Port more NIO changes over from trunk

git-svn-id: https://svn.apache.org/repos/asf/poi/branches/NIO_32_BRANCH@1055375 13f79535-47bb-0310-9956-ffa450edef68

src/java/org/apache/poi/hssf/usermodel/HSSFWorkbook.java

@@ -249,7 +249,25 @@ public class HSSFWorkbook extends POIDocument
     public HSSFWorkbook(DirectoryNode directory, POIFSFileSystem fs, boolean preserveNodes)
             throws IOException
     {
-        super(directory, fs);
+       this(directory, preserveNodes);
+    }
+    /**
+     * given a POI POIFSFileSystem object, and a specific directory
+     *  within it, read in its Workbook and populate the high and
+     *  low level models.  If you're reading in a workbook...start here.
+     *
+     * @param directory the POI filesystem directory to process from
+     * @param preserveNodes whether to preseve other nodes, such as
+     *        macros.  This takes more memory, so only say yes if you
+     *        need to. If set, will store all of the POIFSFileSystem
+     *        in memory
+     * @see org.apache.poi.poifs.filesystem.POIFSFileSystem
+     * @exception IOException if the stream cannot be read
+     */
+    public HSSFWorkbook(DirectoryNode directory, boolean preserveNodes)
+            throws IOException
+    {
+        super(directory);
         String workbookName = getWorkbookDirEntryName(directory);
 
         this.preserveNodes = preserveNodes;

src/java/org/apache/poi/poifs/filesystem/NPOIFSDocument.java

@@ -51,7 +51,7 @@ public final class NPOIFSDocument implements POIFSViewable {
       this._property = property;
       this._filesystem = filesystem;
 
-      if(property.getSize() <= POIFSConstants.BIG_BLOCK_MINIMUM_DOCUMENT_SIZE) {
+      if(property.getSize() < POIFSConstants.BIG_BLOCK_MINIMUM_DOCUMENT_SIZE) {
          _stream = new NPOIFSStream(_filesystem.getMiniStore(), property.getStartBlock());
          _block_size = _filesystem.getMiniStore().getBlockStoreBlockSize();
       } else {

src/java/org/apache/poi/poifs/filesystem/NPOIFSFileSystem.java

@@ -107,7 +107,7 @@ public class NPOIFSFileSystem extends BlockStore
 
     /**
      * Creates a POIFSFileSystem from a <tt>File</tt>. This uses less memory than
-     *  creating from an <tt>InputStream</tt>.
+     *  creating from an <tt>InputStream</tt>. The File will be opened read-only
      *  
      * Note that with this constructor, you will need to call {@link #close()}
      *  when you're done to have the underlying file closed, as the file is
@@ -119,22 +119,71 @@ public class NPOIFSFileSystem extends BlockStore
      */
     public NPOIFSFileSystem(File file)
          throws IOException
+    {
+       this(file, true);
+    }
+    
+    /**
+     * Creates a POIFSFileSystem from a <tt>File</tt>. This uses less memory than
+     *  creating from an <tt>InputStream</tt>.
+     *  
+     * Note that with this constructor, you will need to call {@link #close()}
+     *  when you're done to have the underlying file closed, as the file is
+     *  kept open during normal operation to read the data out. 
+     *  
+     * @param file the File from which to read the data
+     *
+     * @exception IOException on errors reading, or on invalid data
+     */
+    public NPOIFSFileSystem(File file, boolean readOnly)
+         throws IOException
+    {
+       this(
+           (new RandomAccessFile(file, readOnly? "r" : "rw")).getChannel(),
+           true
+       );
+    }
+    
+    /**
+     * Creates a POIFSFileSystem from an open <tt>FileChannel</tt>. This uses 
+     *  less memory than creating from an <tt>InputStream</tt>.
+     *  
+     * Note that with this constructor, you will need to call {@link #close()}
+     *  when you're done to have the underlying Channel closed, as the channel is
+     *  kept open during normal operation to read the data out. 
+     *  
+     * @param channel the FileChannel from which to read the data
+     *
+     * @exception IOException on errors reading, or on invalid data
+     */
+    public NPOIFSFileSystem(FileChannel channel)
+         throws IOException
+    {
+       this(channel, false);
+    }
+    
+    private NPOIFSFileSystem(FileChannel channel, boolean closeChannelOnError)
+         throws IOException
     {
        this();
 
-       // Open the underlying channel
+       try {
-       FileChannel channel = (new RandomAccessFile(file, "r")).getChannel();
+          // Get the header
+          ByteBuffer headerBuffer = ByteBuffer.allocate(POIFSConstants.SMALLER_BIG_BLOCK_SIZE);
+          IOUtils.readFully(channel, headerBuffer);
           
-       // Get the header
+          // Have the header processed
-       ByteBuffer headerBuffer = ByteBuffer.allocate(POIFSConstants.SMALLER_BIG_BLOCK_SIZE);
+          _header = new HeaderBlock(headerBuffer);
-       IOUtils.readFully(channel, headerBuffer);
           
