455 lines
19 KiB
JavaScript
455 lines
19 KiB
JavaScript
|
/**
|
||
|
|
* @fileoverview Rule to enforce concise object methods and properties.
|
||
|
|
* @author Jamund Ferguson
|
||
|
|
*/
|
||
|
|
|
||
|
|
"use strict";
|
||
|
|
|
||
|
|
const OPTIONS = {
|
||
|
|
always: "always",
|
||
|
|
never: "never",
|
||
|
|
methods: "methods",
|
||
|
|
properties: "properties",
|
||
|
|
consistent: "consistent",
|
||
|
|
consistentAsNeeded: "consistent-as-needed"
|
||
|
|
};
|
||
|
|
|
||
|
|
//------------------------------------------------------------------------------
|
||
|
|
// Requirements
|
||
|
|
//------------------------------------------------------------------------------
|
||
|
|
const astUtils = require("../ast-utils");
|
||
|
|
|
||
|
|
//------------------------------------------------------------------------------
|
||
|
|
// Rule Definition
|
||
|
|
//------------------------------------------------------------------------------
|
||
|
|
module.exports = {
|
||
|
|
meta: {
|
||
|
|
docs: {
|
||
|
|
description: "require or disallow method and property shorthand syntax for object literals",
|
||
|
|
category: "ECMAScript 6",
|
||
|
|
recommended: false
|
||
|
|
},
|
||
|
|
|
||
|
|
fixable: "code",
|
||
|
|
|
||
|
|
schema: {
|
||
|
|
anyOf: [
|
||
|
|
{
|
||
|
|
type: "array",
|
||
|
|
items: [
|
||
|
|
{
|
||
|
|
enum: ["always", "methods", "properties", "never", "consistent", "consistent-as-needed"]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
minItems: 0,
|
||
|
|
maxItems: 1
|
||
|
|
},
|
||
|
|
{
|
||
|
|
type: "array",
|
||
|
|
items: [
|
||
|
|
{
|
||
|
|
enum: ["always", "methods", "properties"]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
type: "object",
|
||
|
|
properties: {
|
||
|
|
avoidQuotes: {
|
||
|
|
type: "boolean"
|
||
|
|
}
|
||
|
|
},
|
||
|
|
additionalProperties: false
|
||
|
|
}
|
||
|
|
],
|
||
|
|
minItems: 0,
|
||
|
|
maxItems: 2
|
||
|
|
},
|
||
|
|
{
|
||
|
|
type: "array",
|
||
|
|
items: [
|
||
|
|
{
|
||
|
|
enum: ["always", "methods"]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
type: "object",
|
||
|
|
properties: {
|
||
|
|
ignoreConstructors: {
|
||
|
|
type: "boolean"
|
||
|
|
},
|
||
|
|
avoidQuotes: {
|
||
|
|
type: "boolean"
|
||
|
|
},
|
||
|
|
avoidExplicitReturnArrows: {
|
||
|
|
type: "boolean"
|
||
|
|
}
|
||
|
|
},
|
||
|
|
additionalProperties: false
|
||
|
|
}
|
||
|
|
],
|
||
|
|
minItems: 0,
|
||
|
|
maxItems: 2
|
||
|
|
}
|
||
|
|
]
|
||
|
|
}
|
||
|
|
},
|
||
|
|
|
||
|
|
create(context) {
|
||
|
|
const APPLY = context.options[0] || OPTIONS.always;
|
||
|
|
const APPLY_TO_METHODS = APPLY === OPTIONS.methods || APPLY === OPTIONS.always;
|
||
|
|
const APPLY_TO_PROPS = APPLY === OPTIONS.properties || APPLY === OPTIONS.always;
|
||
|
|
const APPLY_NEVER = APPLY === OPTIONS.never;
|
||
|
|
const APPLY_CONSISTENT = APPLY === OPTIONS.consistent;
|
||
|
|
const APPLY_CONSISTENT_AS_NEEDED = APPLY === OPTIONS.consistentAsNeeded;
|
||
|
|
|
||
|
|
const PARAMS = context.options[1] || {};
|
||
|
|
const IGNORE_CONSTRUCTORS = PARAMS.ignoreConstructors;
|
||
|
|
const AVOID_QUOTES = PARAMS.avoidQuotes;
|
||
|
|
const AVOID_EXPLICIT_RETURN_ARROWS = !!PARAMS.avoidExplicitReturnArrows;
|
||
|
|
const sourceCode = context.getSourceCode();
|
||
|
|
|
||
|
|
//--------------------------------------------------------------------------
|
||
|
|
// Helpers
|
||
|
|
//--------------------------------------------------------------------------
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Determines if the first character of the name is a capital letter.
