The MutableSection the sample application retrieved from
the MutablePropertySet is still empty. It contains no
properties and does not have a format ID. As you have read above the format ID of the first section in a property set
- determines the property set's type. Since our property set should become
- a SummaryInformation property set we have to set the format ID of its
- first (and only) section to
- F29F85E0-4FF9-1068-AB-91-08-00-2B-27-B3-D9. However, you
- won't have to remember that ID: HPSF has it defined as the well-known
- constant SectionIDMap.SUMMARY_INFORMATION_ID. The sample
- application writes it to the section using the
- setFormatID(byte[]) method:
F29F85E0-4FF9-1068-AB-91-08-00-2B-27-B3-D9. However, you
+ won't have to remember that ID: HPSF has it defined as the well-known
+ constant SectionIDMap.SUMMARY_INFORMATION_ID. The sample
+ application writes it to the section using the
+ setFormatID(byte[]) method:
This class copies a POI file system to a new file and compares the copy + * with the original.
+ * + *Property set streams are copied logically, i.e. the application + * establishes a {@link org.apache.poi.hpsf.PropertySet} of an original property + * set, creates a {@link org.apache.poi.hpsf.MutablePropertySet} from the + * {@link org.apache.poi.hpsf.PropertySet} and writes the + * {@link org.apache.poi.hpsf.MutablePropertySet} to the destination POI file + * system. - Streams which are no property set streams are copied bit by + * bit.
+ * + *The comparison of the POI file systems is done logically. That means that + * the two disk files containing the POI file systems do not need to be + * exactly identical. However, both POI file systems must contain the same + * files, and most of these files must be bitwise identical. Property set + * streams, however, are compared logically: they must have the same sections + * with the same attributs, and the sections must contain the same properties. + * Details like the ordering of the properties do not matter.
+ * + * @author Rainer Klute <klute@rainer-klute.de> + * @version $Id$ + * @since 2003-09-19 + */ +public class CopyCompare +{ + /** + *Runs the example program. The application expects one or two + * arguments:
+ * + *The first argument is the disk file name of the POI filesystem to + * copy.
The second argument is optional. If it is given, it is the name of + * a disk file the copy of the POI filesystem will be written to. If it is + * not given, the copy will be written to a temporary file which will be + * deleted at the end of the program.
Compares two {@link DirectoryEntry} instances of a POI file system. + * The directories must contain the same streams with the same names and + * contents.
+ * + * @param d1 The first directory. + * @param d2 The second directory. + * @param msg The method may append human-readable comparison messages to + * this string buffer. + * @returntrue if the directories are equal, else
+ * false.
+ * @exception MarkUnsupportedException if a POI document stream does not
+ * support the mark() operation.
+ * @exception NoPropertySetStreamException if the application tries to
+ * create a property set from a POI document stream that is not a property
+ * set stream.
+ * @exception IOException if any I/O exception occurs.
+ */
+ private static boolean equal(final DirectoryEntry d1,
+ final DirectoryEntry d2,
+ final StringBuffer msg)
+ throws MarkUnsupportedException, NoPropertySetStreamException, IOException
+ {
+ boolean equal = true;
+ /* Iterate over d1 and compare each entry with its counterpart in d2. */
+ for (final Iterator i = d1.getEntries(); equal && i.hasNext();)
+ {
+ final Entry e1 = (Entry) i.next();
+ final String n1 = e1.getName();
+ Entry e2 = null;
+ try
+ {
+ e2 = d2.getEntry(n1);
+ }
+ catch (FileNotFoundException ex)
+ {
+ msg.append("Document \"" + e1 + "\" exitsts, document \"" +
+ e2 + "\" does not.\n");
+ equal = false;
+ break;
+ }
+
+ if (e1.isDirectoryEntry() && e2.isDirectoryEntry())
+ equal = equal((DirectoryEntry) e1, (DirectoryEntry) e2, msg);
+ else if (e1.isDocumentEntry() && e2.isDocumentEntry())
+ equal = equal((DocumentEntry) e1, (DocumentEntry) e2, msg);
+ else
+ {
+ msg.append("One of \"" + e1 + "\" and \"" + e2 + "\" is a " +
+ "document while the other one is a directory.\n");
+ equal = false;
+ }
+ }
+
+ /* Iterate over d2 just to make sure that there are no entries in d2
+ * that are not in d1. */
+ for (final Iterator i = d2.getEntries(); equal && i.hasNext();)
+ {
+ final Entry e2 = (Entry) i.next();
+ final String n2 = e2.getName();
+ Entry e1 = null;
+ try
+ {
+ e1 = d1.getEntry(n2);
+ }
+ catch (FileNotFoundException ex)
+ {
+ msg.append("Document \"" + e2 + "\" exitsts, document \"" +
+ e1 + "\" does not.\n");
+ equal = false;
+ break;
+ }
+ }
+ return equal;
+ }
+
+
+
+ /**
+ * Compares two {@link DocumentEntry} instances of a POI file system. + * Documents that are not property set streams must be bitwise identical. + * Property set streams must be logically equal.