-       // Have the header processed
+          // Now process the various entries
-       _header = new HeaderBlock(headerBuffer);
+          _data = new FileBackedDataSource(channel);
-       
+          readCoreContents();
-       // Now process the various entries
+       } catch(IOException e) {
-       _data = new FileBackedDataSource(channel);
+          if(closeChannelOnError) {
-       readCoreContents();
+             channel.close();
+          }
+          throw e;
+       }
     }
     
     /**
@@ -435,7 +484,7 @@ public class NPOIFSFileSystem extends BlockStore
              // Oh joy, we need a new XBAT too...
              xbat = createBAT(offset+1, false);
              xbat.setValueAt(0, offset);
-             bat.setValueAt(offset+1, POIFSConstants.DIFAT_SECTOR_BLOCK);
+             bat.setValueAt(1, POIFSConstants.DIFAT_SECTOR_BLOCK);
              
              // Will go one place higher as XBAT added in
              offset++;

src/scratchpad/src/org/apache/poi/hslf/EncryptedSlideShow.java

@@ -54,9 +54,8 @@ public class EncryptedSlideShow
 	public static boolean checkIfEncrypted(HSLFSlideShow hss) {
 		// Easy way to check - contains a stream
 		//  "EncryptedSummary"
-		POIFSFileSystem fs = hss.getPOIFSFileSystem();
 		try {
-			fs.getRoot().getEntry("EncryptedSummary");
+			hss.getPOIFSDirectory().getEntry("EncryptedSummary");
 			return true;
 		} catch(FileNotFoundException fnfe) {
 			// Doesn't have encrypted properties

src/scratchpad/src/org/apache/poi/hslf/HSLFSlideShow.java

@@ -78,6 +78,9 @@ public final class HSLFSlideShow extends POIDocument {
 	protected POIFSFileSystem getPOIFSFileSystem() {
 		return directory.getFileSystem();
 	}
+	protected DirectoryNode getPOIFSDirectory() {
+		return directory;
+	}
 
 	/**
 	 * Constructs a Powerpoint document from fileName. Parses the document 

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

@@ -163,14 +163,28 @@ public class HWPFDocument extends POIDocument
    *  in a POIFSFileSystem, probably not the default.
    * Used typically to open embeded documents.
    *
+   * @param directory The Directory that contains the Word document.
    * @param pfilesystem The POIFSFileSystem that contains the Word document.
    * @throws IOException If there is an unexpected IOException from the passed
    *         in POIFSFileSystem.
    */
   public HWPFDocument(DirectoryNode directory, POIFSFileSystem pfilesystem) throws IOException
+  {
+    this(directory);
+  }
+  /**
+   * This constructor loads a Word document from a specific point
+   *  in a POIFSFileSystem, probably not the default.
+   * Used typically to open embeded documents.
+   *
+   * @param directory The Directory that contains the Word document.
+   * @throws IOException If there is an unexpected IOException from the passed
+   *         in POIFSFileSystem.
+   */
+  public HWPFDocument(DirectoryNode directory) throws IOException
   {
     // Sort out the hpsf properties
-	super(directory, pfilesystem);
+	super(directory);
     readProperties();
     
     // read in the main stream.

src/scratchpad/src/org/apache/poi/hwpf/extractor/WordExtractor.java

@@ -28,6 +28,7 @@ import org.apache.poi.hwpf.model.TextPiece;
 import org.apache.poi.hwpf.usermodel.HeaderStories;
 import org.apache.poi.hwpf.usermodel.Paragraph;
 import org.apache.poi.hwpf.usermodel.Range;
+import org.apache.poi.poifs.filesystem.DirectoryNode;
 import org.apache.poi.poifs.filesystem.POIFSFileSystem;
 
 /**
@@ -59,6 +60,15 @@ public class WordExtractor extends POIOLE2TextExtractor {
 		this.fs = fs;
 	}
 	
+	/**
+	 * Create a new Word Extractor
+	 * @param dir DirectoryNode containing the word file
+	 */
+	public WordExtractor(DirectoryNode dir) throws IOException {
+		this(new HWPFDocument(dir));
+		this.fs = fs;
+	}
+	
 	/**
 	 * Create a new Word Extractor
 	 * @param doc The HWPFDocument to extract from

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

@@ -16,12 +16,14 @@
 */
 package org.apache.poi.hwpf.extractor;
 
+import java.io.File;
 import java.io.FileInputStream;
 
 import junit.framework.TestCase;
 
 import org.apache.poi.hwpf.HWPFDocument;
 import org.apache.poi.poifs.filesystem.DirectoryNode;
+import org.apache.poi.poifs.filesystem.NPOIFSFileSystem;
 import org.apache.poi.poifs.filesystem.POIFSFileSystem;
 
