bug#51686 - ConcurrentModificationException in Tika's OfficeParser

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1160137 13f79535-47bb-0310-9956-ffa450edef68

src/scratchpad/src/org/apache/poi/hwpf/HWPFDocumentCore.java

@@ -141,8 +141,7 @@ public abstract class HWPFDocumentCore extends POIDocument
    * @throws IOException If there is an unexpected IOException from the passed
    *         in POIFSFileSystem.
    */
-  public HWPFDocumentCore(DirectoryNode directory) throws IOException
+  public HWPFDocumentCore(DirectoryNode directory) throws IOException {
-  {
     // Sort out the hpsf properties
     super(directory);
 
@@ -159,19 +158,11 @@ public abstract class HWPFDocumentCore extends POIDocument
       throw new EncryptedDocumentException("Cannot process encrypted word files!");
     }
 
-        {
+    try {
-            DirectoryEntry objectPoolEntry;
+      DirectoryEntry objectPoolEntry = (DirectoryEntry) directory
-            try
-            {
-                objectPoolEntry = (DirectoryEntry) directory
               .getEntry(STREAM_OBJECT_POOL);
-            }
-            catch ( FileNotFoundException exc )
-            {
-                objectPoolEntry = directory
-                        .createDirectory( STREAM_OBJECT_POOL );
-            }
       _objectPool = new ObjectPoolImpl(objectPoolEntry);
+    } catch (FileNotFoundException exc) {
     }
   }
 

src/scratchpad/testcases/org/apache/poi/hwpf/extractor/TestWordExtractor.java

@@ -24,9 +24,13 @@ import org.apache.poi.hwpf.HWPFDocument;
 import org.apache.poi.hwpf.HWPFTestDataSamples;
 import org.apache.poi.hwpf.OldWordFileFormatException;
 import org.apache.poi.poifs.filesystem.DirectoryNode;
+import org.apache.poi.poifs.filesystem.Entry;
 import org.apache.poi.poifs.filesystem.NPOIFSFileSystem;
 import org.apache.poi.poifs.filesystem.POIFSFileSystem;
 
+import java.io.IOException;
+import java.io.InputStream;
+
 /**
  * Test the different routes to extracting text
  *
@@ -353,4 +357,21 @@ public final class TestWordExtractor extends TestCase {
           assertEquals(p_text1_block, extractor.getText());
        }
     }
+
+    public void testRootEntiesNavigation() throws IOException {
+        InputStream is = POIDataSamples.getDocumentInstance().openResourceAsStream("testWORD.doc");
+
+        POIFSFileSystem fs = new POIFSFileSystem(is);
+
+        String text = null;
+
+        for (Entry entry : fs.getRoot()) {
+            if ("WordDocument".equals(entry.getName())) {
+                WordExtractor ex = new WordExtractor(fs);
+                text = ex.getText();
+            }
+        }
+
+        assertNotNull(text);
+    }
 }