+ * + * @param d1 The first document. + * @param d2 The second document. + * @param msg The method may append human-readable comparison messages to + * this string buffer. + * @returntrue if the documents are equal, else
+ * false.
+ * @exception MarkUnsupportedException if a POI document stream does not
+ * support the mark() operation.
+ * @exception NoPropertySetStreamException if the application tries to
+ * create a property set from a POI document stream that is not a property
+ * set stream.
+ * @exception IOException if any I/O exception occurs.
+ */
+ private static boolean equal(final DocumentEntry d1, final DocumentEntry d2,
+ final StringBuffer msg)
+ throws MarkUnsupportedException, NoPropertySetStreamException, IOException
+ {
+ boolean equal = true;
+ final DocumentInputStream dis1 = new DocumentInputStream(d1);
+ final DocumentInputStream dis2 = new DocumentInputStream(d2);
+ if (PropertySet.isPropertySetStream(dis1) &&
+ PropertySet.isPropertySetStream(dis2))
+ {
+ final PropertySet ps1 = PropertySetFactory.create(dis1);
+ final PropertySet ps2 = PropertySetFactory.create(dis2);
+ equal = ps1.equals(ps2);
+ if (!equal)
+ {
+ msg.append("Property sets are not equal.\n");
+ return equal;
+ }
+ }
+ else
+ {
+ int i1;
+ int i2;
+ do
+ {
+ i1 = dis1.read();
+ i2 = dis2.read();
+ if (i1 != i2)
+ {
+ equal = false;
+ msg.append("Documents are not equal.\n");
+ break;
+ }
+ }
+ while (equal && i1 == -1);
+ }
+ return true;
+ }
+
+
+
+ /**
+ * This class does all the work. Its method {@link + * #processPOIFSReaderEvent(POIFSReaderEvent)} is called for each file in + * the original POI file system. Except for property set streams it copies + * everything unmodified to the destination POI filesystem. Property set + * streams are copied by creating a new {@link PropertySet} from the + * original property set by using the {@link + * MutablePropertySet#MutablePropertySet(PropertySet) constructor.
+ */ + static class CopyFile implements POIFSReaderListener + { + String dstName; + OutputStream out; + POIFSFileSystem poiFs; + + + /** + *The constructor of a {@link CopyFile} instance creates the target + * POIFS. It also stores the name of the file the POIFS will be written + * to once it is complete.
+ * + * @param dstName The name of the disk file the destination POIFS is to + * be written to. + * @throws FileNotFoundException + */ + public CopyFile(final String dstName) + { + this.dstName = dstName; + poiFs = new POIFSFileSystem(); + } + + + /** + *The method is called by POI's eventing API for each file in the + * origin POIFS.
