muesli/lexer.go

package muesli

import (
	"bufio"
	_ "bytes"
	_ "errors"
	"fmt"
	"io"
	_ "io"
	"os"
	_ "reflect"
	_ "runtime"
	"strconv"
	"strings"
	"unicode"
	_ "unicode"
	// "gitlab.com/beoran/woe/graphviz"
	// _ "gitlab.com/beoran/woe/monolog"
)

/* A Lexer splits scanned input into tokens.
 */
type Lexer struct {
	Position
	Index int
	Start int
	io.RuneScanner
	buffer  []rune
	Current rune
	LoggerWrapper
}

func (lexer *Lexer) SetLogger(logger Logger) {
	lexer.LoggerWrapper = LoggerWrapper{logger}
}

func (lexer *Lexer) ClearBuffer() {
	lexer.buffer = make([]rune, 0)
}

func (lexer *Lexer) MakeIntegerToken() Token {
	var sbuffer = string(lexer.buffer)
	i, err := strconv.ParseInt(sbuffer, 0, 64)
	if err == nil {
		lexer.ClearBuffer()
		return NewToken(TokenKindInteger, IntValue(i), lexer.Position)
	} else {
		lexer.ClearBuffer()
		return lexer.MakeErrorToken(err)
	}
}

func (lexer *Lexer) MakeFloatToken() Token {
	var sbuffer = string(lexer.buffer)
	f, err := strconv.ParseFloat(sbuffer, 64)
	if err == nil {
		lexer.ClearBuffer()
		return NewToken(TokenKindFloat, FloatValue(f), lexer.Position)
	} else {
		lexer.ClearBuffer()
		return lexer.MakeErrorToken(err)
	}
}

func (lexer *Lexer) MakeBooleanToken(b bool) Token {
    lexer.ClearBuffer()
    if b { 
        return NewToken(TokenKindBoolean, TrueValue, lexer.Position)
    } else { 
        return NewToken(TokenKindBoolean, FalseValue, lexer.Position)
    }
}

func (lexer *Lexer) MakeNilToken() Token {
    lexer.ClearBuffer()
    return NewToken(TokenKindNil, NilValue, lexer.Position)
}



func (lexer *Lexer) MakeBuiltinToken() Token {
	var sbuffer = string(lexer.buffer)
    lexer.ClearBuffer()
	if sbuffer == "true" {
    	lexer.ClearBuffer()
        return NewToken(TokenKindBoolean, TrueValue, lexer.Position)
    } else if sbuffer == "false" {
        return NewToken(TokenKindBoolean, FalseValue, lexer.Position)
    } else if sbuffer == "nil" {
        return NewToken(TokenKindNil, NilValue, lexer.Position)
    } else {
        return lexer.MakeErrorfToken("Not a builtin: %s", sbuffer)
    }
}

func (lexer *Lexer) MakeStringValueToken(kind TokenKind) Token {
	var sbuffer = string(lexer.buffer)
	return NewToken(kind, StringValue(sbuffer), lexer.Position)
}

func (lexer *Lexer) MakeTypeValueToken(kind TokenKind) Token {
	var sbuffer = string(lexer.buffer)
	return NewToken(kind, TypeValue(sbuffer), lexer.Position)
}

func (lexer *Lexer) MakeErrorValueToken(kind TokenKind) Token {
	var sbuffer = string(lexer.buffer)
	return NewToken(kind, NewErrorValuef("%s", sbuffer), lexer.Position)
}

func (lexer *Lexer) MakeWordValueToken(kind TokenKind) Token {
	var sbuffer = string(lexer.buffer)
	return NewToken(kind, WordValue(sbuffer), lexer.Position)
}



