基于AST构建JavaScript元循环求值器详解
本文深入解析如何使用抽象语法树(AST)构建一个JavaScript元循环求值器。通过拆解Tokenizer、Parser、Evaluate和Apply四个核心模块,重点阐述递归计算过程中的evaluate-apply机制以及Scope作用域类的变量声明与查找逻辑。文章提供了基于@babel/parser的代码结构示例,并展示了如何处理var、let、const的不同作用域行为,帮助开发者理解解释器内部工作原理。
引言与背景
在上一篇关于微信小程序组件多端编译的文章中,我们利用AST完成了相关转换。本文将进一步深入,探讨如何利用AST构建一个JavaScript元循环求值器。
元循环求值器结构
结构
一个完整的元循环求值器通常包含以下核心组件:
- Tokenizer:负责代码文本的词法与语法分析,将源代码分割为多个Token。
- Parser:依据Token序列生成抽象语法树(AST)。
- Evaluate:根据AST节点的类型,调用对应的apply方法进行处理。
- Apply:基于当前环境执行实际的求值计算。
- Scope:定义代码执行时的上下文环境。
代码目录结构
代码目录
根据上述结构设计,我们将代码目录大致划分为以下文件模块:
- Parser
- Eval
- Scope
Tokenize和Parser两个步骤并非本文重点,因此统一交由@babel/parser处理。Evaluate和Apply过程合并至eval文件,后续将重点分析该部分。Scope模块则独立存放于scope文件中。
Evaluate-Apply机制详解
Evaluate-Apply机制
这本质上是一个递归计算的过程。
首先,evaluate函数接收两个参数:当前遍历的AST节点node和当前环境scope。随后,evaluate根据node的type属性判断节点类型,并执行apply方法来求值该节点代表的表达式。在apply内部,会再次递归调用evaluate以计算当前节点的子节点,直至完整遍历整棵AST树。
具体代码实现如下:
const evaluate = (node: t.Node, scope) => {
const evalFunc = evaluateMap[node.type];
if (!evalFunc) {
throw `${node.loc} ${node.type} 还未实现`;
}
return evalFunc(node, scope);
}
上述即为evaluate函数的主要逻辑。
其中,evaluateMap包含了目前已实现的功能集合,具体代码见下:
const evaluateMap: EvaluateMap = {
File(node: t.File, scope) {
evaluate(node.program, scope);
},
Program(node: t.Program, scope) {
for (const n of node.body) {
evaluate(n, scope);
}
},
Identifier(node: t.Identifier, scope) {
const $var = scope.$find(node.name);
if (!$var) {
throw `[Error] ${node.loc}, '${node.name}' 未定义`;
}
return $var.$get();
},
StringLiteral(node: t.StringLiteral, scope) {
return node.value;
},
NumericLiteral(node: t.NumericLiteral, scope) {
return node.value;
},
BooleanLiteral(node: t.BooleanLiteral, scope) {
return node.value;
},
NullLiteral(node: t.NullLiteral, scope) {
return null;
},
BlockStatement(block: t.BlockStatement, scope) {
const blockScope = scope.shared ? scope : new Scope('block', scope);
for (const node of block.body) {
const res = evaluate(node, blockScope);
if (res === BREAK || res === CONTINUE || res === RETURN) {
return res;
}
}
},
DebuggerStatement(node: t.DebuggerStatement, scope) {
debugger;
},
ExpressionStatement(node: t.ExpressionStatement, scope) {
evaluate(node.expression, scope);
},
ReturnStatement(node: t.ReturnStatement, scope) {
RETURN.result = (node.argument ? evaluate(node.argument, scope) : void 0);
return RETURN;
},
BreakStatement(node: t.BreakStatement, scope) {
return BREAK;
},
ContinueStatement(node: t.ContinueStatement, scope) {
return CONTINUE;
},
IfStatement(node: t.IfStatement, scope) {
if (evaluate(node.test, scope)) {
return evaluate(node.consequent, scope);
}
if (node.alternate) {
const ifScope = new Scope('block', scope, true);
return evaluate(node.alternate, ifScope)
}
},
SwitchStatement(node: t.SwitchStatement, scope) {
const discriminant = evaluate(node.discriminant, scope);
const switchScope = new Scope('switch', scope);
for (const ca of node.cases){
if (ca.test === null || evaluate(ca.test, switchScope) === discriminant) {
const res = evaluate(ca, switchScope);
if (res === BREAK) {
break;
} else if (res === RETURN) {
return res;
}
}
}
},
SwitchCase(node: t.SwitchCase, scope) {
for (const item of node.consequent) {
const res = evaluate(item, scope);
if (res === BREAK || res === RETURN) {
return res;
}
}
},
ThrowStatement(node: t.ThrowStatement, scope) {
throw evaluate(node.argument, scope);
},
TryStatement(node: t.TryStatement, scope) {
try {
return evaluate(node.block, scope);
} catch (error) {
if (node.handler) {
const catchScope = new Scope('block', scope, true);
catchScope.$let((<t.Identifier>node.handler.param).name, error);
return evaluate(node.handler, catchScope);
} else {
throw error;
}
} finally {
if (node.finalizer) {
return evaluate(node.finalizer, scope);
}
}
},
CatchClause(node: t.CatchClause, scope) {
return evaluate(node.body, scope);
