct Example]

const x = 1;

const x: Set<string> = new Set<string>();

For details, see @typescript-eslint/no-unsafe-assignment.

Do Not Call a Variable with the any Type

[Category] Rule

[Incorrect Example]

declare const anyVar: any;
declare const nestedAny: { prop: any };

anyVar();
anyVar.a.b();

nestedAny.prop();
nestedAny.prop['a']();

[Correct Example]

declare const typedVar: () => void;
declare const typedNested: { prop: { a: () => void } };

typedVar();
typedNested.prop.a();

For details, see @typescript-eslint/no-unsafe-call.

Do Not Access Members of an Object with the any Type

[Category] Rule

[Incorrect Example]

declare const anyVar: any;
declare const nestedAny: { prop: any };

anyVar.a;
anyVar.a.b;

nestedAny.prop.a;
nestedAny.prop['a'];

[Correct Example]

declare const properlyTyped: { prop: { a: string } };

properlyTyped.prop.a;
properlyTyped.prop['a'];

For details, see @typescript-eslint/no-unsafe-member-access.

Do Not Declare the Return Value Type of a Function as any or any[]

[Category] Rule

[Incorrect Example]

function foo1() {
  return 1 as any;
}

[Correct Example]

function foo1() : number {
  return 1;
}

For details, see @typescript-eslint/no-unsafe-return.

References

  1. JavaScript Coding Style Guide
  2. ESLint Rules
  3. High Performance JavaScript

你可能感兴趣的鸿蒙文章

harmony 鸿蒙FAQs

harmony 鸿蒙32- and 64-Bit Portability Coding Guide

harmony 鸿蒙Java Secure Coding Guide

harmony 鸿蒙JavaScript Coding Style Guide

harmony 鸿蒙Logging Guide

harmony 鸿蒙OpenHarmony Build Specifications

harmony 鸿蒙C Coding Style Guide

harmony 鸿蒙C&C++ Secure Coding Guide

harmony 鸿蒙OpenHarmony Compilation Specifications

harmony 鸿蒙C++ Coding Style Guide

0  赞