|
||
|
|
* @param {string} name The name of the node to evaluate.
|
||
|
|
* @returns {boolean} True if the first character of the property name is a capital letter, false if not.
|
||
|
|
* @private
|
||
|
|
*/
|
||
|
|
function isConstructor(name) {
|
||
|
|
const firstChar = name.charAt(0);
|
||
|
|
|
||
|
|
return firstChar === firstChar.toUpperCase();
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Determines if the property can have a shorthand form.
|
||
|
|
* @param {ASTNode} property Property AST node
|
||
|
|
* @returns {boolean} True if the property can have a shorthand form
|
||
|
|
* @private
|
||
|
|
*
|
||
|
|
*/
|
||
|
|
function canHaveShorthand(property) {
|
||
|
|
return (property.kind !== "set" && property.kind !== "get" && property.type !== "SpreadProperty" && property.type !== "ExperimentalSpreadProperty");
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Checks whether a node is a string literal.
|
||
|
|
* @param {ASTNode} node - Any AST node.
|
||
|
|
* @returns {boolean} `true` if it is a string literal.
|
||
|
|
*/
|
||
|
|
function isStringLiteral(node) {
|
||
|
|
return node.type === "Literal" && typeof node.value === "string";
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Determines if the property is a shorthand or not.
|
||
|
|
* @param {ASTNode} property Property AST node
|
||
|
|
* @returns {boolean} True if the property is considered shorthand, false if not.
|
||
|
|
* @private
|
||
|
|
*
|
||
|
|
*/
|
||
|
|
function isShorthand(property) {
|
||
|
|
|
||
|
|
// property.method is true when `{a(){}}`.
|
||
|
|
return (property.shorthand || property.method);
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Determines if the property's key and method or value are named equally.
|
||
|
|
* @param {ASTNode} property Property AST node
|
||
|
|
* @returns {boolean} True if the key and value are named equally, false if not.
|
||
|
|
* @private
|
||
|
|
*
|
||
|
|
*/
|
||
|
|
function isRedundant(property) {
|
||
|
|
const value = property.value;
|
||
|
|
|
||
|
|
if (value.type === "FunctionExpression") {
|
||
|
|
return !value.id; // Only anonymous should be shorthand method.
|
||
|
|
}
|
||
|
|
if (value.type === "Identifier") {
|
||
|
|
return astUtils.getStaticPropertyName(property) === value.name;
|
||
|
|
}
|
||
|
|
|
||
|
|
return false;
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Ensures that an object's properties are consistently shorthand, or not shorthand at all.
|
||
|
|
* @param {ASTNode} node Property AST node
|
||
|
|
* @param {boolean} checkRedundancy Whether to check longform redundancy
|
||
|
|
* @returns {void}
|
||
|
|
*
|
||
|
|
*/
|
||
|
|
function checkConsistency(node, checkRedundancy) {
|
||
|
|
|
||
|
|
// We are excluding getters/setters and spread properties as they are considered neither longform nor shorthand.
|
||
|
|
const properties = node.properties.filter(canHaveShorthand);
|
||
|
|
|
||
|
|
// Do we still have properties left after filtering the getters and setters?
|
||
|
|
if (properties.length > 0) {
|
||
|
|
const shorthandProperties = properties.filter(isShorthand);
|
||
|
|
|
||
|
|
/*
|
||
|
|
* If we do not have an equal number of longform properties as
|
||
|
|
* shorthand properties, we are using the annotations inconsistently
|
||
|
|
*/
|
||
|
|
if (shorthandProperties.length !== properties.length) {
|
||
|
|
|
||
|
|
// We have at least 1 shorthand property
|
||
|
|
if (shorthandProperties.length > 0) {
|
||
|
|
context.report({ node, message: "Unexpected mix of shorthand and non-shorthand properties." });
|
||
|
|
} else if (checkRedundancy) {
|
||
|
|
|
||
|
|
/*
|
||
|
|
* If all properties of the object contain a method or value with a name matching it's key,
|
||
|
|
* all the keys are redundant.
|
||
|
|
*/
|
||
|
|
const canAlwaysUseShorthand = properties.every(isRedundant);
|
||
|
|
|
||
|
|
if (canAlwaysUseShorthand) {
|
||
|
|
context.report({ node, message: "Expected shorthand for all properties." });
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Fixes a FunctionExpression node by making it into a shorthand property.