},
WhileStatement(node: t.WhileStatement, scope) {
while (evaluate(node.test, scope)) {
const whileScope = new Scope('loop', scope, true);
const res = evaluate(node.body, whileScope);
if (res === CONTINUE) continue;
if (res === BREAK) break;
if (res === RETURN) return res;
}
},
ForStatement(node: t.ForStatement, scope) {
for (
const forScope = new Scope('loop', scope),
initVal = evaluate(node.init, forScope);
evaluate(node.test, forScope);
evaluate(node.update, forScope)
) {
const res = evaluate(node.body, forScope);
if (res === CONTINUE) continue;
if (res === BREAK) break;
if (res === RETURN) return res;
}
},
ForInStatement(node: t.ForInStatement, scope) {
const kind = (<t.VariableDeclaration>node.left).kind;
const decl = (<t.VariableDeclaration>node.left).declarations[0];
const name = (<t.Identifier>decl.id).name;
for (const value in evaluate(node.right, scope)) {
const forScope = new Scope('loop', scope, true);
scope.$define(kind, name, value);
const res = evaluate(node.body, forScope);
if (res === CONTINUE) continue;
if (res === BREAK) break;
if (res === RETURN) return res;
}
},
ForOfStatement(node: t.ForOfStatement, scope) {
const kind = (<t.VariableDeclaration>node.left).kind;
const decl = (<t.VariableDeclaration>node.left).declarations[0];
const name = (<t.Identifier>decl.id).name;
for (const value of evaluate(node.right, scope)) {
const forScope = new Scope('loop', scope, true);
scope.$define(kind, name, value);
const res = evaluate(node.body, forScope);
if (res === CONTINUE) continue;
if (res === BREAK) break;
if (res === RETURN) return res;
}
},
FunctionDeclaration(node: t.FunctionDeclaration, scope) {
const func = evaluateMap.FunctionExpression(node, scope);
scope.$var(node.id.name, func);
},
VariableDeclaration(node: t.VariableDeclaration, scope) {
const { kind, declarations } = node;
for (const decl of declarations) {
const varName = (<t.Identifier>decl.id).name;
const value = decl.init ? evaluate(decl.init, scope) : void 0;
if (!scope.$define(kind, varName, value)) {
throw `[Error] ${name} 重复定义`
}
}
},
ThisExpression(node: t.ThisExpression, scope) {
const _this = scope.$find('this');
return _this ? _this.$get() : null;
},
ArrayExpression(node: t.ArrayExpression, scope) {
return node.elements.map(item => evaluate(item, scope));
},
ObjectExpression(node: t.ObjectExpression, scope) {
let res = Object.create(null);
node.properties.forEach((prop) => {
let key;
let value;
if(prop.type === 'ObjectProperty'){
key = prop.key.name;
value = evaluate(prop.value, scope);
res[key] = value;
}else if (prop.type === 'ObjectMethod'){
const kind = prop.kind;
key = prop.key.name;
value = evaluate(prop.body, scope);
if(kind === 'method') {
res[key] = value;
}else if(kind === 'get') {
Object.defineProperty(res, key, { get: value });
}else if(kind === 'set') {
Object.defineProperty(res, key, { set: value });
}
}else if(prop.type === 'SpreadElement'){
const arg = evaluate(prop.argument, scope);
res = Object.assign(res, arg);
}
});
return res;
},
FunctionExpression(node: t.FunctionExpression, scope) {
return function (...args: any) {
const funcScope = new Scope('function', scope, true);
node.params.forEach((param: t.Identifier, idx) => {
const { name: paramName } = param;
funcScope.$let(paramName, args[idx]);
});
funcScope.$const('this', this);
funcScope.$const('arguments', arguments);
const res = evaluate(node.body, funcScope);
if (res === RETURN) {
return res.result;
}
}
},
ArrowFunctionExpression(node: t.ArrowFunctionExpression, scope) {
return (...args) => {
const funcScope = new Scope('function', scope, true);
node.params.forEach((param: t.Identifier, idx) => {
const { name: paramName } = param;
funcScope.$let(paramName, args[idx]);
});
const _this = funcScope.$find('this');
funcScope.$const('this', _this ? _this.$get() : null);
