Merge pull request LibrePDF#66 from rik86/56-html-parser-fix

56 html parser fix

pdf-html/src/test/java/com/lowagie/text/html/HtmlParserTest.java

@@ -0,0 +1,38 @@
+package com.lowagie.text.html;
+
+import static org.junit.Assert.assertNotNull;
+
+import java.io.StringReader;
+
+import org.junit.Test;
+
+import com.lowagie.text.Document;
+import com.lowagie.text.TextElementArray;
+
+public class HtmlParserTest {
+
+	/**
+	 * Bug fix scenario: a table with a space throws a {@link TextElementArray}
+	 * class cast Exception. A table without spaces is parsed correctly.
+	 */
+	@Test
+	public void testParse_tableWithNoSpaces() {
+		Document doc1 = new Document();
+		doc1.open();
+		HtmlParser.parse(doc1, new StringReader("<table><tr><td>test</td></tr></table>")); // succeeds
+		assertNotNull(doc1);
+	}
+
+	/**
+	 * Bug fix scenario: a table with a space throws a {@link TextElementArray}
+	 * class cast Exception.
+	 */
+	@Test
+	public void testParse_tableWithSpaces() {
+		Document doc1 = new Document();
+		doc1.open();
+		HtmlParser.parse(doc1, new StringReader("<table> <tr><td>test</td></tr></table>")); // was throwin exception
+		assertNotNull(doc1);
+	}
+
+}

pdf-xml/src/main/java/com/lowagie/text/xml/SAXiTextHandler.java

@@ -206,6 +206,10 @@ public void startElement(String uri, String lname, String name,
         handleStartingTags(name, attributes);
     }
 
+    private boolean isNotBlank(String text) {
+    	return text != null && !text.trim().isEmpty();
+    }
+
     /**
      * This method deals with the starting tags.
      * 
@@ -223,7 +227,7 @@ public void handleStartingTags(String name, Properties attributes) {
         }
 
         // maybe there is some meaningful data that wasn't between tags
-        if (currentChunk != null) {
+        if (currentChunk != null && isNotBlank(currentChunk.getContent())) {
             TextElementArray current;
             try {
                 current = (TextElementArray) stack.pop();