|
||
|
|
* @param {SourceCodeFixer} fixer The fixer object
|
||
|
|
* @param {ASTNode} node A `Property` node that has a `FunctionExpression` or `ArrowFunctionExpression` as its value
|
||
|
|
* @returns {Object} A fix for this node
|
||
|
|
*/
|
||
|
|
function makeFunctionShorthand(fixer, node) {
|
||
|
|
const firstKeyToken = node.computed
|
||
|
|
? sourceCode.getFirstToken(node, astUtils.isOpeningBracketToken)
|
||
|
|
: sourceCode.getFirstToken(node.key);
|
||
|
|
const lastKeyToken = node.computed
|
||
|
|
? sourceCode.getFirstTokenBetween(node.key, node.value, astUtils.isClosingBracketToken)
|
||
|
|
: sourceCode.getLastToken(node.key);
|
||
|
|
const keyText = sourceCode.text.slice(firstKeyToken.range[0], lastKeyToken.range[1]);
|
||
|
|
let keyPrefix = "";
|
||
|
|
|
||
|
|
if (node.value.generator) {
|
||
|
|
keyPrefix = "*";
|
||
|
|
} else if (node.value.async) {
|
||
|
|
keyPrefix = "async ";
|
||
|
|
}
|
||
|
|
|
||
|
|
if (node.value.type === "FunctionExpression") {
|
||
|
|
const functionToken = sourceCode.getTokens(node.value).find(token => token.type === "Keyword" && token.value === "function");
|
||
|
|
const tokenBeforeParams = node.value.generator ? sourceCode.getTokenAfter(functionToken) : functionToken;
|
||
|
|
|
||
|
|
return fixer.replaceTextRange(
|
||
|
|
[firstKeyToken.range[0], node.range[1]],
|
||
|
|
keyPrefix + keyText + sourceCode.text.slice(tokenBeforeParams.range[1], node.value.range[1])
|
||
|
|
);
|
||
|
|
}
|
||
|
|
const arrowToken = sourceCode.getTokens(node.value).find(token => token.value === "=>");
|
||
|
|
const tokenBeforeArrow = sourceCode.getTokenBefore(arrowToken);
|
||
|
|
const hasParensAroundParameters = tokenBeforeArrow.type === "Punctuator" && tokenBeforeArrow.value === ")";
|
||
|
|
const oldParamText = sourceCode.text.slice(sourceCode.getFirstToken(node.value, node.value.async ? 1 : 0).range[0], tokenBeforeArrow.range[1]);
|
||
|
|
const newParamText = hasParensAroundParameters ? oldParamText : `(${oldParamText})`;
|
||
|
|
|
||
|
|
return fixer.replaceTextRange(
|
||
|
|
[firstKeyToken.range[0], node.range[1]],
|
||
|
|
keyPrefix + keyText + newParamText + sourceCode.text.slice(arrowToken.range[1], node.value.range[1])
|
||
|
|
);
|
||
|
|
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Fixes a FunctionExpression node by making it into a longform property.
|
||
|
|
* @param {SourceCodeFixer} fixer The fixer object
|
||
|
|
* @param {ASTNode} node A `Property` node that has a `FunctionExpression` as its value
|
||
|
|
* @returns {Object} A fix for this node
|
||
|
|
*/
|
||
|
|
function makeFunctionLongform(fixer, node) {
|
||
|
|
const firstKeyToken = node.computed ? sourceCode.getTokens(node).find(token => token.value === "[") : sourceCode.getFirstToken(node.key);
|
||
|
|
const lastKeyToken = node.computed ? sourceCode.getTokensBetween(node.key, node.value).find(token => token.value === "]") : sourceCode.getLastToken(node.key);
|
||
|
|
const keyText = sourceCode.text.slice(firstKeyToken.range[0], lastKeyToken.range[1]);
|
||
|
|
let functionHeader = "function";
|
||
|
|
|
||
|
|
if (node.value.generator) {
|
||
|
|
functionHeader = "function*";
|
||
|
|
} else if (node.value.async) {
|
||
|
|
functionHeader = "async function";
|
||
|
|
}
|
||
|
|
|
||
|
|
return fixer.replaceTextRange([node.range[0], lastKeyToken.range[1]], `${keyText}: ${functionHeader}`);
|
||
|
|
}
|
||
|
|
|
||
|
|
/*
|
||
|
|
* To determine whether a given arrow function has a lexical identifier (`this`, `arguments`, `super`, or `new.target`),
|
||
|
|
* create a stack of functions that define these identifiers (i.e. all functions except arrow functions) as the AST is
|
||
|
|
* traversed. Whenever a new function is encountered, create a new entry on the stack (corresponding to a different lexical
|
||
|
|
* scope of `this`), and whenever a function is exited, pop that entry off the stack. When an arrow function is entered,
|
||
|
|
* keep a reference to it on the current stack entry, and remove that reference when the arrow function is exited.