funcScope.$const('arguments', args);
const res = evaluate(node.body, funcScope);
if (res === RETURN) {
return res.result;
}
}
},
UnaryExpression(node: t.UnaryExpression, scope) {
const expressionMap = {
'~': () => ~evaluate(node.argument, scope),
'+': () => +evaluate(node.argument, scope),
'-': () => -evaluate(node.argument, scope),
'!': () => !evaluate(node.argument, scope),
'void': () => void evaluate(node.argument, scope),
'typeof': () => {
if (node.argument.type === 'Identifier') {
const $var = scope.$find(node.argument.name);
const value = $var ? $var.$get() : void 0;
return typeof value;
}
return typeof evaluate(node.argument, scope);
},
'delete': () => {
if (node.argument.type === 'MemberExpression') {
const { object, property, computed } = node.argument;
const obj = evaluate(object, scope);
let prop;
if (computed) {
prop = evaluate(property, scope);
} else {
prop = property.name;
}
return delete obj[prop];
} else {
throw '[Error] 出现错误'
}
},
}
return expressionMap[node.operator]();
},
UpdateExpression(node: t.UpdateExpression, scope) {
const { prefix, argument, operator } = node;
let $var: IVariable;
if (argument.type === 'Identifier') {
$var = scope.$find(argument.name);
if (!$var) throw `${argument.name} 未定义`;
} else if (argument.type === 'MemberExpression') {
const obj = evaluate(argument.object, scope);
let prop;
if (argument.computed) {
prop = evaluate(argument.property, scope);
} else {
prop = argument.property.name;
}
$var = {
$set(value: any) {
obj[prop] = value;
return true;
},
$get() {
return obj[prop];
}
}
} else {
throw '[Error] 出现错误'
}
const expressionMap = {
'++': v => {
$var.$set(v + 1);
return prefix ? ++v : v++
},
'--': v => {
$var.$set(v - 1);
return prefix ? --v : v--
},
}
return expressionMap[operator]($var.$get());
},
BinaryExpression(node: t.BinaryExpression, scope) {
const { left, operator, right } = node;
const expressionMap = {
'==': (a, b) => a == b,
'===': (a, b) => a === b,
'>': (a, b) => a > b,
'<': (a, b) => a < b,
'!=': (a, b) => a != b,
'!==': (a, b) => a !== b,
'>=': (a, b) => a >= b,
'<=': (a, b) => a <= b,
'<<': (a, b) => a << b,
'>>': (a, b) => a >> b,
'>>>': (a, b) => a >>> b,
'+': (a, b) => a + b,
'-': (a, b) => a - b,
'*': (a, b) => a * b,
'/': (a, b) => a / b,
'&': (a, b) => a & b,
'%': (a, b) => a % b,
'|': (a, b) => a | b,
'^': (a, b) => a ^ b,
'in': (a, b) => a in b,
'instanceof': (a, b) => a instanceof b,
}
return expressionMap[operator](evaluate(left, scope), evaluate(right, scope));
},
AssignmentExpression(node: t.AssignmentExpression, scope) {
const { left, right, operator } = node;
let $var: IVariable;
if (left.type === 'Identifier') {
$var = scope.$find(left.name);
if(!$var) throw `${left.name} 未定义`;
} else if (left.type === 'MemberExpression') {
const obj = evaluate(left.object, scope);
let prop;
if (left.computed) {
prop = evaluate(left.property, scope);
} else {
prop = left.property.name;
}
$var = {
$set(value: any) {
obj[prop] = value;
return true;
},
$get() {
return obj[prop];
}
}
} else {
throw '[Error] 出现错误'
}
const expressionMap = {
'=': v => { $var.$set(v); return $var.$get() },
'+=': v => { $var.$set($var.$get() + v); return $var.$get() },
'-=': v => { $var.$set($var.$get() - v); return $var.$get() },
'*=': v => { $var.$set($var.$get() * v); return $var.$get() },
'/=': v => { $var.$set($var.$get() / v); return $var.$get() },
'%=': v => { $var.$set($var.$get() % v); return $var.$get() },
'<<=': v => { $var.$set($var.$get() << v); return $var.$get() },
'>>=': v => { $var.$set($var.$get() >> v); return $var.$get() },
'>>>=': v => { $var.$set($var.$get() >>> v); return $var.$get() },
'|=': v => { $var.$set($var.$get() | v); return $var.$get() },
'&=': v => { $var.$set($var.$get() & v); return $var.$get() },
'^=': v => { $var.$set($var.$get() ^ v); return $var.$get() },
}
