From 3d56a5d5f1c9a802938f5f8b564ea75e577a6f1e Mon Sep 17 00:00:00 2001
From: zhouzhihong <zhouzhihong@ijunhai.com>
Date: Thu, 25 Aug 2022 14:33:22 +0800
Subject: [PATCH] Add text line and text column calculator.

---
 frontend/frontend_test.go |  2 +-
 lexer.go                  | 48 +++++++++++++++++++++++++++++++++++++++
 lexer_test.go             | 21 +----------------
 machines/machine.go       |  2 ++
 machines/machine_test.go  | 14 ++++++------
 5 files changed, 59 insertions(+), 28 deletions(-)

diff --git a/frontend/frontend_test.go b/frontend/frontend_test.go
index de345c9..22f9c07 100644
--- a/frontend/frontend_test.go
+++ b/frontend/frontend_test.go
@@ -23,7 +23,7 @@ func TestParse(x *testing.T) {
 }
 
 func tMatch(program inst.Slice, text string, t *test.T) {
-	expected := []machines.Match{{len(program) - 1, 0, 1, 1, 1, len(text), []byte(text)}}
+	expected := []machines.Match{{PC: len(program) - 1, TC: 0, StartLine: 1, StartColumn: 1, EndLine: 1, EndColumn: len(text), Bytes: []byte(text), TSLine: 1, TSColumn: 0, TELine: 1, TEColumn: 1}}
 	if expected[0].EndColumn == 0 {
 		expected[0].EndColumn = 1
 	}
diff --git a/lexer.go b/lexer.go
index 73241d7..ea6d37e 100644
--- a/lexer.go
+++ b/lexer.go
@@ -3,6 +3,7 @@ package lexmachine
 import (
 	"bytes"
 	"fmt"
+	"unicode/utf8"
 
 	dfapkg "gitea.xintech.co/zhouzhihong/lexmachine/dfa"
 	"gitea.xintech.co/zhouzhihong/lexmachine/frontend"
@@ -112,6 +113,12 @@ type Scanner struct {
 	sColumn int
 	eLine   int
 	eColumn int
+
+	lpp map[int]lastPostion
+}
+
+type lastPostion struct {
+	l, c int
 }
 
 // Next iterates through the string being scanned returning one token at a time
@@ -155,6 +162,38 @@ func (s *Scanner) Next() (tok interface{}, err error, eos bool) {
 		s.eLine = match.EndLine
 		s.eColumn = match.EndColumn
 
+		p := s.pTC
+		l, c := s.lpp[p].l, s.lpp[p].c
+		stc := s.TC - len(match.Bytes)
+
+		for {
+			if s.Text[p] == '\n' {
+				l++
+				c = 0
+			} else {
+				c++
+			}
+
+			if p == stc {
+				match.TSLine = l
+				match.TSColumn = c
+			}
+
+			match.TELine = l
+			match.TEColumn = c
+
+			_, sz := utf8.DecodeRune(s.Text[p:])
+			p += sz
+			if p >= s.TC {
+				break
+			}
+		}
+
+		s.lpp[s.TC] = lastPostion{
+			l: l,
+			c: c,
+		}
+
 		pattern := s.lexer.patterns[s.matches[match.PC]]
 		token, err = pattern.action(s, match)
 		if err != nil {
@@ -204,6 +243,7 @@ func (l *Lexer) Scanner(text []byte) (*Scanner, error) {
 			scan:    machines.DFALexerEngine(l.dfa.Start, l.dfa.Error, l.dfa.Trans, l.dfa.Accepting, textCopy),
 			Text:    textCopy,
 			TC:      0,
+			lpp:     make(map[int]lastPostion),
 		}
 	} else {
 		s = &Scanner{
@@ -212,8 +252,16 @@ func (l *Lexer) Scanner(text []byte) (*Scanner, error) {
 			scan:    machines.LexerEngine(l.program, textCopy),
 			Text:    textCopy,
 			TC:      0,
+			lpp:     make(map[int]lastPostion),
 		}
 	}
+
+	//init
+	s.lpp[0] = lastPostion{
+		l: 1,
+		c: 0,
+	}
+
 	return s, nil
 }
 
