/// <summary>
/// Creates an instance of this class from an embedded OLE Object. The OLE Object is expected
/// to include a stream "{01}Ole10Native" which contains the actual
/// data relevant for this class.
/// </summary>
/// <param name="directory">directory POI Filesystem object</param>
/// <returns>Returns an instance of this class</returns>
public static Ole10Native CreateFromEmbeddedOleObject(DirectoryNode directory)
{
DocumentEntry nativeEntry =
(DocumentEntry)directory.GetEntry(OLE10_NATIVE);
byte[] data = new byte[nativeEntry.Size];
directory.CreateDocumentInputStream(nativeEntry).Read(data);
return(new Ole10Native(data, 0));
}
/// <summary>
/// Creates an instance of this class from an embedded OLE Object. The OLE Object is expected
/// to include a stream "{01}Ole10Native" which contains the actual
/// data relevant for this class.
/// </summary>
/// <param name="directory">directory POI Filesystem object</param>
/// <returns>Returns an instance of this class</returns>
public static Ole10Native CreateFromEmbeddedOleObject(DirectoryNode directory)
{
bool plain = false;
try
{
directory.GetEntry("\u0001Ole10ItemName");
plain = true;
}
catch (FileNotFoundException)
{
plain = false;
}
DocumentEntry nativeEntry =
(DocumentEntry)directory.GetEntry(OLE10_NATIVE);
byte[] data = new byte[nativeEntry.Size];
directory.CreateDocumentInputStream(nativeEntry).Read(data);
return new Ole10Native(data, 0, plain);
}
/// <summary>
/// Creates an instance of this class from an embedded OLE Object. The OLE Object is expected
/// to include a stream "{01}Ole10Native" which contains the actual
/// data relevant for this class.
/// </summary>
/// <param name="directory">directory POI Filesystem object</param>
/// <returns>Returns an instance of this class</returns>
public static Ole10Native CreateFromEmbeddedOleObject(DirectoryNode directory)
{
bool plain = false;
try
{
directory.GetEntry("\u0001Ole10ItemName");
plain = true;
}
catch (FileNotFoundException)
{
plain = false;
}
DocumentEntry nativeEntry =
(DocumentEntry)directory.GetEntry(OLE10_NATIVE);
byte[] data = new byte[nativeEntry.Size];
directory.CreateDocumentInputStream(nativeEntry).Read(data);
return(new Ole10Native(data, 0));
}
private bool HasEntry(DirectoryNode dirNode, String entryName)
{
try
{
dirNode.GetEntry(entryName);
return true;
}
catch (FileNotFoundException)
{
return false;
}
}
/// <summary>
/// Initializes a new instance of the <see cref="HWPFDocument"/> class.
/// </summary>
/// <param name="directory">The directory.</param>
public HWPFDocument(DirectoryNode directory)
: base(directory)
{
_endnotes = new NotesImpl(_endnotesTables);
_footnotes = new NotesImpl(_footnotesTables);
// Load the main stream and FIB
// Also handles HPSF bits
// Do the CP Split
_cpSplit = new CPSplitCalculator(_fib);
// Is this document too old for us?
if (_fib.GetNFib() < 106)
{
throw new OldWordFileFormatException("The document is too old - Word 95 or older. Try HWPFOldDocument instead?");
}
// use the fib to determine the name of the table stream.
String name = "0Table";
if (_fib.IsFWhichTblStm())
{
name = "1Table";
}
// Grab the table stream.
DocumentEntry tableProps;
try
{
tableProps =
(DocumentEntry)directory.GetEntry(name);
}
catch (FileNotFoundException)
{
throw new InvalidOperationException("Table Stream '" + name + "' wasn't found - Either the document is corrupt, or is Word95 (or earlier)");
}
// read in the table stream.
