Monkey/parser/parser.go

package parser

import (
	"fmt"
	"github/runnignwater/monkey/ast"
	"github/runnignwater/monkey/lexer"
	"github/runnignwater/monkey/token"
	"strconv"
)

/**
 * @Author: simon
 * @Author: ynwdlxm@163.com
 * @Date: 2022/10/2 下午9:55
 * @Desc:
 */

const (
	// 优先级常量
	_ int = iota
	LOWEST
	EQUALS      // ==
	LESSGREATER // > OR <
	SUM         // +
	PRODUCT     // *
	PREFIX      // -X OR !X
	CALL        // myFunction(X)
)

// 指派 token 类型的优先级
var precedences = map[token.TypeToken]int{
	token.EQ:       EQUALS,
	token.NOT_EQ:   EQUALS,
	token.GT:       LESSGREATER,
	token.LT:       LESSGREATER,
	token.PLUS:     SUM,
	token.MINUS:    SUM,
	token.ASTERISK: PRODUCT,
	token.SLASH:    PRODUCT,
}

type (
	prefixParseFn func() ast.Expression
	infixParseFn  func(expression ast.Expression) ast.Expression
)

type Parser struct {
	l      *lexer.Lexer // point to the instance of the lexer
	errors []string

	curToken  token.Token // point to the current token
	peekToken token.Token // point to the next token

	prefixParseFns map[token.TypeToken]prefixParseFn
	infixParseFns  map[token.TypeToken]infixParseFn
}

func New(l *lexer.Lexer) *Parser {
	p := &Parser{
		l:      l,
		errors: []string{},
	}

	p.prefixParseFns = make(map[token.TypeToken]prefixParseFn)
	p.registerPrefix(token.IDENT, p.parseIdentifier)
	p.registerPrefix(token.INT, p.parseIntegerLiteral)
	p.registerPrefix(token.BANG, p.parsePrefixExpression)
	p.registerPrefix(token.MINUS, p.parsePrefixExpression)
	p.registerPrefix(token.TRUE, p.parseBoolean)
	p.registerPrefix(token.FALSE, p.parseBoolean)
	p.registerPrefix(token.LPAREN, p.parseGroupedExpression)
	p.registerPrefix(token.IF, p.parseIfExpression)

	p.infixParseFns = make(map[token.TypeToken]infixParseFn)
	p.registerInfix(token.PLUS, p.parseInfixExpression)
	p.registerInfix(token.MINUS, p.parseInfixExpression)
	p.registerInfix(token.ASTERISK, p.parseInfixExpression)
	p.registerInfix(token.SLASH, p.parseInfixExpression)
	p.registerInfix(token.EQ, p.parseInfixExpression)
	p.registerInfix(token.NOT_EQ, p.parseInfixExpression)
	p.registerInfix(token.GT, p.parseInfixExpression)
	p.registerInfix(token.LT, p.parseInfixExpression)

	// Read two tokens, so curToken and peekToken are both set
	p.nextToken()
	p.nextToken()

	return p
}

func (p *Parser) parseIdentifier() ast.Expression {
	return &ast.Identifier{Token: p.curToken, Value: p.curToken.Literal}
}

func (p *Parser) parseBoolean() ast.Expression {
	return &ast.Boolean{Token: p.curToken, Value: p.curTokenIs(token.TRUE)}
}
func (p *Parser) parseGroupedExpression() ast.Expression {
	defer untrace(trace("parseGroupedExpression"))
	p.nextToken()

	exp := p.parseExpression(LOWEST)

	if !p.expectPeek(token.RPAREN) {
		return nil
	}

	return exp
}
func (p *Parser) parseIfExpression() ast.Expression {
	defer untrace(trace("parseIfExpression"))
	exp := &ast.IfExpression{Token: p.curToken}

	// (
	if !p.expectPeek(token.LPAREN) {
		return nil
	}
	p.nextToken()
	exp.Condition = p.parseExpression(LOWEST)
	// )
	if !p.expectPeek(token.RPAREN) {
		return nil
	}

	// {
	if !p.expectPeek(token.LBRACE) {
		return nil
	}

	exp.Consequence = p.parseBlockStatement()

	// else expression
	if p.peekTokenIs(token.ELSE) {
		p.nextToken()

		if !p.expectPeek(token.LBRACE) {
			return nil
		}
		exp.Alternative = p.parseBlockStatement()
	}
	return exp
}
func (p *Parser) parseIntegerLiteral() ast.Expression {
	defer untrace(trace("parseIntegerLiteral"))

	lit := &ast.IntegerLiteral{Token: p.curToken}

	value, err := strconv.ParseInt(p.curToken.Literal, 0, 64)
	if err != nil {
		msg := fmt.Sprintf("could not parse %q as integer", p.curToken.Literal)
		p.errors = append(p.errors, msg)
		return nil
	}
	lit.Value = value

	return lit
}
func (p *Parser) parseInfixExpression(left ast.Expression) ast.Expression {
	defer untrace(trace("parseInfixExpression"))
	exp := &ast.InfixExpression{
		Token:    p.curToken,
		Left:     left,
		Operator: p.curToken.Literal,
	}

	precedence := p.curPrecedence()
	p.nextToken()
	exp.Right = p.parseExpression(precedence)

	return exp
}
func (p *Parser) parsePrefixExpression() ast.Expression {
	defer untrace(trace("parsePrefixExpression"))
	exp := &ast.PrefixExpression{
		Token:    p.curToken,
		Operator: p.curToken.Literal,
	}

	p.nextToken()

	exp.Right = p.parseExpression(PREFIX)

	return exp
}
func (p *Parser) parseBlockStatement() *ast.BlockStatement {
	defer untrace(trace("parseBlockStatement"))
	block := &ast.BlockStatement{Token: p.curToken}
	block.Statements = []ast.Statement{}

	p.nextToken()