return expressionMap[operator](evaluate(right, scope));
},
LogicalExpression(node: t.LogicalExpression, scope) {
const { left, right, operator } = node;
const expressionMap = {
'&&': () => evaluate(left, scope) && evaluate(right, scope),
'||': () => evaluate(left, scope) || evaluate(right, scope),
}
return expressionMap[operator]();
},
MemberExpression(node: t.MemberExpression, scope) {
const { object, property, computed } = node;
const obj = evaluate(object, scope);
let prop;
if (computed) {
prop = evaluate(property, scope);
} else {
prop = property.name;
}
return obj[prop];
},
ConditionalExpression(node: t.ConditionalExpression, scope) {
const { test, consequent, alternate } = node;
return evaluate(test, scope) ? evaluate(consequent, scope) : evaluate(alternate, scope);
},
CallExpression(node: t.CallExpression, scope) {
const func = evaluate(node.callee, scope);
const args = node.arguments.map(arg => evaluate(arg, scope));
let _this;
if (node.callee.type === 'MemberExpression') {
_this = evaluate(node.callee.object, scope);
} else {
const $var = scope.$find('this');
_this = $var ? $var.$get() : null;
}
return func.apply(_this, args);
},
NewExpression(node: t.NewExpression, scope) {
const func = evaluate(node.callee, scope);
const args = node.arguments.map(arg => evaluate(arg, scope));
return new (func.bind(func, ...args));
},
SequenceExpression(node: t.SequenceExpression, scope) {
let last;
node.expressions.forEach(expr => {
last = evaluate(expr, scope);
})
return last;
},
}
至此,evaluate-apply流程讲解完毕。
Scope作用域实现
Scope作用域实现
接下来探讨Scope的具体实现方式。
class Scope implements IScope {
public readonly variables: EmptyObj = Object.create(null);
constructor(
private readonly scopeType: ScopeType,
private parent: Scope = null,
public readonly shared = false,
) { }
}
我们构造一个类来模拟Scope行为。观察可知,Scope类包含以下四个属性:
- variables:记录当前环境下存在的变量。
- scopeType:标识当前环境的类型。
- parent:指向当前环境的父级环境。
- shared:用于标识某些情况下无需重复构造子环境。
接下来介绍如何在环境中声明变量。
首先构造一个类来模拟变量行为。
class Variable implements IVariable {
constructor(
private kind: Kind,
private value: any
){ }
$get() {
return this.value
}
$set(value: any) {
if (this.kind === 'const') {
return false
}
this.value = value;
return true;
}
}
该类包含两个属性和两个方法:
- kind:用于标识变量是通过var、let还是const声明。
- value:表示变量的具体值。
- $get和$set:分别用于获取和设置变量的值。
有了Variable类后,即可编写Scope类中声明变量的方法。
let和const的声明逻辑基本一致。
$const(varName: string, value: any) {
const variable = this.variables[varName];
if (!variable) {
this.variables[varName] = new Variable('const', value);
return true;
}
return false;
}
$let(varName: string, value: any) {
const variable = this.variables[varName];
if (!variable) {
this.variables[varName] = new Variable('let', value);
return true;
}
return false;
}
var的声明方式略有不同,因为在JavaScript中,除函数作用域外,使用var声明的变量会提升至父级作用域(这是JS的历史遗留特性)。具体代码实现如下:
$var(varName: string, value: any) {
let scope: Scope = this;
while (!!scope.parent && scope.scopeType !== 'function') {
scope = scope.parent;
}
const variable = scope.variables[varName];
if (!variable) {
scope.variables[varName] = new Variable('var', value);
} else {
scope.variables[varName] = variable.$set(value);
}
return true
}
除了声明,我们还需要一个查找变量的方法。该方法从当前环境开始,沿作用域链一直向上查找,直至最外层环境。代码实现如下:
$find(varName: string): null | IVariable {
if (Reflect.has(this.variables, varName)) {
return Reflect.get(this.variables, varName);
}
if (this.parent) {
return this.parent.$find(varName);
}
return null;
}
至此,一个基础的JavaScript元循环求值器即告完成。
最后
大家可以在codesandbox在线体验。完整项目地址为nvwajs,欢迎提出建议及Star支持。
参考
《SICP》 微信小程序也要强行热更代码,鹅厂不服你来肛我呀
到此这篇关于编写一个javascript元循环求值器的方法的文章就介绍到这了,更多相关javascript元循环求值器内容请搜索本站以前的文章或继续浏览下面的相关文章希望大家以后多多支持本站!
JS处理含Emoji或中文的字符串转数组乱码问题
2026-11-05 16:12:28
JS怎么使用展开运算符合并对象
2026-11-05 14:42:18
ES6+环境下判断JS对象为空的简洁写法
2026-11-05 12:12:23
怎么通过replacer函数自定义JS中JSON的格式化输出
2026-11-05 11:19:44
JS正则表达式基础:字面量与RegExp构造函数的区别及常用元字符
2026-11-05 11:16:58