_tableStream = new byte[tableProps.Size];
directory.CreatePOIFSDocumentReader(name).Read(_tableStream);
_fib.FillVariableFields(_mainStream, _tableStream);
// read in the data stream.
try
{
DocumentEntry dataProps =
(DocumentEntry)directory.GetEntry("Data");
_dataStream = new byte[dataProps.Size];
directory.CreatePOIFSDocumentReader("Data").Read(_dataStream);
}
catch (FileNotFoundException)
{
_dataStream = new byte[0];
}
// Get the cp of the start of text in the main stream
// The latest spec doc says this is always zero!
int fcMin = 0;
//fcMin = _fib.GetFcMin()
// Start to load up our standard structures.
_dop = new DocumentProperties(_tableStream, _fib.GetFcDop());
_cft = new ComplexFileTable(_mainStream, _tableStream, _fib.GetFcClx(), fcMin);
TextPieceTable _tpt = _cft.GetTextPieceTable();
// Now load the rest of the properties, which need to be adjusted
// for where text really begin
_cbt = new CHPBinTable(_mainStream, _tableStream, _fib.GetFcPlcfbteChpx(), _fib.GetLcbPlcfbteChpx(), _tpt);
_pbt = new PAPBinTable(_mainStream, _tableStream, _dataStream, _fib.GetFcPlcfbtePapx(), _fib.GetLcbPlcfbtePapx(), _tpt);
_text = _tpt.Text;
/*
* in this mode we preserving PAPX/CHPX structure from file, so text may
* miss from output, and text order may be corrupted
*/
bool preserveBinTables = false;
try
{
preserveBinTables = Boolean.Parse(
ConfigurationManager.AppSettings[PROPERTY_PRESERVE_BIN_TABLES]);
}
catch (Exception)
{
// ignore;
}
if (!preserveBinTables)
{
_cbt.Rebuild(_cft);
_pbt.Rebuild(_text, _cft);
}
/*
* Property to disable text rebuilding. In this mode changing the text
* will lead to unpredictable behavior
*/
bool preserveTextTable = false;
try
{
preserveTextTable = Boolean.Parse(
ConfigurationManager.AppSettings[PROPERTY_PRESERVE_TEXT_TABLE]);
}
catch (Exception)
{
// ignore;
}
if (!preserveTextTable)
{
_cft = new ComplexFileTable();
_tpt = _cft.GetTextPieceTable();
TextPiece textPiece = new SinglentonTextPiece(_text);
_tpt.Add(textPiece);
_text = textPiece.GetStringBuilder();
}
// Read FSPA and Escher information
// _fspa = new FSPATable(_tableStream, _fib.getFcPlcspaMom(),
// _fib.getLcbPlcspaMom(), getTextTable().getTextPieces());
_fspaHeaders = new FSPATable(_tableStream, _fib,
FSPADocumentPart.HEADER);
_fspaMain = new FSPATable(_tableStream, _fib, FSPADocumentPart.MAIN);
if (_fib.GetFcDggInfo() != 0)
{
_dgg = new EscherRecordHolder(_tableStream, _fib.GetFcDggInfo(), _fib.GetLcbDggInfo());
}
else
{
_dgg = new EscherRecordHolder();
}
// read in the pictures stream
_pictures = new PicturesTable(this, _dataStream, _mainStream, _fspa, _dgg);
// And the art shapes stream
_officeArts = new ShapesTable(_tableStream, _fib);
// And escher pictures
_officeDrawingsHeaders = new OfficeDrawingsImpl(_fspaHeaders, _dgg, _mainStream);
_officeDrawingsMain = new OfficeDrawingsImpl(_fspaMain, _dgg, _mainStream);