	// }
	for !p.curTokenIs(token.RBRACE) {
		stmt := p.parseStatement()
		if stmt != nil {
			block.Statements = append(block.Statements, stmt)
		}
		p.nextToken()
	}

	return block
}

func (p *Parser) Errors() []string {
	return p.errors
}

func (p *Parser) peekError(t token.TypeToken) {
	msg := fmt.Sprintf("exepected next token to be %s, got %s instead.", t, p.peekToken.Type)
	p.errors = append(p.errors, msg)
}

func (p *Parser) nextToken() {
	p.curToken = p.peekToken
	p.peekToken = p.l.NextToken()
}

func (p *Parser) ParseProgram() *ast.Program {
	program := &ast.Program{}
	program.Statements = []ast.Statement{}

	for !p.curTokenIs(token.EOF) {
		stmt := p.parseStatement()
		if stmt != nil {
			program.Statements = append(program.Statements, stmt)
		}
		p.nextToken()

	}
	return program
}

func (p *Parser) parseStatement() ast.Statement {
	switch p.curToken.Type {
	case token.LET:
		return p.parseLetStatement()
	case token.RETURN:
		return p.parseReturnStatement()
	default:
		return p.parseExpressionStatement()
	}
}

// let <identifier> = <expression>;
func (p *Parser) parseLetStatement() *ast.LetStatement {
	stmt := &ast.LetStatement{Token: p.curToken}

	// let
	if !p.expectPeek(token.IDENT) {
		return nil
	}

	// identifier
	stmt.Name = &ast.Identifier{Token: p.curToken, Value: p.curToken.Literal}
	// =
	if !p.expectPeek(token.ASSIGN) {
		return nil
	}

	// TODO: we're skipping the expression until we
	//       we encounter a semicolon

	// ;
	for !p.curTokenIs(token.SEMICOLON) {
		p.nextToken()
	}
	return stmt
}

// return <expression>;
func (p *Parser) parseReturnStatement() *ast.ReturnStatement {
	stmt := &ast.ReturnStatement{Token: p.curToken}

	p.nextToken()

	// TODO: we're skipping the expressions until we
	//       encounter a semicolon
	for !p.curTokenIs(token.SEMICOLON) {
		p.nextToken()
	}

	return stmt
}

func (p *Parser) parseExpressionStatement() *ast.ExpressionStatement {
	defer untrace(trace("parseExpressionStatement"))

	stmt := &ast.ExpressionStatement{Token: p.curToken}

	stmt.Expression = p.parseExpression(LOWEST)

	if p.peekTokenIs(token.SEMICOLON) {
		p.nextToken()
	}

	return stmt
}

func (p *Parser) parseExpression(precedence int) ast.Expression {
	defer untrace(trace("parseExpression"))
	prefix := p.prefixParseFns[p.curToken.Type]
	if prefix == nil {
		p.noPrefixParseFnError(p.curToken.Type)
		return nil
	}
	leftExp := prefix()

	for !p.peekTokenIs(token.SEMICOLON) && precedence < p.peekPrecedence() {
		infix := p.infixParseFns[p.peekToken.Type]
		if infix == nil {
			return leftExp
		}
		p.nextToken()

		leftExp = infix(leftExp)
	}

	return leftExp
}

func (p *Parser) curTokenIs(t token.TypeToken) bool {
	return p.curToken.Type == t
}

func (p *Parser) peekTokenIs(t token.TypeToken) bool {
	return p.peekToken.Type == t
}

func (p *Parser) expectPeek(t token.TypeToken) bool {
	if p.peekTokenIs(t) {
		p.nextToken()
		return true
	} else {
		p.peekError(t)
		return false
	}
}

func (p *Parser) registerPrefix(tokenType token.TypeToken, fn prefixParseFn) {
	p.prefixParseFns[tokenType] = fn
}

func (p *Parser) registerInfix(tokenType token.TypeToken, fn infixParseFn) {
	p.infixParseFns[tokenType] = fn
}

func (p *Parser) noPrefixParseFnError(t token.TypeToken) {
	msg := fmt.Sprintf("no prefix parse function for %s found", t)
	p.errors = append(p.errors, msg)
}
func (p *Parser) peekPrecedence() int {
	if p, ok := precedences[p.peekToken.Type]; ok {
		return p
	}
	return LOWEST
}
func (p *Parser) curPrecedence() int {
	if p, ok := precedences[p.curToken.Type]; ok {
		return p
	}
	return LOWEST
}