1
0
Fork 0
kifflom/lexer/lexer.go

284 lines
5.0 KiB
Go
Raw Normal View History

2015-02-12 11:12:19 +00:00
package lexer
import (
"fmt"
"strings"
"unicode/utf8"
)
type (
// lexer holds the state of the scanner.
Lexer struct {
2015-02-16 11:18:14 +00:00
input string // the string being scanned
state stateFn // the next lexing function to enter
lineNum int // Line number
pos int // current position in the input
start int // start position of this item
width int // width of last rune read from input
items chan Item // channel of scanned items
2015-02-12 11:12:19 +00:00
}
2015-02-16 11:18:14 +00:00
// Item represents a token or text string returned from the scanner.
2015-02-16 08:54:53 +00:00
Item struct {
2015-02-16 11:18:14 +00:00
Token Token // The type of this item.
Pos int // The starting position, in bytes, of this item in the input string.
Val string // The value of this item.
2015-02-12 11:12:19 +00:00
}
2015-02-16 11:18:14 +00:00
// Token identifies the type of lex items.
Token int
// stateFn represents the state of the scanner as a function that returns the next state.
stateFn func(*Lexer) stateFn
2015-02-12 11:12:19 +00:00
)
const (
// Special
2015-02-16 11:18:14 +00:00
Error Token = iota // error occurred; value is text of error
EOF
2015-02-12 11:12:19 +00:00
// Symbols
2015-02-16 11:18:14 +00:00
BraceOpen // {
BraceClose // }
BracketOpen // [
BracketClose // [
Quote // "
Colon // :
Comma // ,
2015-02-12 11:12:19 +00:00
// Types
2015-02-16 11:18:14 +00:00
Null // null
Bool // true, false
Number // 0, 2.5
String // "foo"
2015-02-12 11:12:19 +00:00
)
// lex creates a new scanner for the input string.
2015-02-16 09:02:31 +00:00
func New(input string) *Lexer {
2015-02-12 11:12:19 +00:00
l := &Lexer{
input: input,
2015-02-16 08:54:53 +00:00
items: make(chan Item),
2015-02-12 11:12:19 +00:00
}
return l
}
// run runs the state machine for the lexer.
func (l *Lexer) Run() {
for l.state = lexInitial; l.state != nil; {
l.state = l.state(l)
}
}
2015-02-16 11:18:14 +00:00
func (l *Lexer) NextItem() (item Item, ok bool) {
item, ok = <-l.items
return
2015-02-16 08:54:53 +00:00
}
2015-02-12 11:12:19 +00:00
//
// Lexer stuff
//
2015-02-16 08:54:53 +00:00
func (i Item) String() string {
2015-02-16 11:18:14 +00:00
switch i.Token {
case EOF:
2015-02-12 11:12:19 +00:00
return "EOF"
2015-02-16 11:18:14 +00:00
case Error:
return "Error: " + i.Val
case BraceOpen:
2015-02-16 08:54:53 +00:00
return "{"
2015-02-16 11:18:14 +00:00
case BraceClose:
2015-02-16 08:54:53 +00:00
return "}"
2015-02-16 11:18:14 +00:00
case BracketOpen:
2015-02-16 08:54:53 +00:00
return "["
2015-02-16 11:18:14 +00:00
case BracketClose:
2015-02-16 08:54:53 +00:00
return "]"
2015-02-16 11:18:14 +00:00
case Quote:
2015-02-16 08:54:53 +00:00
return "\""
2015-02-16 11:18:14 +00:00
case Colon:
2015-02-16 08:54:53 +00:00
return ":"
2015-02-16 11:18:14 +00:00
case Comma:
2015-02-16 08:54:53 +00:00
return ","
2015-02-16 11:18:14 +00:00
case Null:
2015-02-16 08:54:53 +00:00
return "NULL"
2015-02-16 11:18:14 +00:00
case Bool:
return "Bool: " + i.Val
case Number:
return "Number: " + i.Val
case String:
return "String: " + i.Val
2015-02-16 08:54:53 +00:00
default:
panic("Unreachable")
2015-02-12 11:12:19 +00:00
}
}
// next returns the next rune in the input.
func (l *Lexer) next() rune {
if int(l.pos) >= len(l.input) {
l.width = 0
2015-02-16 11:18:14 +00:00
return 0
2015-02-12 11:12:19 +00:00
}
r, w := utf8.DecodeRuneInString(l.input[l.pos:])
2015-02-16 08:54:53 +00:00
l.width = w
2015-02-12 11:12:19 +00:00
l.pos += l.width
return r
}
// peek returns but does not consume the next rune in the input.
func (l *Lexer) peek() rune {
r := l.next()
l.backup()
return r
}
// backup steps back one rune. Can only be called once per call of next.
func (l *Lexer) backup() {
l.pos -= l.width
}
// emit passes an item back to the client.
