Начальный релизHEADv0.0.1master

4 files changed, 122 insertions, 0 deletions

diff --git a/go.mod b/go.mod
new file mode 100644
index 0000000..719088b
--- /dev/null
+++ b/go.mod
@@ -0,0 +1,3 @@
+module go.neonxp.ru/lex
+
+go 1.25.0
diff --git a/states.go b/states.go
new file mode 100644
index 0000000..6b75a7b
--- /dev/null
+++ b/states.go
@@ -0,0 +1,3 @@
+package lex
+
+type StateFunc func(l *Lexer) StateFunc
diff --git a/token.go b/token.go
new file mode 100644
index 0000000..ef3c92b
--- /dev/null
+++ b/token.go
@@ -0,0 +1,10 @@
+package lex
+
+type Token struct {
+	Typ   Typ
+	Pos   int
+	Value string
+	Error error
+}
+
+type Typ int
diff --git a/tokenizer.go b/tokenizer.go
new file mode 100644
index 0000000..1ddd8d0
--- /dev/null
+++ b/tokenizer.go
@@ -0,0 +1,106 @@
+// Package lex представляет собой достаточно простой лексер произвольных выражений.
+// Практически полностью аналогичен лексеру от Роба Пайка.
+package lex
+
+import (
+	"strings"
+	"unicode/utf8"
+)
+
+func Do(initState StateFunc, input string) <-chan Token {
+	t := &Lexer{
+		pos:   0,
+		input: input,
+		start: 0,
+		ch:    make(chan Token),
+	}
+
+	go func() {
+		defer close(t.ch)
+		for state := initState; state != nil; {
+			state = state(t)
+		}
+	}()
+
+	return t.ch
+}
+
+type Lexer struct {
+	pos   int
+	start int
+	input string
+	width int
+	ch    chan Token
+}
+
+func (l *Lexer) Next() (ch rune) {
+	if l.pos >= len(l.input) {
+		l.width = 0
+		return 0
+	}
+	ch, l.width = utf8.DecodeRuneInString(l.input[l.pos:])
+	l.pos += l.width
+	return ch
+}
+
+func (l *Lexer) Peek() rune {
+	ch := l.Next()
+	l.Back()
+	return ch
+}
+
+func (l *Lexer) Back() {
+	l.pos -= l.width
+}
+
+func (l *Lexer) Skip() {
+	l.start = l.pos
+}
+
+func (l *Lexer) EmitToken(typ Typ) {
+	value := ""
+	if l.pos > l.start {
+		value = l.input[l.start:l.pos]
+	}
+	l.ch <- Token{
+		Typ:   typ,
+		Value: value,
+		Pos:   l.pos,
+	}
+	l.start = l.pos
+}
+
+func (l *Lexer) EmitError(err error) {
+	value := ""
+	if l.pos > l.start {
+		value = l.input[l.start:l.pos]
+	}
+	l.ch <- Token{
+		Value: value,
+		Pos:   l.pos,
+		Error: err,
+	}
+}
+
+func (l *Lexer) Accept(valid string) bool {
+	if strings.ContainsRune(valid, l.Next()) {
+		return true
+	}
+	l.Back()
+
+	return false
+}
+
+func (l *Lexer) AcceptRun(valid string) {
+	for strings.ContainsRune(valid, l.Next()) {
+	}
+	l.Back()
+}
+
+func (l *Lexer) AcceptNotRun(invalid string, greedy bool) {
+	for ch := l.Next(); !strings.ContainsRune(invalid, ch) && ch != 0; ch = l.Next() {
+	}
+	if !greedy {
+		l.Back()
+	}
+}