TIKA-4244 -- improve ics detection (#1731)

(cherry picked from commit f78dc99)
apache · Apr 25, 2024 · b419cf5 · b419cf5
1 parent 0976b1f
commit b419cf5
Show file tree

Hide file tree

Showing 3 changed files with 18 additions and 1 deletion.
diff --git a/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml b/tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml
@@ -7213,7 +7213,7 @@
   <mime-type type="text/calendar">
     <magic priority="50">
       <match value="BEGIN:VCALENDAR" type="string" offset="0">
-        <match value="VERSION:2.0" type="string" offset="15:30"/>
+        <match value="\nVERSION:2.0" type="string" offset="15:360"/>
       </match>
     </magic>
     <glob pattern="*.ics"/>

diff --git a/...ndard/tika-parsers-standard-package/src/test/java/org/apache/tika/mime/TestMimeTypes.java b/...ndard/tika-parsers-standard-package/src/test/java/org/apache/tika/mime/TestMimeTypes.java
@@ -1235,6 +1235,10 @@ public void testVandICalendars() throws Exception {
         assertType("text/x-vcalendar", "testVCalendar.vcs");
         assertTypeByData("text/calendar", "testICalendar.ics");
         assertTypeByData("text/x-vcalendar", "testVCalendar.vcs");
+        //TIKA-4244
+        //this tests detection with content intervening between the BEGIN:VCALENDAR and the VERSION:2.0 entry
+        assertType("text/calendar", "testICalendar_w_prodId.ics");
+        assertTypeByData("text/calendar", "testICalendar_w_prodId.ics");
     }
 
     @Test

diff --git a/...ika-parsers-standard-package/src/test/resources/test-documents/testICalendar_w_prodId.ics b/...ika-parsers-standard-package/src/test/resources/test-documents/testICalendar_w_prodId.ics
@@ -0,0 +1,13 @@
+BEGIN:VCALENDAR
+PRODID:-//Example Corp//iCalendar Export//EN
+VERSION:2.0
+BEGIN:VEVENT
+UID:[email protected]
+DTSTAMP:20240101T080000Z
+DTSTART:20240101T100000Z
+DTEND:20240101T120000Z
+SUMMARY:Sample HTML Event
+DESCRIPTION:This is a sample event with an HTML description.
+X-ALT-DESC;FMTTYPE=text/html:<html><body><h1>Sample HTML Event</h1><p>This is a sample event with an <strong>HTML</strong> description.</p></body></html>
+END:VEVENT
+END:VCALENDAR