_st = new SectionTable(_mainStream, _tableStream, _fib.GetFcPlcfsed(), _fib.GetLcbPlcfsed(), fcMin, _tpt, _cpSplit);
_ss = new StyleSheet(_tableStream, _fib.GetFcStshf());
_ft = new FontTable(_tableStream, _fib.GetFcSttbfffn(), _fib.GetLcbSttbfffn());
int listOffset = _fib.GetFcPlcfLst();
int lfoOffset = _fib.GetFcPlfLfo();
if (listOffset != 0 && _fib.GetLcbPlcfLst() != 0)
{
_lt = new ListTables(_tableStream, _fib.GetFcPlcfLst(), _fib.GetFcPlfLfo());
}
int sbtOffset = _fib.GetFcSttbSavedBy();
int sbtLength = _fib.GetLcbSttbSavedBy();
if (sbtOffset != 0 && sbtLength != 0)
{
_sbt = new SavedByTable(_tableStream, sbtOffset, sbtLength);
}
int rmarkOffset = _fib.GetFcSttbfRMark();
int rmarkLength = _fib.GetLcbSttbfRMark();
if (rmarkOffset != 0 && rmarkLength != 0)
{
_rmat = new RevisionMarkAuthorTable(_tableStream, rmarkOffset, rmarkLength);
}
_bookmarksTables = new BookmarksTables(_tableStream, _fib);
_bookmarks = new BookmarksImpl(_bookmarksTables);
_endnotesTables = new NotesTables(NoteType.ENDNOTE, _tableStream, _fib);
_endnotes = new NotesImpl(_endnotesTables);
_footnotesTables = new NotesTables(NoteType.FOOTNOTE, _tableStream, _fib);
_footnotes = new NotesImpl(_footnotesTables);
_fieldsTables = new FieldsTables(_tableStream, _fib);
_fields = new FieldsImpl(_fieldsTables);
}
private static String GetWorkbookDirEntryName(DirectoryNode directory)
{
String[] potentialNames = WORKBOOK_DIR_ENTRY_NAMES;
for (int i = 0; i < potentialNames.Length; i++)
{
String wbName = potentialNames[i];
try
{
directory.GetEntry(wbName);
return wbName;
}
catch (FileNotFoundException)
{
// continue - to try other options
}
}
// Check for previous version of file format
try
{
directory.GetEntry("Book");
throw new OldExcelFormatException("The supplied spreadsheet seems to be Excel 5.0/7.0 (BIFF5) format. "
+ "POI only supports BIFF8 format (from Excel versions 97/2000/XP/2003)");
}
catch (FileNotFoundException)
{
// fall through
}
throw new ArgumentException("The supplied POIFSFileSystem does not contain a BIFF8 'Workbook' entry. "
+ "Is it really an excel file?");
}
private void Open(DirectoryNode directory)
{
DocumentNode book = (DocumentNode)directory.GetEntry("Book");
if (book == null)
{
throw new IOException("No Excel 5/95 Book stream found");
}
ris = new RecordInputStream(directory.CreateDocumentInputStream(book));
Prepare();
}
/// <summary>
/// This constructor loads a Word document from a specific point in a POIFSFileSystem, probably not the default.Used typically to open embeded documents.
/// </summary>
/// <param name="directory">The POIFSFileSystem that Contains the Word document.</param>
/// <param name="pfilesystem">If there is an unexpected IOException from the passed in POIFSFileSystem.</param>
public HWPFDocumentCore(DirectoryNode directory)
: base(directory)
{
// Sort out the hpsf properties
// read in the main stream.
