Pull atom xhtml title from nested elements

The Atom spec says that any title marked with a type of "xhtml" should be
contained in a div element[1] so we need to use the full XML text when
extracting the text.

[1] https://www.rfc-editor.org/rfc/rfc4287#section-3.1

src/parser/atom.go

@@ -47,6 +47,8 @@ type atomLinks []atomLink
 func (a *atomText) Text() string {
 	if a.Type == "html" {
 		return htmlutil.ExtractText(a.Data)
+	} else if a.Type == "xhtml" {
+		return htmlutil.ExtractText(a.XML)
 	}
 	return a.Data
 }