TIKA-1782 allow XHTMLContentHandler to pass attributes of html elemen…

…t via Markus Jelsma git-svn-id: https://svn.apache.org/repos/asf/tika/trunk@1710799 13f79535-47bb-0310-9956-ffa450edef68
apache · Oct 27, 2015 · f43de5a · f43de5a
1 parent 63351d1
commit f43de5a
Show file tree

Hide file tree

Showing 3 changed files with 21 additions and 1 deletion.
diff --git a/CHANGES.txt b/CHANGES.txt
@@ -1,5 +1,8 @@
 Release 1.12 - Current Development
 
+  * Allow XHTMLContentHandler to pass attributes of html element 
+    via Markus Jelsma (TIKA-1782).
+
   * Fix regression with spacing in PPT via Andreas Beeker (TIKA-1777).
 
 

diff --git a/tika-core/src/main/java/org/apache/tika/sax/XHTMLContentHandler.java b/tika-core/src/main/java/org/apache/tika/sax/XHTMLContentHandler.java
@@ -60,7 +60,7 @@ public class XHTMLContentHandler extends SafeContentHandler {
      * skip them if they get sent to startElement/endElement by mistake.
      */
     private static final Set<String> AUTO =
-        unmodifiableSet("html", "head", "frameset");
+        unmodifiableSet("head", "frameset");
 
     /**
      * The elements that get prepended with the {@link #TAB} character.

diff --git a/tika-core/src/test/java/org/apache/tika/sax/XHTMLContentHandlerTest.java b/tika-core/src/test/java/org/apache/tika/sax/XHTMLContentHandlerTest.java
@@ -140,6 +140,23 @@ public void testAttributesOnBody() throws Exception {
 
         assertTrue(toHTMLContentHandler.toString().contains("itemscope"));
     }
+
+    @Test
+    public void testAttributesOnHtml() throws Exception {
+        ToHTMLContentHandler toHTMLContentHandler = new ToHTMLContentHandler();
+        XHTMLContentHandler xhtmlContentHandler = new XHTMLContentHandler(toHTMLContentHandler, new Metadata());
+        AttributesImpl attributes = new AttributesImpl();
+
+        attributes.addAttribute(XHTMLContentHandler.XHTML, "itemscope", "itemscope", "", "");
+        attributes.addAttribute(XHTMLContentHandler.XHTML, "itemtype", "itemtype", "", "http://schema.org/Event");
+
+        xhtmlContentHandler.startDocument();
+        xhtmlContentHandler.startElement(XHTMLContentHandler.XHTML, "html", "html", attributes);
+        xhtmlContentHandler.endElement("html");
+        xhtmlContentHandler.endDocument();
+
+        assertTrue(toHTMLContentHandler.toString().contains("itemscope"));
+    }
 
     /**
      * Return array of non-zerolength words. Splitting on whitespace will get us