func (lexer *Lexer) MakeToken(kind TokenKind) Token {
	switch kind {
	case TokenKindInteger:
		return lexer.MakeIntegerToken()
	case TokenKindFloat:
		return lexer.MakeFloatToken()
	case TokenKindString:
		return lexer.MakeStringValueToken(kind)
	case TokenKindSymbol:
		return lexer.MakeWordValueToken(kind)
	case TokenKindType:
		return lexer.MakeTypeValueToken(kind)
	case TokenKindError:
		return lexer.MakeErrorValueToken(kind)
	case TokenKindWord:
		return lexer.MakeWordValueToken(kind)
    case TokenKindNil:
        fallthrough
	case TokenKindBoolean:
		return lexer.MakeBuiltinToken()
	case TokenKindGet:
		fallthrough
	case TokenKindSet:
		fallthrough
	case TokenKindOpenBlock:
		fallthrough
	case TokenKindCloseBlock:
		fallthrough
	case TokenKindOpenList:
		fallthrough
	case TokenKindCloseList:
		fallthrough
	case TokenKindOpenParen:
		fallthrough
	case TokenKindCloseParen:
		fallthrough
	case TokenKindEOX:
		fallthrough
	case TokenKindEOF:
		val := StringValue(string(lexer.buffer))
		lexer.ClearBuffer()
		return NewToken(kind, val, lexer.Position)
	default:
		return lexer.MakeErrorfToken("Internal error on token type %s", kind)
	}
}

func (lexer Lexer) MakeErrorToken(err error) Token {
	return NewToken(TokenKindError, ErrorValue{err}, lexer.Position)
}

func (lexer Lexer) MakeErrorfToken(format string, va ...interface{}) Token {
	err := fmt.Errorf(format, va...)
	return lexer.MakeErrorToken(err)
}

func (lexer Lexer) MakeEOFToken() Token {
	return NewToken(TokenKindEOF, &EmptyValue{}, lexer.Position)
}

func (lexer *Lexer) Peek() (rune, error) {
	r, _, err := lexer.RuneScanner.ReadRune()
	err2 := lexer.RuneScanner.UnreadRune()

	if err == nil {
		err = err2
	}
	return r, err
}

/* Advances the lexer's position based on the rune r read. */
func (lexer *Lexer) advance(r rune) {
	lexer.Current = r
	lexer.Index++
	lexer.Position.Column++
	if r == '\n' {
		lexer.Position.Column = 1
		lexer.Position.Line++
	}
}

/* Append a rune to the lexer's buffer. */
func (lexer *Lexer) appendRune(r rune) {
	lexer.buffer = append(lexer.buffer, r)
}

/* Advances the lexer's input buffer but does not store the rune read,
 * but just returns it. */
func (lexer *Lexer) Skip() (rune, error) {
	r, _, err := lexer.RuneScanner.ReadRune()
	if err != nil {
		return 0, err
	}
	lexer.advance(r)
	return r, nil
}

/* Actually reads the next rune from the lexer's input source and stores
 * them in the lexer's token buffer.
 * Shorthand for r, err := lexer.Skip() ; lexer.appendRune(r)  */
func (lexer *Lexer) Next() (rune, error) {
	r, err := lexer.Skip()
	if err == nil {
		lexer.appendRune(r)
	}
	return r, nil
}

func (lexer *Lexer) DoIf(predicate func(rune) bool,
	todo func(*Lexer) (rune, error)) (bool, error) {
	r, err := lexer.Peek()
	if err != nil {
		return false, err
	}
	if predicate(r) {
		r, err = todo(lexer)
		if err != nil {
			return true, err
		}
		return true, nil
	}
	return false, nil
}

func (lexer *Lexer) NextIf(predicate func(rune) bool) (bool, error) {
	return lexer.DoIf(predicate, (*Lexer).Next)
}

func (lexer *Lexer) SkipIf(predicate func(rune) bool) (bool, error) {
	return lexer.DoIf(predicate, (*Lexer).Skip)
}

func (lexer *Lexer) NextWhile(predicate func(rune) bool) (bool, error) {
	result := true
	ok, err := lexer.NextIf(predicate)
	result = result || ok
	for ; ok && (err == nil); ok, err = lexer.NextIf(predicate) {
		result = result || ok
	}
	return result, err
}

