Handle < and > for HTML in autodescribe

These are probably the two most common character entities, so might as well convert them. Using an HTML parser would be better (or calibre) but the simple text scanning method for HTML is much more likely to have all the required dependencies available, and works just as well in most real world cases.
dfandrich · Sep 2, 2024 · cd806c5 · cd806c5
1 parent 29dc0f5
commit cd806c5
Show file tree

Hide file tree

Showing 3 changed files with 3 additions and 3 deletions.
diff --git a/autodescribe b/autodescribe
@@ -1059,7 +1059,7 @@ comment_first_line () {
 
 # File type: html (HTML text)
 comment_html () {
-	COMMENT=$(sed -n 's,^.*<[tT][iI][tT][lL][eE]\>[^>]*>\(.*\)</.*$,\1,p' < "$1" | head -1)
+	COMMENT=$(sed -n 's,^.*<[tT][iI][tT][lL][eE]\>[^>]*>\(.*\)</.*$,\1,p' < "$1" | sed -e 's/&lt;/</g' -e 's/&gt;/>/g' | head -1)
 }
 
 # File type: kdenlive (Kdenlive file)

diff --git a/test-autodescribe-expected b/test-autodescribe-expected
@@ -28,7 +28,7 @@
 'testfiles/type1.gif' 'GIF comment'
 'testfiles/type1.gnumeric' 'Gnumeric Title'
 'testfiles/type1.gpx' 'GPX metadata name'
-'testfiles/type1.html' 'HTML title is description'
+'testfiles/type1.html' 'HTML title <is> description'
 'testfiles/type1.ics' 'iCalendar summary'
 'testfiles/type1.iso' 'volume_id'
 'testfiles/type1.jar' 'Jar Application Name'

diff --git a/testfiles/type1.html b/testfiles/type1.html
@@ -1,6 +1,6 @@
 <html>
   <head>
-    <TITLE>HTML title is description</TITLE>
+	  <TITLE>HTML title &lt;is&gt; description</TITLE>
   </head>
   <body>
     Nothing to see here...