DocumentEntry documentProps = (DocumentEntry)
directory.GetEntry(STREAM_WORD_DOCUMENT);
_mainStream = new byte[documentProps.Size];
directory.CreatePOIFSDocumentReader(STREAM_WORD_DOCUMENT).Read(_mainStream);
// Create our FIB, and check for the doc being encrypted
_fib = new FileInformationBlock(_mainStream);
if (_fib.IsFEncrypted())
{
throw new EncryptedDocumentException("Cannot process encrypted word files!");
}
DirectoryEntry objectPoolEntry;
try
{
objectPoolEntry = (DirectoryEntry)directory.GetEntry(STREAM_OBJECT_POOL);
}
catch (FileNotFoundException exc)
{
objectPoolEntry = null;
}
_objectPool = new ObjectPoolImpl(objectPoolEntry);
}
public void TestEmptyConstructor()
{
POIFSFileSystem fs = new POIFSFileSystem();
DirectoryProperty property1 = new DirectoryProperty("parent");
DirectoryProperty property2 = new DirectoryProperty("child");
DirectoryNode parent = new DirectoryNode(property1, fs, null);
DirectoryNode node = new DirectoryNode(property2, fs, parent);
Assert.AreEqual(0, parent.Path.Length);
Assert.AreEqual(1, node.Path.Length);
Assert.AreEqual("child", node.Path.GetComponent(0));
// Verify that GetEntries behaves correctly
int count = 0;
IEnumerator iter = node.Entries;
while (iter.MoveNext())
{
count++;
}
Assert.AreEqual(0, count);
// Verify behavior of IsEmpty
Assert.IsTrue(node.IsEmpty);
// Verify behavior of EntryCount
Assert.AreEqual(0, node.EntryCount);
// Verify behavior of Entry
try
{
node.GetEntry("foo");
Assert.Fail("Should have caught FileNotFoundException");
}
catch (FileNotFoundException )
{
// as expected
}
// Verify behavior of isDirectoryEntry
Assert.IsTrue(node.IsDirectoryEntry);
// Verify behavior of GetName
Assert.AreEqual(property2.Name, node.Name);
// Verify behavior of isDocumentEntry
Assert.IsTrue(!node.IsDocumentEntry);
// Verify behavior of GetParent
Assert.AreEqual(parent, node.Parent);
}
public void TestNonEmptyConstructor()
{
DirectoryProperty property1 = new DirectoryProperty("parent");
DirectoryProperty property2 = new DirectoryProperty("child1");
property1.AddChild(property2);
property1.AddChild(new DocumentProperty("child2", 2000));
property2.AddChild(new DocumentProperty("child3", 30000));
DirectoryNode node = new DirectoryNode(property1, new POIFSFileSystem(), null);
// Verify that GetEntries behaves correctly
int count = 0;
IEnumerator iter = node.Entries;
while (iter.MoveNext())
{
count++;
//iter.Current;
}
Assert.AreEqual(2, count);
// Verify behavior of IsEmpty
Assert.IsTrue(!node.IsEmpty);
// Verify behavior of EntryCount
Assert.AreEqual(2, node.EntryCount);
// Verify behavior of Entry
DirectoryNode child1 = (DirectoryNode)node.GetEntry("child1");
child1.GetEntry("child3");
node.GetEntry("child2");
try
{
node.GetEntry("child3");
Assert.Fail("Should have caught FileNotFoundException");
}
catch (FileNotFoundException)
{
// as expected
}
// Verify behavior of isDirectoryEntry
Assert.IsTrue(node.IsDirectoryEntry);
// Verify behavior of GetName
Assert.AreEqual(property1.Name, node.Name);
// Verify behavior of isDocumentEntry
Assert.IsTrue(!node.IsDocumentEntry);
// Verify behavior of GetParent
Assert.IsNull(node.Parent);
}
/// <summary>
/// Creates an instance of this class from an embedded OLE Object. The OLE Object is expected
/// to include a stream "{01}Ole10Native" which contains the actual
/// data relevant for this class.
/// </summary>
/// <param name="directory">directory POI Filesystem object</param>
/// <returns>Returns an instance of this class</returns>
public static Ole10Native CreateFromEmbeddedOleObject(DirectoryNode directory)
{
DocumentEntry nativeEntry =
(DocumentEntry)directory.GetEntry(OLE10_NATIVE);
byte[] data = new byte[nativeEntry.Size];
directory.CreateDocumentInputStream(nativeEntry).Read(data);
return new Ole10Native(data, 0);
}