func (lexer *Lexer) SkipWhile(predicate func(rune) bool) (bool, error) {
	result := true
	ok, err := lexer.SkipIf(predicate)
	result = result || ok
	for ; ok && (err == nil); ok, err = lexer.SkipIf(predicate) {
		result = result || ok
	}
	return result, err
}

func isEOX(r rune) bool {
	return r == '\n' || r == '.'
}


func isSpace(r rune) bool {
	return r == ' ' || r == '\t' || r == '\v' || r == '\r'
}

func isSpaceOrEOX(r rune) bool {
	return r == ' ' || r == '\t' || r == '\v' || r == '\r' || r == '\n' || r == '.'
}

func isComment(r rune) bool {
	return r == '#'
}

func (lexer *Lexer) SkipSpace() error {
	_, err := lexer.SkipWhile(isSpace)
	return err
}

func (lexer *Lexer) SkipBlockComment() error {
	var err error
	var r rune
	lexer.LogDebug("Skipping block comment.")
	for block := 1; block > 0 && err == nil; {
		_, err = lexer.Skip()
		if err != nil {
			return err
		}
		r, err = lexer.Peek()
		if r == '{' {
			block++
		} else if r == '}' {
			block--
		}
		lexer.LogDebug("Skipping block comment: %d", block)
	}
	_, err = lexer.Skip()
	return err
}

func (lexer *Lexer) SkipComment() error {
	r, err := lexer.Skip()
	lexer.LogDebug("Skipping %c.", r)
	if err != nil {
		return err
	}
	r, err = lexer.Peek()
	if r == '{' {
		return lexer.SkipBlockComment()
	}

	for r != '\n' && err == nil {
		lexer.LogDebug("Skipping comment %c.", r)
		_, err = lexer.Skip()
		if err != nil {
			return err
		}
		r, err = lexer.Peek()
	}
	if err != nil {
		return err
	}
	_, err = lexer.Skip()
	return err
}

/* Handles errors including EOF by either returning an error token or an
 * EOF token.
 */
func (lexer *Lexer) handleError(err error) Token {
	if err == io.EOF {
		return lexer.MakeEOFToken()
	} else {
		return lexer.MakeErrorToken(err)
	}
}

func (lexer *Lexer) LexNumber() Token {
	isFloat := false

	// skip any first - or +
	_, err := lexer.NextIf(func(r rune) bool {
		return r == '-' || r == '+'
	})

	_, err = lexer.NextWhile(func(r rune) bool {
		if unicode.IsDigit(r) {
			return true
		} else if r == '.' {
			if isFloat {
				return false // double point in floating point
			} else {
				isFloat = true
				return true
			}
		} else {
			return false
		}
	})

	if err != nil {
		return lexer.MakeErrorfToken("when parsing number: %s", err)
	}
	if isFloat {
		return lexer.MakeToken(TokenKindFloat)
	} else {
		return lexer.MakeToken(TokenKindInteger)
	}
}

func isDoubleQuote(r rune) bool {
	return r == '"'
}

func (lexer *Lexer) handleEscapeHexChars(amount int) error {
	buffer := make([]byte, 0)
	r, err := lexer.Skip()
	for index := 0; err == nil && index < amount; {
		if unicode.Is(unicode.ASCII_Hex_Digit, r) {
			buffer = append(buffer, byte(r))
		} else {
			return fmt.Errorf("Not a hexadecimal digit: %c", r)
		}
		index++
		if index < amount {
			r, err = lexer.Skip()
		}
	}
	if err != nil {
		return err
	}
	i, err := strconv.ParseInt(string(buffer), 16, 32)
	if err != nil {
		return err
	}
	lexer.appendRune(rune(i))
	_, err = lexer.Peek()
	return err
}