|
||
|
|
* When a lexical identifier is encountered, mark all the arrow functions on the current stack entry by adding them
|
||
|
|
* to an `arrowsWithLexicalIdentifiers` set. Any arrow function in that set will not be reported by this rule,
|
||
|
|
* because converting it into a method would change the value of one of the lexical identifiers.
|
||
|
|
*/
|
||
|
|
const lexicalScopeStack = [];
|
||
|
|
const arrowsWithLexicalIdentifiers = new WeakSet();
|
||
|
|
const argumentsIdentifiers = new WeakSet();
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Enters a function. This creates a new lexical identifier scope, so a new Set of arrow functions is pushed onto the stack.
|
||
|
|
* Also, this marks all `arguments` identifiers so that they can be detected later.
|
||
|
|
* @returns {void}
|
||
|
|
*/
|
||
|
|
function enterFunction() {
|
||
|
|
lexicalScopeStack.unshift(new Set());
|
||
|
|
context.getScope().variables.filter(variable => variable.name === "arguments").forEach(variable => {
|
||
|
|
variable.references.map(ref => ref.identifier).forEach(identifier => argumentsIdentifiers.add(identifier));
|
||
|
|
});
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Exits a function. This pops the current set of arrow functions off the lexical scope stack.
|
||
|
|
* @returns {void}
|
||
|
|
*/
|
||
|
|
function exitFunction() {
|
||
|
|
lexicalScopeStack.shift();
|
||
|
|
}
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Marks the current function as having a lexical keyword. This implies that all arrow functions
|
||
|
|
* in the current lexical scope contain a reference to this lexical keyword.
|
||
|
|
* @returns {void}
|
||
|
|
*/
|
||
|
|
function reportLexicalIdentifier() {
|
||
|
|
lexicalScopeStack[0].forEach(arrowFunction => arrowsWithLexicalIdentifiers.add(arrowFunction));
|
||
|
|
}
|
||
|
|
|
||
|
|
//--------------------------------------------------------------------------
|
||
|
|
// Public
|
||
|
|
//--------------------------------------------------------------------------
|
||
|
|
|
||
|
|
return {
|
||
|
|
Program: enterFunction,
|
||
|
|
FunctionDeclaration: enterFunction,
|
||
|
|
FunctionExpression: enterFunction,
|
||
|
|
"Program:exit": exitFunction,
|
||
|
|
"FunctionDeclaration:exit": exitFunction,
|
||
|
|
"FunctionExpression:exit": exitFunction,
|
||
|
|
|
||
|
|
ArrowFunctionExpression(node) {
|
||
|
|
lexicalScopeStack[0].add(node);
|
||
|
|
},
|
||
|
|
"ArrowFunctionExpression:exit"(node) {
|
||
|
|
lexicalScopeStack[0].delete(node);
|
||
|
|
},
|
||
|
|
|
||
|
|
ThisExpression: reportLexicalIdentifier,
|
||
|
|
Super: reportLexicalIdentifier,
|
||
|
|
MetaProperty(node) {
|
||
|
|
if (node.meta.name === "new" && node.property.name === "target") {
|
||
|
|
reportLexicalIdentifier();
|
||
|
|
}
|
||
|
|
},
|
||
|
|
Identifier(node) {
|
||
|
|
if (argumentsIdentifiers.has(node)) {
|
||
|
|
reportLexicalIdentifier();
|
||
|
|
}
|
||
|
|
},
|
||
|
|
|
||
|
|
ObjectExpression(node) {
|
||
|
|
if (APPLY_CONSISTENT) {
|
||
|
|
checkConsistency(node, false);
|
||
|
|
} else if (APPLY_CONSISTENT_AS_NEEDED) {
|
||
|
|
checkConsistency(node, true);
|
||
|
|
}
|
||
|
|
},
|
||
|
|
|
||
|
|
"Property:exit"(node) {
|
||
|
|
const isConciseProperty = node.method || node.shorthand;
|
||
|
|
|
||
|
|
// Ignore destructuring assignment
|
||
|
|
if (node.parent.type === "ObjectPattern") {
|
||
|
|
return;
|
||
|
|
}
|
||
|
|
|
||
|
|
// getters and setters are ignored
|
||
|
|
if (node.kind === "get" || node.kind === "set") {
|
||
|
|
return;
|
||
|
|
}
|
||
|
|
|
||
|
|
// only computed methods can fail the following checks
|
||
|
|
if (node.computed && node.value.type !== "FunctionExpression" && node.value.type !== "ArrowFunctionExpression") {
|
||
|
|
return;
|
||
|
|
}
|
||
|
|
|
||
|
|
//--------------------------------------------------------------
|
||
|
|
// Checks for property/method shorthand.