+ */ + public void processPOIFSReaderEvent(final POIFSReaderEvent event) + { + /* The following declarations are shortcuts for accessing the + * "event" object. */ + final POIFSDocumentPath path = event.getPath(); + final String name = event.getName(); + final DocumentInputStream stream = event.getStream(); + + Throwable t = null; + + try + { + /* Find out whether the current document is a property set + * stream or not. */ + if (PropertySet.isPropertySetStream(stream)) + { + /* Yes, the current document is a property set stream. + * Let's create a PropertySet instance from it. */ + PropertySet ps = null; + try + { + ps = PropertySetFactory.create(stream); + } + catch (NoPropertySetStreamException ex) + { + /* This exception will not be thrown because we already + * checked above. */ + } + + /* Copy the property set to the destination POI file + * system. */ + copy(poiFs, path, name, ps); + } + else + /* No, the current document is not a property set stream. We + * copy it unmodified to the destination POIFS. */ + copy(poiFs, event.getPath(), event.getName(), stream); + } + catch (MarkUnsupportedException ex) + { + t = ex; + } + catch (IOException ex) + { + t = ex; + } + catch (WritingNotSupportedException ex) + { + t = ex; + } + + /* According to the definition of the processPOIFSReaderEvent method + * we cannot pass checked exceptions to the caller. The following + * lines check whether a checked exception occured and throws an + * unchecked exception. The message of that exception is that of + * the underlying checked exception. */ + if (t != null) + { + throw new HPSFRuntimeException + ("Could not read file \"" + path + "/" + name + + "\". Reason: " + Util.toString(t)); + } + } + + + + /** + *Writes a {@link PropertySet} to a POI filesystem.
+ * + * @param poiFs The POI filesystem to write to. + * @param path The file's path in the POI filesystem. + * @param name The file's name in the POI filesystem. + * @param ps The property set to write. + */ + public void copy(final POIFSFileSystem poiFs, + final POIFSDocumentPath path, + final String name, + final PropertySet ps) + throws WritingNotSupportedException, IOException + { + final DirectoryEntry de = getPath(poiFs, path); + final MutablePropertySet mps = new MutablePropertySet(ps); + de.createDocument(name, mps.toInputStream()); + } + + + + /** + *Copies the bytes from a {@link DocumentInputStream} to a new + * stream in a POI filesystem.
+ * + * @param poiFs The POI filesystem to write to. + * @param path The source document's path. + * @param stream The stream containing the source document. + */ + public void copy(final POIFSFileSystem poiFs, + final POIFSDocumentPath path, + final String name, + final DocumentInputStream stream) throws IOException + { + final DirectoryEntry de = getPath(poiFs, path); + final ByteArrayOutputStream out = new ByteArrayOutputStream(); + int c; + while ((c = stream.read()) != -1) + out.write(c); + stream.close(); + out.close(); + final InputStream in = + new ByteArrayInputStream(out.toByteArray()); + de.createDocument(name, in); + } + + + /** + *Writes the POI file system to a disk file.
+ * + * @throws FileNotFoundException + * @throws IOException + */ + public void close() throws FileNotFoundException, IOException + { + out = new FileOutputStream(dstName); + poiFs.writeFilesystem(out); + out.close(); + } + + + + /** Contains the directory paths that have already been created in the + * output POI filesystem and maps them to their corresponding + * {@link org.apache.poi.poifs.filesystem.DirectoryNode}s. */ + private final Map paths = new HashMap(); + + + + /** + *Ensures that the directory hierarchy for a document in a POI + * fileystem is in place. When a document is to be created somewhere in + * a POI filesystem its directory must be created first. This method + * creates all directories between the POI filesystem root and the + * directory the document should belong to which do not yet exist.
+ * + *Unfortunately POI does not offer a simple method to interrogate + * the POIFS whether a certain child node (file or directory) exists in + * a directory. However, since we always start with an empty POIFS which + * contains the root directory only and since each directory in the + * POIFS is created by this method we can maintain the POIFS's directory + * hierarchy ourselves: The {@link DirectoryEntry} of each directory + * created is stored in a {@link Map}. The directories' path names map + * to the corresponding {@link DirectoryEntry} instances.
