Enforce UTF-8 encoding for input bytes (#115)

* Detect UTF-8 for vtt/srt, error if encoded otherwise

* Switch to isValidUTF8Reader

* Remove isValidUTF8Reader function

* Enforce valid UTF8 on each SRT and WebVTT line

* Add benchmark tests for webvtt and srt

* Update benchmark test

* Add non-utf-8 tests and files

* Remove benchmarks

srt.go

@@ -7,6 +7,7 @@ import (
 	"strconv"
 	"strings"
 	"time"
+	"unicode/utf8"
 )
 
 // Constants
@@ -43,6 +44,10 @@ func ReadFromSRT(i io.Reader) (o *Subtitles, err error) {
 		// Fetch line
 		line = strings.TrimSpace(scanner.Text())
 		lineNum++
+		if !utf8.ValidString(line) {
+			err = fmt.Errorf("astisub: line %d is not valid utf-8", lineNum)
+			return
+		}
 
 		// Remove BOM header
 		if lineNum == 1 {

srt_test.go

@@ -46,3 +46,8 @@ func TestSRTMissingSequence(t *testing.T) {
 	assert.NoError(t, err)
 	assert.Equal(t, string(c), w.String())
 }
+
+func TestNonUTF8SRT(t *testing.T) {
+	_, err := astisub.OpenFile("./testdata/example-in-non-utf8.srt")
+	assert.Error(t, err)
+}

webvtt.go

@@ -11,6 +11,7 @@ import (
 	"strconv"
 	"strings"
 	"time"
+	"unicode/utf8"
 
 	"golang.org/x/net/html"
 )
@@ -128,6 +129,10 @@ func ReadFromWebVTT(i io.Reader) (o *Subtitles, err error) {
 		lineNum++
 		line = scanner.Text()
 		line = strings.TrimPrefix(line, string(BytesBOM))
+		if !utf8.ValidString(line) {
+			err = fmt.Errorf("astisub: line %d is not valid utf-8", lineNum)
+			return
+		}
 		if fs := strings.Fields(line); len(fs) > 0 && fs[0] == "WEBVTT" {
 			break
 		}
@@ -144,6 +149,10 @@ func ReadFromWebVTT(i io.Reader) (o *Subtitles, err error) {
 		// Fetch line
 		line = strings.TrimSpace(scanner.Text())
 		lineNum++
+		if !utf8.ValidString(line) {
+			err = fmt.Errorf("astisub: line %d is not valid utf-8", lineNum)
+			return
+		}
 
 		switch {
 		// Comment

webvtt_test.go

@@ -48,6 +48,11 @@ func TestBroken1WebVTT(t *testing.T) {
 	assert.Nil(t, err)
 }
 
+func TestNonUTF8WebVTT(t *testing.T) {
+	_, err := astisub.OpenFile("./testdata/example-in-non-utf8.vtt")
+	assert.Error(t, err)
+}
+
 func TestWebVTTWithVoiceName(t *testing.T) {
 	testData := `WEBVTT