2015-02-16 11:18:14 +00:00
func (l *Lexer) emit(t Token) {
2015-02-16 08:54:53 +00:00
l.items <- Item{t, l.start, l.input[l.start:l.pos]}
2015-02-12 11:12:19 +00:00
l.start = l.pos
2015-02-16 11:18:14 +00:00
if t == EOF {
close(l.items)
}
2015-02-12 11:12:19 +00:00
}
// ignore skips over the pending input before this point.
func (l *Lexer) ignore() {
l.start = l.pos
}
2015-02-16 08:54:53 +00:00
func (l *Lexer) acceptString(s string) (ok bool) {
if strings.HasPrefix(l.input[l.pos:], s) {
l.pos += len(s)
2015-02-12 11:12:19 +00:00
return true
}
return false
}
func (l *Lexer) errorf(format string, args ...interface{}) stateFn {
2015-02-16 11:18:14 +00:00
l.items <- Item{Error, l.start, fmt.Sprintf(format, args...)}
return nil // Stop lexing
}
//
// States
//
func lexInitial(l *Lexer) stateFn {
for {
switch r := l.next(); r {
case ' ', '\t':
return lexSpace(l)
case '\n':
l.lineNum++
2015-02-16 11:25:33 +00:00
l.ignore()
2015-02-16 11:18:14 +00:00
case 'n':
l.backup()
return lexNull(l)
case 't', 'f':
l.backup()
return lexBool(l)
case '1', '2', '3', '4', '5', '6', '7', '8', '9', '0':
l.backup()
return lexNumber(l)
case '"':
return lexString(l)
case '[':
l.emit(BracketOpen)
case ']':
l.emit(BracketClose)
case '{':
l.emit(BraceOpen)
case '}':
l.emit(BraceClose)
case ':':
l.emit(Colon)
case ',':
l.emit(Comma)
case 0:
2015-02-16 11:25:33 +00:00
l.emit(EOF)
return nil
2015-02-16 11:18:14 +00:00
default:
panic("Unexpected symbol: " + string(r))
}
}
2015-02-12 11:12:19 +00:00
}
2015-02-16 11:18:14 +00:00
// Skip all spaces
func lexSpace(l *Lexer) stateFn {
for {
if r := l.peek(); r == ' ' || r == '\t' {
l.next()
} else {
break
}
}
l.ignore()
return lexInitial
}
func lexNull(l *Lexer) stateFn {
if l.acceptString("null") {
l.emit(Null)
} else {
return l.errorf("Unexpected token")
}
return lexInitial
}
func lexBool(l *Lexer) stateFn {
if l.acceptString("true") || l.acceptString("false") {
l.emit(Bool)
}
return lexInitial
}
func lexNumber(l *Lexer) stateFn {
hasDot := false
for {
switch r := l.peek(); r {
case '1', '2', '3', '4', '5', '6', '7', '8', '9', '0':
l.next()
case '.':
if hasDot {
return l.errorf("Invalid number")
} else {
hasDot = true
l.next()
}
default:
2015-02-16 11:25:33 +00:00
l.emit(Number)
return lexInitial
2015-02-16 11:18:14 +00:00
}
}
}
func lexString(l *Lexer) stateFn {
l.ignore()
2015-02-16 11:18:14 +00:00
escaped := false
for {
switch r := l.next(); r {
case '\\':
2015-02-16 11:25:33 +00:00
escaped = !escaped
2015-02-16 11:18:14 +00:00
case '"':
if escaped {
escaped = false
} else {
l.backup() // Going before closing quote
2015-02-16 11:18:14 +00:00
l.emit(String)
l.next() // Skipping closing quote
2015-02-16 11:25:33 +00:00
return lexInitial
2015-02-16 11:18:14 +00:00
}
case '\n':
l.lineNum++
case 0:
return l.errorf("Unterminated string")
default:
escaped = false
}
}
}