[ttml] Rename ttml decoder to token reader and correct test data (#108)

asticode · Jun 18, 2024 · 593487a · 593487a
1 parent 35e8bad
commit 593487a
Show file tree

Hide file tree

Showing 3 changed files with 18 additions and 15 deletions.
diff --git a/testdata/example-in-breaklines.ttml b/testdata/example-in-breaklines.ttml
@@ -4,20 +4,16 @@
   <body>
     <div>
       <p xml:id="1" begin="00:00:00.000" end="00:00:01.000">
-        <span>First line<br/>
-Second line</span>
+        <span>First line<br/>Second line</span>
       </p>
       <p xml:id="2" begin="00:00:01.000" end="00:00:02.000">
         <span>Third line<br></br>Fourth line</span>
       </p>
       <p xml:id="3" begin="00:00:02.000" end="00:00:03.000">
-                Fifth line
-        <br/>
-                Sixth <span>middle</span> line
+        Fifth line<br/>Sixth <span>middle</span> line
       </p>
       <p xml:id="4" begin="00:00:03.000" end="00:00:04.000">
-        Seventh line 
-        <br></br>Eighth <span>middle</span> line
+        Seventh line <br></br>Eighth <span>middle</span> line
       </p>
     </div>
   </body>

diff --git a/testdata/example-out-breaklines.ttml b/testdata/example-out-breaklines.ttml
@@ -11,7 +11,9 @@
                 <span>Second line</span>
             </p>
             <p begin="00:00:01.000" end="00:00:02.000">
-                <span>Third lineFourth line</span>
+                <span>Third line</span>
+                <br></br>
+                <span>Fourth line</span>
             </p>
             <p begin="00:00:02.000" end="00:00:03.000">
                 <span>Fifth line</span>

diff --git a/ttml.go b/ttml.go
@@ -197,22 +197,22 @@ func (i *TTMLInItems) UnmarshalXML(d *xml.Decoder, start xml.StartElement) (err
 	return nil
 }
 
-type ttmlXmlDecoder struct {
-	xml.Decoder
-	holdingToken xml.Token
+type ttmlXmlTokenReader struct {
+	xmlTokenReader xml.TokenReader
+	holdingToken   xml.Token
 }
 
 // Token implements the TokenReader interface, when it meets the "br" tag, it will hold the token and return a newline
 // instead. This is to work around the fact that the go xml unmarshaler will ignore the "br" tag if it's within a
 // character data field.
-func (r *ttmlXmlDecoder) Token() (xml.Token, error) {
+func (r *ttmlXmlTokenReader) Token() (xml.Token, error) {
 	if r.holdingToken != nil {
 		returnToken := r.holdingToken
 		r.holdingToken = nil
 		return returnToken, nil
 	}
 
-	t, err := r.Decoder.Token()
+	t, err := r.xmlTokenReader.Token()
 	if err != nil {
 		return nil, err
 	}
@@ -225,8 +225,13 @@ func (r *ttmlXmlDecoder) Token() (xml.Token, error) {
 	return t, nil
 }
 
-func newTTMLXmlDecoder(ts TTMLInSubtitle) *ttmlXmlDecoder {
-	return &ttmlXmlDecoder{Decoder: *xml.NewDecoder(strings.NewReader("<p>" + ts.Items + "</p>")), holdingToken: nil}
+func newTTMLXmlDecoder(ts TTMLInSubtitle) *xml.Decoder {
+	return xml.NewTokenDecoder(
+		&ttmlXmlTokenReader{
+			xmlTokenReader: xml.NewDecoder(strings.NewReader("<p>" + ts.Items + "</p>")),
+			holdingToken:   nil,
+		},
+	)
 }
 
 // TTMLInItem represents an input TTML item