+ * + * @param poiFs The POI filesystem the directory hierarchy is created + * in, if needed. + * @param path The document's path. This method creates those directory + * components of this hierarchy which do not yet exist. + * @return The directory entry of the document path's parent. The caller + * should use this {@link DirectoryEntry} to create documents in it. + */ + public DirectoryEntry getPath(final POIFSFileSystem poiFs, + final POIFSDocumentPath path) + { + try + { + /* Check whether this directory has already been created. */ + final String s = path.toString(); + DirectoryEntry de = (DirectoryEntry) paths.get(s); + if (de != null) + /* Yes: return the corresponding DirectoryEntry. */ + return de; + + /* No: We have to create the directory - or return the root's + * DirectoryEntry. */ + int l = path.length(); + if (l == 0) + /* Get the root directory. It does not have to be created + * since it always exists in a POIFS. */ + de = poiFs.getRoot(); + else + { + /* Create a subordinate directory. The first step is to + * ensure that the parent directory exists: */ + de = getPath(poiFs, path.getParent()); + /* Now create the target directory: */ + de = de.createDirectory(path.getComponent + (path.length() - 1)); + } + paths.put(s, de); + return de; + } + catch (IOException ex) + { + /* This exception will be thrown if the directory already + * exists. However, since we have full control about directory + * creation we can ensure that this will never happen. */ + ex.printStackTrace(System.err); + throw new RuntimeException(ex); + } + } + } + +} diff --git a/src/examples/src/org/apache/poi/hpsf/examples/ReadCustomPropertySets.java b/src/examples/src/org/apache/poi/hpsf/examples/ReadCustomPropertySets.java index 93853f45ad..dc0376e5e6 100644 --- a/src/examples/src/org/apache/poi/hpsf/examples/ReadCustomPropertySets.java +++ b/src/examples/src/org/apache/poi/hpsf/examples/ReadCustomPropertySets.java @@ -70,13 +70,13 @@ import org.apache.poi.poifs.eventfilesystem.POIFSReaderListener; import org.apache.poi.util.HexDump; /** - *Sample application showing how to read a custom property set of - * a document. Call it with the document's file name as command line - * parameter.
+ *Sample application showing how to read a document's custom property set. + * Call it with the document's file name as command-line parameter.
* *Explanations can be found in the HPSF HOW-TO.
* - * @author Rainer Klute (klute@rainer-klute.de) + * @author Rainer Klute <klute@rainer-klute.de> * @version $Id$ * @since 2003-02-01 */ diff --git a/src/examples/src/org/apache/poi/hpsf/examples/ReadTitle.java b/src/examples/src/org/apache/poi/hpsf/examples/ReadTitle.java index fcd61542e1..5d4ef3b947 100644 --- a/src/examples/src/org/apache/poi/hpsf/examples/ReadTitle.java +++ b/src/examples/src/org/apache/poi/hpsf/examples/ReadTitle.java @@ -70,7 +70,8 @@ import org.apache.poi.poifs.eventfilesystem.POIFSReaderListener; * *Explanations can be found in the HPSF HOW-TO.
* - * @author Rainer Klute (klute@rainer-klute.de) + * @author Rainer Klute <klute@rainer-klute.de> * @version $Id$ * @since 2003-02-01 */ diff --git a/src/examples/src/org/apache/poi/hpsf/examples/WriteAuthorAndTitle.java b/src/examples/src/org/apache/poi/hpsf/examples/WriteAuthorAndTitle.java index d0b203dfdd..dd550dae01 100644 --- a/src/examples/src/org/apache/poi/hpsf/examples/WriteAuthorAndTitle.java +++ b/src/examples/src/org/apache/poi/hpsf/examples/WriteAuthorAndTitle.java @@ -118,7 +118,8 @@ import org.apache.poi.poifs.filesystem.POIFSFileSystem; * *Further explanations can be found in the HPSF HOW-TO.
* - * @author Rainer Klute (klute@rainer-klute.de) + * @author Rainer Klute <klute@rainer-klute.de> * @version $Id$ * @since 2003-09-01 */ diff --git a/src/examples/src/org/apache/poi/hpsf/examples/WriteTitle.java b/src/examples/src/org/apache/poi/hpsf/examples/WriteTitle.java index df85917e68..2b1367cb19 100644 --- a/src/examples/src/org/apache/poi/hpsf/examples/WriteTitle.java +++ b/src/examples/src/org/apache/poi/hpsf/examples/WriteTitle.java @@ -72,7 +72,8 @@ import org.apache.poi.poifs.filesystem.POIFSFileSystem; *This class is a simple sample application showing how to create a property * set and write it to disk.