 /**
@@ -225,4 +227,31 @@ public class TestWordExtractor extends TestCase {
     			text.indexOf("The footer, with") > -1
     	);
     }
+    
+    /**
+     * Tests that we can work with both {@link POIFSFileSystem}
+     *  and {@link NPOIFSFileSystem}
+     */
+    public void testDifferentPOIFS() throws Exception {
+		 String dirname = System.getProperty("HWPF.testdata.path");
+       File f = new File(dirname, "test2.doc");
+       
+       // Open the two filesystems
+       DirectoryNode[] files = new DirectoryNode[2];
+       files[0] = (new POIFSFileSystem(new FileInputStream(f))).getRoot();
+       files[1] = (new NPOIFSFileSystem(f)).getRoot();
+       
+       // Open directly 
+       for(DirectoryNode dir : files) {
+          WordExtractor extractor = new WordExtractor(dir);
+          assertEquals(p_text1_block, extractor.getText());
+       }
+
+       // Open via a HWPFDocument
+       for(DirectoryNode dir : files) {
+          HWPFDocument doc = new HWPFDocument(dir);
+          WordExtractor extractor = new WordExtractor(doc);
+          assertEquals(p_text1_block, extractor.getText());
+       }
+    }
 }

src/testcases/org/apache/poi/hssf/HSSFTestDataSamples.java

@@ -48,7 +48,14 @@ public final class HSSFTestDataSamples {
 	 * @return an open <tt>InputStream</tt> for the specified sample file
 	 */
 	public static InputStream openSampleFileStream(String sampleFileName) {
-		
+      File f = getSampeFile(sampleFileName);
+		try {
+			return new FileInputStream(f);
+		} catch (FileNotFoundException e) {
+			throw new RuntimeException(e);
+		}
+   }
+	public static File getSampeFile(String sampleFileName) {
 		if(!_isInitialised) {
 			try {
 				initialise();
@@ -56,16 +63,6 @@ public final class HSSFTestDataSamples {
 				_isInitialised = true;
 			}
 		}
-		if (_sampleDataIsAvaliableOnClassPath) {
-			InputStream result = openClasspathResource(sampleFileName);
-			if(result == null) {
-				throw new RuntimeException("specified test sample file '" + sampleFileName 
-						+ "' not found on the classpath");
-			}
-//			System.out.println("opening cp: " + sampleFileName);
-			// wrap to avoid temp warning method about auto-closing input stream
-			return new NonSeekableInputStream(result);
-		}
 		if (_resolvedDataDir == null) {
 			throw new RuntimeException("Must set system property '"
 					+ TEST_DATA_DIR_SYS_PROPERTY_NAME
@@ -78,11 +75,7 @@ public final class HSSFTestDataSamples {
 					+ "' not found in data dir '" + _resolvedDataDir.getAbsolutePath() + "'");
 		}
 //		System.out.println("opening " + f.getAbsolutePath());
-		try {
+      return f;
-			return new FileInputStream(f);
-		} catch (FileNotFoundException e) {
-			throw new RuntimeException(e);
-		}
 	}
 
 	private static void initialise() {

src/testcases/org/apache/poi/hssf/usermodel/TestHSSFWorkbook.java

@@ -32,6 +32,9 @@ import org.apache.poi.hssf.record.Record;
 import org.apache.poi.hssf.record.RecordFormatException;
 import org.apache.poi.hssf.record.RecordInputStream;
 import org.apache.poi.hssf.record.formula.Area3DPtg;
+import org.apache.poi.poifs.filesystem.DirectoryNode;
+import org.apache.poi.poifs.filesystem.NPOIFSFileSystem;
+import org.apache.poi.poifs.filesystem.POIFSFileSystem;
 import org.apache.poi.util.LittleEndian;
 import org.apache.poi.util.TempFile;
 /**
@@ -550,4 +553,31 @@ public final class TestHSSFWorkbook extends TestCase {
         nr = wb.getWorkbook().getNameRecord(2);
         assertEquals("Sheet2!E:F,Sheet2!$A$9:$IV$12", nr.getAreaReference(wb)); // E:F,9:12
     }
+
+    /**
+     * Tests that we can work with both {@link POIFSFileSystem}
+     *  and {@link NPOIFSFileSystem}
+     */
+    public void testDifferentPOIFS() throws Exception {
+       // Open the two filesystems
+       DirectoryNode[] files = new DirectoryNode[2];
+       files[0] = (new POIFSFileSystem(HSSFTestDataSamples.openSampleFileStream("Simple.xls"))).getRoot();
+       files[1] = (new NPOIFSFileSystem(HSSFTestDataSamples.getSampeFile("Simple.xls"))).getRoot();
+       
+       // Open without preserving nodes 
+       for(DirectoryNode dir : files) {
+          HSSFWorkbook workbook = new HSSFWorkbook(dir, false);
+          HSSFSheet sheet = workbook.getSheetAt(0);
+          HSSFCell cell = sheet.getRow(0).getCell(0);
+          assertEquals("replaceMe", cell .getRichStringCellValue().getString());
+       }
+
+       // Now re-check with preserving
+       for(DirectoryNode dir : files) {
+          HSSFWorkbook workbook = new HSSFWorkbook(dir, true);
+          HSSFSheet sheet = workbook.getSheetAt(0);
+          HSSFCell cell = sheet.getRow(0).getCell(0);
+          assertEquals("replaceMe", cell .getRichStringCellValue().getString());
+       }
+    }
 }