func (lexer *Lexer) handleEscape() error {
	r, err := lexer.Skip()
	if err != nil {
		return err
	}
	switch r {
	case 'a':
		lexer.appendRune('\a')
	case 'b':
		lexer.appendRune('\b')
	case 'e':
		lexer.appendRune('\033')
	case 'f':
		lexer.appendRune('\f')
	case 'n':
		lexer.appendRune('\n')
	case 'r':
		lexer.appendRune('\r')
	case 't':
		lexer.appendRune('\t')
	case '\\':
		lexer.appendRune('\\')
	case '"':
		lexer.appendRune('"')
	// case 'o':   fallthrough   // No octals, for now.
	case 'x':
		err = lexer.handleEscapeHexChars(2)
	case 'u':
		err = lexer.handleEscapeHexChars(4)
	case 'U':
		err = lexer.handleEscapeHexChars(6)
	default:
		return fmt.Errorf("Unknown escape sequence character %c: %d", r, r)
	}

	return err
}

func (lexer *Lexer) LexString() Token {
	var err error
	var r rune

	_, err = lexer.Skip() // Skip first "
	if err != nil {
		return lexer.handleError(err)
	}

	r, err = lexer.Skip()
	for r != '"' && err == nil {
		if r == '\\' {
			err = lexer.handleEscape()
			if err != nil {
				return lexer.handleError(err)
			}
		} else {
			lexer.appendRune(r)
			// still inside the string
		}
		r, err = lexer.Skip()
	}
	if err != nil {
		return lexer.MakeErrorfToken("when parsing string: %s", err)
	}
    
    if err != nil {
		return lexer.handleError(err)
	}

	return lexer.MakeToken(TokenKindString)
}

func (lexer *Lexer) LexLongString() Token {
	var err error

	_, err = lexer.Skip()
	if err != nil {
		return lexer.handleError(err)
	}

	_, err = lexer.NextWhile(func(r rune) bool {
		return r != '`'
	})

	if err != nil {
		return lexer.MakeErrorfToken("when parsing long string: %s", err)
	}

	_, err = lexer.Skip()
	if err != nil {
		return lexer.handleError(err)
	}

	return lexer.MakeToken(TokenKindString)
}

func (lexer *Lexer) LexWordOrType(kind TokenKind) Token {
	var err error
	first := true

	_, err = lexer.Next()
	if err != nil {
		return lexer.handleError(err)
	}

	_, err = lexer.NextWhile(func(r rune) bool {
		if first {
			first = false
			return unicode.IsLetter(r) || r == '_'
		} else {
			return unicode.IsLetter(r) || unicode.IsNumber(r) || r == '_'
		}
	})

	if err != nil {
		return lexer.handleError(err)
	}

	sbuffer := string(lexer.buffer)

	// handle key words
	switch sbuffer {
	case "true":
		return lexer.MakeBooleanToken(true)
	case "false":
		return lexer.MakeBooleanToken(false)
    case "nil":
        return lexer.MakeNilToken()
	default:
	}

	return lexer.MakeToken(kind)
}

func (lexer *Lexer) LexWord() Token {
	return lexer.LexWordOrType(TokenKindWord)
}

func (lexer *Lexer) LexType() Token {
	return lexer.LexWordOrType(TokenKindType)
}

func (lexer *Lexer) LexSymbol() Token {
	var err error

	_, err = lexer.Skip()
	if err != nil {
		return lexer.handleError(err)
	}

	_, err = lexer.NextWhile(func(r rune) bool {
		return !isSpaceOrEOX(r)
	})

	if err != nil {
		return lexer.handleError(err)
	}

	return lexer.MakeToken(TokenKindSymbol)
}


func (lexer *Lexer) LexBuiltin() Token {
	var err error

	_, err = lexer.Skip()
	if err != nil {
		return lexer.handleError(err)
	}

	_, err = lexer.NextWhile(func(r rune) bool {
		return !isSpaceOrEOX(r)
	})

