From dbda43cc7c86906f01e6569a78b9adf3e8254cf5 Mon Sep 17 00:00:00 2001
From: gobwas <gobwas@gmail.com>
Date: Sat, 26 Dec 2015 12:14:30 +0300
Subject: [PATCH] utf-8 support

---
 lexer.go      | 27 +++++++++++++--------------
 lexer_test.go | 20 ++++++++++----------
 2 files changed, 23 insertions(+), 24 deletions(-)

diff --git a/lexer.go b/lexer.go
index e7095a4..76685c9 100644
--- a/lexer.go
+++ b/lexer.go
@@ -122,9 +122,8 @@ func (l *lexer) unread() {
 	l.runes--
 }
 
-func (l *lexer) shift(i int) {
-	l.pos += i
-	l.start = l.pos
+func (l *lexer) reset() {
+	l.pos = l.start
 	l.runes = 0
 }
 
@@ -160,7 +159,7 @@ func (l *lexer) emit(t itemType) {
 	l.width = 0
 }
 
-func (l *lexer) flush(t itemType) {
+func (l *lexer) emitMaybe(t itemType) {
 	if l.pos > l.start {
 		l.emit(t)
 	}
@@ -202,15 +201,15 @@ func lexText(l *lexer) stateFn {
 			}
 		case single:
 			l.unread()
-			l.flush(item_text)
+			l.emitMaybe(item_text)
 			return lexSingle
 		case any:
 			l.unread()
-			l.flush(item_text)
+			l.emitMaybe(item_text)
 			return lexAny
 		case range_open:
 			l.unread()
-			l.flush(item_text)
+			l.emitMaybe(item_text)
 			return lexRangeOpen
 		}
 
@@ -236,22 +235,22 @@ func lexInsideRange(l *lexer) stateFn {
 		switch c {
 		case inside_range_not:
 			// only first char makes sense
-			if l.pos-1 == l.start {
+			if l.pos-l.width == l.start {
 				l.emit(item_range_not)
 			}
 
 		case inside_range_minus:
-			if l.pos-l.start != 2 {
+			if l.runes != 2 {
 				l.errorf("unexpected length of lo char inside range")
 				return nil
 			}
 
-			l.shift(-2)
+			l.reset()
 			return lexRangeHiLo
 
 		case range_close:
 			l.unread()
-			l.flush(item_range_chars)
+			l.emitMaybe(item_range_chars)
 			return lexRangeClose
 		}
 	}
@@ -275,7 +274,7 @@ func lexRangeHiLo(l *lexer) stateFn {
 
 		switch c {
 		case inside_range_minus:
-			if l.pos-l.start != 1 {
+			if l.runes != 1 {
 				l.errorf("unexpected length of range: single character expected before minus")
 				return nil
 			}
@@ -285,7 +284,7 @@ func lexRangeHiLo(l *lexer) stateFn {
 		case range_close:
 			l.unread()
 
-			if l.pos-l.start != 1 {
+			if l.runes != 1 {
 				l.errorf("unexpected length of range: single character expected before close")
 				return nil
 			}
@@ -298,7 +297,7 @@ func lexRangeHiLo(l *lexer) stateFn {
 				continue
 			}
 
-			if l.pos-l.start != 1 {
+			if l.runes != 1 {
 				l.errorf("unexpected length of range: single character expected at the begining")
 				return nil
 			}
diff --git a/lexer_test.go b/lexer_test.go
index 69f3703..63fc5b1 100644
--- a/lexer_test.go
+++ b/lexer_test.go
@@ -42,43 +42,43 @@ func TestLexGood(t *testing.T) {
 			},
 		},
 		{
-			pattern: "[a-b]",
+			pattern: "[日-語]",
 			items: []item{
 				item{item_range_open, "["},
-				item{item_range_lo, "a"},
+				item{item_range_lo, "日"},
 				item{item_range_minus, "-"},
-				item{item_range_hi, "b"},
+				item{item_range_hi, "語"},
 				item{item_range_close, "]"},
 				item{item_eof, ""},
 			},
 		},
 		{
-			pattern: "[!a-b]",
+			pattern: "[!日-語]",
 			items: []item{
 				item{item_range_open, "["},
 				item{item_range_not, "!"},
-				item{item_range_lo, "a"},
+				item{item_range_lo, "日"},
 				item{item_range_minus, "-"},
-				item{item_range_hi, "b"},
+				item{item_range_hi, "語"},
 				item{item_range_close, "]"},
 				item{item_eof, ""},
 			},
 		},
 		{
-			pattern: "[abc]",
+			pattern: "[日本語]",
 			items: []item{
 				item{item_range_open, "["},
-				item{item_range_chars, "abc"},
+				item{item_range_chars, "日本語"},
 				item{item_range_close, "]"},
 				item{item_eof, ""},
 			},
 		},
 		{
-			pattern: "[!abc]",
+			pattern: "[!日本語]",
 			items: []item{
 				item{item_range_open, "["},
 				item{item_range_not, "!"},
-				item{item_range_chars, "abc"},
+				item{item_range_chars, "日本語"},
 				item{item_range_close, "]"},
 				item{item_eof, ""},
 			},