umx_compiler

commit e4e1fdf1534021f113946a65701fdba78d1be9d9
parent b2cb2a1c4d76faa74a9a398f498bc0d866628f0b
Author: bsandro <email@bsandro.tech>
Date:   Sun, 12 Jun 2022 23:13:04 +0300

Basic lexer error checking

Diffstat:
A
ast/ast.go
 | 
52
++++++++++++++++++++++++++++++++++++++++++++++++++++
M
main.go
 | 
2
+-
A
parser/parser.go
 | 
96
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
A
parser/parser_test.go
 | 
79
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
M
repl/repl.go
 | 
2
+-

5 files changed, 229 insertions(+), 2 deletions(-)
diff --git a/ast/ast.go b/ast/ast.go
@@ -0,0 +1,52 @@
+package ast
+
+import (
+	"interp/token"
+)
+
+type Node interface {
+	TokenLiteral() string
+}
+
+type Statement interface {
+	Node
+	statementNode()
+}
+
+type Expression interface {
+	Node
+	expressionNode()
+}
+
+type Identifier struct {
+	Token token.Token
+	Value string
+}
+
+func (i *Identifier) expressionNode() {}
+func (i *Identifier) TokenLiteral() string {
+	return i.Token.Literal
+}
+
+type LetStatement struct {
+	Token token.Token
+	Name  *Identifier
+	Value Expression
+}
+
+func (l *LetStatement) statementNode() {}
+func (l *LetStatement) TokenLiteral() string {
+	return l.Token.Literal
+}
+
+type Program struct {
+	Statements []Statement
+}
+
+func (p *Program) TokenLiteral() string {
+	if len(p.Statements) > 0 {
+		return p.Statements[0].TokenLiteral()
+	} else {
+		return ""
+	}
+}
diff --git a/main.go b/main.go
@@ -2,9 +2,9 @@ package main
 
 import (
 	"fmt"
+	"interp/repl"
 	"os"
 	"os/user"
-	"interp/repl"
 )
 
 func main() {
diff --git a/parser/parser.go b/parser/parser.go
@@ -0,0 +1,96 @@
+package parser
+
+import (
+	"fmt"
+	"interp/ast"
+	"interp/lexer"
+	"interp/token"
+)
+
+type Parser struct {
+	l         *lexer.Lexer
+	curToken  token.Token
+	peekToken token.Token
+	errors    []string
+}
+
+func New(l *lexer.Lexer) *Parser {
+	p := &Parser{
+		l: l,
+		errors: []string{},
+	}
+	p.nextToken()
+	p.nextToken()
+	return p
+}
+
+func (p *Parser) Errors() []string {
+	return p.errors
+}
+
+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.curToken.Type != token.EOF {
+		statement := p.parseStatement()
+		if statement != nil {
+			program.Statements = append(program.Statements, statement)
+		}
+		p.nextToken()
+	}
+	return program
+}
+
+func (p *Parser) parseStatement() ast.Statement {
+	switch p.curToken.Type {
+	case token.LET:
+		return p.parseLetStatement()
+	default:
+		return nil
+	}
+}
+
+func (p *Parser) parseLetStatement() *ast.LetStatement {
+	statement := &ast.LetStatement{Token: p.curToken}
+	if !p.expectPeek(token.IDENT) {
+		return nil
+	}
+	statement.Name = &ast.Identifier{Token: p.curToken, Value: p.curToken.Literal}
+	if !p.expectPeek(token.ASSIGN) {
+		return nil
+	}
+
+	for !p.curTokenIs(token.SEMICOLON) {
+		p.nextToken()
+	}
+
+	return statement
+}
+
+func (p *Parser) curTokenIs(t token.TokenType) bool {
+	return p.curToken.Type == t
+}
+
+func (p *Parser) peekTokenIs(t token.TokenType) bool {
+	return p.peekToken.Type == t
+}
+
+func (p *Parser) expectPeek(t token.TokenType) bool {
+	if p.peekTokenIs(t) {
+		p.nextToken()
+		return true
+	} else {
+		p.peekError(t)
+		return false
+	}
+}
+
+func (p *Parser) peekError(t token.TokenType) {
+	msg := fmt.Sprintf("token expected: %s, got: %s", t, p.peekToken.Type);
+	p.errors = append(p.errors, msg)
+}
diff --git a/parser/parser_test.go b/parser/parser_test.go
@@ -0,0 +1,79 @@
+package parser
+
+import (
+	"interp/ast"
+	"interp/lexer"
+	"testing"
+)
+
+func TestLetStatements(t *testing.T) {
+	input := `
+let x = 5;
+let y = 10;
+let foo = 112233;
+`
+	l := lexer.New(input)
+	p := New(l)
+	program := p.ParseProgram()
+	checkParserErrors(t, p)
+
+	if program == nil {
+		t.Fatalf("ParseProgram() returned nil")
+	}
+	if len(program.Statements) != 3 {
+		t.Fatalf("program.Statements has not 3 elements but %d", len(program.Statements))
+	}
+
+	tests := []struct {
+		expectedIdentifier string
+	}{
+		{"x"},
+		{"y"},
+		{"foo"},
+	}
+
+	for i, tt := range tests {
+		statement := program.Statements[i]
+		if !testLetStatement(t, statement, tt.expectedIdentifier) {
+			return
+		}
+	}
+}
+
+func testLetStatement(t *testing.T, s ast.Statement, name string) bool {
+	if s.TokenLiteral() != "let" {
+		t.Errorf("s.TokenLiteral() is not 'let' but '%s' instead", s.TokenLiteral())
+		return false
+	}
+
+	letStatement, ok := s.(*ast.LetStatement)
+	if !ok {
+		t.Errorf("s is not *ast.LetStatement but %T", s)
+		return false
+	}
+
+	if letStatement.Name.Value != name {
+		t.Errorf("letStatement.Name.Value is not %s but %s", name, letStatement.Name.Value)
+		return false
+	}
+
+	if letStatement.Name.TokenLiteral() != name {
+		t.Errorf("letStatement.Name.TokenLiteral() is not %s but %s", name, letStatement.Name.TokenLiteral())
+		return false
+	}
+
+	return true
+}
+
+func checkParserErrors(t *testing.T, p *Parser) {
+	errors := p.Errors()
+	if len(errors) == 0 {
+		return
+	}
+
+	t.Errorf("%d parse errors:", len(errors))
+	for _, msg := range errors {
+		t.Errorf("%s", msg)
+	}
+	t.FailNow()
+}
diff --git a/repl/repl.go b/repl/repl.go
@@ -3,9 +3,9 @@ package repl
 import (
 	"bufio"
 	"fmt"
-	"io"
 	"interp/lexer"
 	"interp/token"
+	"io"
 )
 
 const PROMPT = "$> "