	if err != nil {
		return lexer.handleError(err)
	}
    
	return lexer.MakeBuiltinToken()
}

func (lexer *Lexer) skipSpaceAndCommentAndPeek() (rune, error) {
	r, err := lexer.Peek()
	if err != nil {
		return r, err
	}
	i := 0
	for isSpace(r) || isComment(r) {
		if isSpace(r) {
			err = lexer.SkipSpace()
		} else if isComment(r) {
			err = lexer.SkipComment()
		}
		if err != nil {
			return r, err
		}
		i++
		r, err = lexer.Peek()
		lexer.LogDebug("Peeked again: >%c< %v %v %d", r, isSpace(r), isComment(r), i)
		if err != nil {
			return r, err
		}
	}
	return r, err
}

func (lexer *Lexer) LexEOX() Token {
    lexer.Next()
    _, err := lexer.skipSpaceAndCommentAndPeek()
	if err != nil {
		return lexer.handleError(err)
	}
    
    _, err = lexer.NextWhile(func(r rune) bool {
        return isSpaceOrEOX(r) || r == '\n' || r == '.'
    })

    if err != nil {
            return lexer.handleError(err)
    }
    
	return lexer.MakeToken(TokenKindEOX)
}


func (lexer *Lexer) lex() Token {
	r, err := lexer.skipSpaceAndCommentAndPeek()
	lexer.LogDebug(" After skip: >%c< >%v<\n", r, err)

	if err != nil {
		return lexer.handleError(err)
	}

	if unicode.IsDigit(r) || r == '-' || r == '+' {
		return lexer.LexNumber()
	}

	if r == '\n' || r == '.' {
	    return lexer.LexEOX()
	}

	if r == '"' {
		return lexer.LexString()
	}

	if r == '`' {
		return lexer.LexLongString()
	}
    
    if r == '!' {
		return lexer.LexBuiltin()
	}

	if r == ':' {
		return lexer.LexSymbol()
	}

	switch TokenKind(r) {
	case TokenKindGet:
		fallthrough
	case TokenKindSet:
		fallthrough
	case TokenKindOpenBlock:
		fallthrough
	case TokenKindCloseBlock:
		fallthrough
	case TokenKindOpenList:
		fallthrough
	case TokenKindCloseList:
		fallthrough
	case TokenKindOpenParen:
		fallthrough
	case TokenKindCloseParen:
		lexer.Next()
		return lexer.MakeToken(TokenKind(r))
	default:
	}

	if unicode.IsLetter(r) {
		if unicode.IsUpper(r) {
			return lexer.LexType()
		} else {
			return lexer.LexWord()
		}
	}
	
	// EOF character
	if r == 0x7f {
		return lexer.MakeEOFToken()
	}

	return lexer.MakeErrorfToken("Unknown character: %c", r)
}

func (lexer *Lexer) Lex() Token {
	res := lexer.lex()
	lexer.ClearBuffer() // ensure buffer is cleared after lexing, always.
	return res
}

func (lexer *Lexer) LexAll() []Token {
	var token Token

	res := make([]Token, 0)

	for token = lexer.Lex(); !token.IsLast(); token = lexer.Lex() {
		res = append(res, token)
	}

	res = append(res, token)

	return res
}

func NewLexer(scanner io.RuneScanner, filename string) *Lexer {
	lexer := &Lexer{}
	lexer.RuneScanner = scanner
	lexer.Position.FileName = filename
	lexer.Position.Column = 1
	lexer.Position.Line = 1
	lexer.LoggerWrapper = LoggerWrapper{nil}
	return lexer
}

func (lexer * Lexer) Report() {
    if lexer == nil {
        fmt.Printf("Lexer: is nil\n") 
    } else {
        fmt.Printf("Lexer: %s:%d:%d\n", 
        lexer.Position.FileName,
        lexer.Position.Column,
        lexer.Position.Line)
    }
}

func NewLexerFromString(input string) *Lexer {
	reader := strings.NewReader(input)
	return NewLexer(reader, "<input>")
}

func NewLexerFromFilename(filename string) (*Lexer, error) {
	read, err := os.Open(filename)
	if err == nil {
		bread := bufio.NewReader(read)
		lex := NewLexer(bread, filename)
		return lex, nil
	}
	return nil, err
}