* - * @author Rainer Klute (klute@rainer-klute.de) + * @author Rainer Klute <klute@rainer-klute.de> * @version $Id$ * @since 2003-09-12 */ diff --git a/src/java/org/apache/poi/hpsf/Section.java b/src/java/org/apache/poi/hpsf/Section.java index efe74015c3..dcbea1b60d 100644 --- a/src/java/org/apache/poi/hpsf/Section.java +++ b/src/java/org/apache/poi/hpsf/Section.java @@ -504,7 +504,77 @@ public class Section return false; if (s.getPropertyCount() != getPropertyCount()) return false; - return Util.equals(s.getProperties(), getProperties()); + + /* Compare all properties except 0 and 1 as they must be handled + * specially. */ + Property[] pa1 = new Property[getProperties().length]; + Property[] pa2 = new Property[s.getProperties().length]; + System.arraycopy(getProperties(), 0, pa1, 0, pa1.length); + System.arraycopy(s.getProperties(), 0, pa2, 0, pa2.length); + + /* Extract properties 0 and 1 and remove them from the copy of the + * arrays. */ + Property p10 = null; + Property p11; + Property p20 = null; + Property p21; + for (int i = 0; i < pa1.length; i++) + { + final long id = pa1[i].getID(); + if (id == 0) + { + p10 = pa1[i]; + pa1 = remove(pa1, i); + i--; + } + if (id == 1) + { + p11 = pa1[i]; + pa1 = remove(pa1, i); + i--; + } + } + for (int i = 0; i < pa2.length; i++) + { + final long id = pa2[i].getID(); + if (id == 0) + { + p20 = pa2[i]; + pa2 = remove(pa2, i); + i--; + } + if (id == 1) + { + p21 = pa2[i]; + pa2 = remove(pa2, i); + i--; + } + } + + boolean dictionaryEqual = true; + if (p10 != null && p20 != null) + dictionaryEqual = p10.getValue().equals(p20.getValue()); + else if (p10 != null || p20 != null) + dictionaryEqual = false; + if (!dictionaryEqual) + return false; + else + return Util.equals(pa1, pa2); + } + + + + /** + *Removes a field from a property array. The resulting array is + * compactified and returned.
+ */ + private Property[] remove(final Property[] pa, final int i) + { + final Property[] h = new Property[pa.length - 1]; + if (i > 0) + System.arraycopy(pa, 0, h, 0, i); + System.arraycopy(pa, i + 1, h, i, h.length - i); + return h; } diff --git a/src/java/org/apache/poi/hpsf/Util.java b/src/java/org/apache/poi/hpsf/Util.java index c3ec949247..dc9454c2c1 100644 --- a/src/java/org/apache/poi/hpsf/Util.java +++ b/src/java/org/apache/poi/hpsf/Util.java @@ -263,13 +263,17 @@ public class Util { for (int i1 = 0; i1 < o1.length; i1++) { + final Object obj1 = o1[i1]; boolean matchFound = false; for (int i2 = 0; !matchFound && i2 < o1.length; i2++) - if (o1[i1].equals(o2[i2])) + { + final Object obj2 = o2[i2]; + if (obj1.equals(obj2)) { matchFound = true; o2[i2] = null; } + } if (!matchFound) return false; } diff --git a/src/testcases/org/apache/poi/hpsf/data/Test0313rur.adm b/src/testcases/org/apache/poi/hpsf/data/Test0313rur.adm new file mode 100644 index 0000000000..1e1f9547be Binary files /dev/null and b/src/testcases/org/apache/poi/hpsf/data/Test0313rur.adm differ diff --git a/src/testcases/org/apache/poi/hpsf/data/TestEditTime.doc b/src/testcases/org/apache/poi/hpsf/data/TestEditTime.doc new file mode 100644 index 0000000000..8b30417dea Binary files /dev/null and b/src/testcases/org/apache/poi/hpsf/data/TestEditTime.doc differ diff --git a/src/testcases/org/apache/poi/hpsf/data/TestMickey.doc b/src/testcases/org/apache/poi/hpsf/data/TestMickey.doc new file mode 100644 index 0000000000..f62ade3161 Binary files /dev/null and b/src/testcases/org/apache/poi/hpsf/data/TestMickey.doc differ