|
||
|
|
if (isConciseProperty) {
|
||
|
|
if (node.method && (APPLY_NEVER || AVOID_QUOTES && isStringLiteral(node.key))) {
|
||
|
|
const message = APPLY_NEVER ? "Expected longform method syntax." : "Expected longform method syntax for string literal keys.";
|
||
|
|
|
||
|
|
// { x() {} } should be written as { x: function() {} }
|
||
|
|
context.report({
|
||
|
|
node,
|
||
|
|
message,
|
||
|
|
fix: fixer => makeFunctionLongform(fixer, node)
|
||
|
|
});
|
||
|
|
} else if (APPLY_NEVER) {
|
||
|
|
|
||
|
|
// { x } should be written as { x: x }
|
||
|
|
context.report({
|
||
|
|
node,
|
||
|
|
message: "Expected longform property syntax.",
|
||
|
|
fix: fixer => fixer.insertTextAfter(node.key, `: ${node.key.name}`)
|
||
|
|
});
|
||
|
|
}
|
||
|
|
} else if (APPLY_TO_METHODS && !node.value.id && (node.value.type === "FunctionExpression" || node.value.type === "ArrowFunctionExpression")) {
|
||
|
|
if (IGNORE_CONSTRUCTORS && node.key.type === "Identifier" && isConstructor(node.key.name)) {
|
||
|
|
return;
|
||
|
|
}
|
||
|
|
if (AVOID_QUOTES && isStringLiteral(node.key)) {
|
||
|
|
return;
|
||
|
|
}
|
||
|
|
|
||
|
|
// {[x]: function(){}} should be written as {[x]() {}}
|
||
|
|
if (node.value.type === "FunctionExpression" ||
|
||
|
|
node.value.type === "ArrowFunctionExpression" &&
|
||
|
|
node.value.body.type === "BlockStatement" &&
|
||
|
|
AVOID_EXPLICIT_RETURN_ARROWS &&
|
||
|
|
!arrowsWithLexicalIdentifiers.has(node.value)
|
||
|
|
) {
|
||
|
|
context.report({
|
||
|
|
node,
|
||
|
|
message: "Expected method shorthand.",
|
||
|
|
fix: fixer => makeFunctionShorthand(fixer, node)
|
||
|
|
});
|
||
|
|
}
|
||
|
|
} else if (node.value.type === "Identifier" && node.key.name === node.value.name && APPLY_TO_PROPS) {
|
||
|
|
|
||
|
|
// {x: x} should be written as {x}
|
||
|
|
context.report({
|
||
|
|
node,
|
||
|
|
message: "Expected property shorthand.",
|
||
|
|
fix(fixer) {
|
||
|
|
return fixer.replaceText(node, node.value.name);
|
||
|
|
}
|
||
|
|
});
|
||
|
|
} else if (node.value.type === "Identifier" && node.key.type === "Literal" && node.key.value === node.value.name && APPLY_TO_PROPS) {
|
||
|
|
if (AVOID_QUOTES) {
|
||
|
|
return;
|
||
|
|
}
|
||
|
|
|
||
|
|
// {"x": x} should be written as {x}
|
||
|
|
context.report({
|
||
|
|
node,
|
||
|
|
message: "Expected property shorthand.",
|
||
|
|
fix(fixer) {
|
||
|
|
return fixer.replaceText(node, node.value.name);
|
||
|
|
}
|
||
|
|
});
|
||
|
|
}
|
||
|
|
}
|
||
|
|
};
|
||
|
|
}
|
||
|
|
};
|