diff --git a/lexer_test.go b/lexer_test.go
index a80815e..79e9797 100644
--- a/lexer_test.go
+++ b/lexer_test.go
@@ -316,26 +316,7 @@ func TestRegression(t *testing.T) {
 func TestRegression2(t *testing.T) {
 
 	text := `# dhcpd.conf
-#
-# Sample configuration file for ISC dhcpd
-#
-
-# option definitions common to all supported networks...
-option domain-name "你好";
-option domain-name-servers ns1.example.org, ns2.example.org;
-
-default-lease-time 600;
-max-lease-time 7200;
-
-# The ddns-updates-style parameter controls whether or not the server will
-# attempt to do a DNS update when a lease is confirmed. We default to the
-# behavior of the version 2 packages ('none', since DHCP v2 didn't
-# have support for DDNS.)
-ddns-update-style none;
-
-# If this DHCP server is the official DHCP server for the local
-# network, the authoritative directive should be uncommented.
-#authoritative;
+option domain-name "你好"
 `
 
 	literals := []string{
diff --git a/machines/machine.go b/machines/machine.go
index b6325b0..e4c44e0 100644
--- a/machines/machine.go
+++ b/machines/machine.go
@@ -67,6 +67,8 @@ type Match struct {
 	EndLine     int
 	EndColumn   int
 	Bytes       []byte // the actual bytes matched during scanning.
+
+	TSLine, TSColumn, TELine, TEColumn int
 }
 
 func computeLineCol(text []byte, prevTC, tc, line, col int) (int, int) {
diff --git a/machines/machine_test.go b/machines/machine_test.go
index 1b42d05..b0ad37e 100644
--- a/machines/machine_test.go
+++ b/machines/machine_test.go
@@ -34,7 +34,7 @@ func TestLexerMatch(t *testing.T) {
 	t.Log(program)
 	mtext := []byte("ababcbcbb")
 	expected := []Match{
-		{16, 0, 1, 1, 1, len(mtext), mtext},
+		{16, 0, 1, 1, 1, len(mtext), mtext, 1, 0, 1, 1},
 	}
 	i := 0
 	for tc, m, err, scan := LexerEngine(program, text)(0); scan != nil; tc, m, err, scan = scan(tc) {
@@ -114,9 +114,9 @@ func TestLexerThreeStrings(t *testing.T) {
 	t.Log(len(text))
 	t.Log(program)
 	expected := []Match{
-		{8, 0, 1, 1, 1, 6, []byte("struct")},
-		{13, 6, 1, 7, 1, 8, []byte("  ")},
-		{15, 8, 1, 9, 1, 9, []byte("*")},
+		{8, 0, 1, 1, 1, 6, []byte("struct"), 1, 0, 1, 1},
+		{13, 6, 1, 7, 1, 8, []byte("  "), 1, 0, 1, 1},
+		{15, 8, 1, 9, 1, 9, []byte("*"), 1, 0, 1, 1},
 	}
 
 	i := 0
@@ -165,9 +165,9 @@ func TestLexerRestart(t *testing.T) {
 	t.Log(len(text))
 	t.Log(program)
 	expected := []Match{
-		{8, 0, 1, 1, 1, 6, []byte("struct")},
-		{19, 6, 2, 0, 2, 2, []byte("\n  ")},
-		{21, 9, 2, 3, 2, 3, []byte("*")},
+		{8, 0, 1, 1, 1, 6, []byte("struct"), 1, 0, 1, 1},
+		{19, 6, 2, 0, 2, 2, []byte("\n  "), 1, 0, 1, 1},
+		{21, 9, 2, 3, 2, 3, []byte("*"), 1, 0, 1, 1},
 	}
 
 	check := func(m *Match, i int, err error) {