<changes>
<release version="3.5-beta7" date="2009-??-??">
- <action dev="POI-DEVELOPERS" type="add">47668 - Improved parsing of OOXML documents</action>
+ <action dev="POI-DEVELOPERS" type="add">45583 - Avoid exception when reading ClipboardData packet in OLE property sets</action>
<action dev="POI-DEVELOPERS" type="add">47652 - Added support for reading encrypted workbooks</action>
<action dev="POI-DEVELOPERS" type="add">47604 - Implementation of an XML to XLSX Importer using Custom XML Mapping</action>
<action dev="POI-DEVELOPERS" type="fix">47620 - Avoid FormulaParseException in XSSFWorkbook.setRepeatingRowsAndColumns when removing repeated rows and columns</action>
*/
public abstract class POIDocument {
/** Holds metadata on our document */
- protected SummaryInformation sInf;
+ private SummaryInformation sInf;
/** Holds further metadata on our document */
- protected DocumentSummaryInformation dsInf;
+ private DocumentSummaryInformation dsInf;
/** The open POIFS FileSystem that contains our document */
protected POIFSFileSystem filesystem;
/** The directory that our document lives in */
protected DirectoryNode directory;
/** For our own logging use */
- protected POILogger logger = POILogFactory.getLogger(this.getClass());
+ private final static POILogger logger = POILogFactory.getLogger(POIDocument.class);
/* Have the property streams been read yet? (Only done on-demand) */
- protected boolean initialized = false;
+ private boolean initialized = false;
protected POIDocument(DirectoryNode dir, POIFSFileSystem fs) {
* if it wasn't found
*/
protected PropertySet getPropertySet(String setName) {
- DocumentInputStream dis;
+ //directory can be null when creating new documents
+ if(directory == null) return null;
+
+ DocumentInputStream dis;
try {
// Find the entry, and get an input stream for it
dis = directory.createDocumentInputStream(setName);
* @param writtenEntries a list of POIFS entries to add the property names too
*/
protected void writeProperties(POIFSFileSystem outFS, List writtenEntries) throws IOException {
- if(sInf != null) {
- writePropertySet(SummaryInformation.DEFAULT_STREAM_NAME,sInf,outFS);
+ SummaryInformation si = getSummaryInformation();
+ if(si != null) {
+ writePropertySet(SummaryInformation.DEFAULT_STREAM_NAME, si, outFS);
if(writtenEntries != null) {
writtenEntries.add(SummaryInformation.DEFAULT_STREAM_NAME);
}
}
- if(dsInf != null) {
- writePropertySet(DocumentSummaryInformation.DEFAULT_STREAM_NAME,dsInf,outFS);
+ DocumentSummaryInformation dsi = getDocumentSummaryInformation();
+ if(dsi != null) {
+ writePropertySet(DocumentSummaryInformation.DEFAULT_STREAM_NAME, dsi, outFS);
if(writtenEntries != null) {
writtenEntries.add(DocumentSummaryInformation.DEFAULT_STREAM_NAME);
}
{
ple = propertyList.get(propertyCount - 1);
ple.length = size - ple.offset;
- if (ple.length <= 0)
- {
- final StringBuffer b = new StringBuffer();
- b.append("The property set claims to have a size of ");
- b.append(size);
- b.append(" bytes. However, it exceeds ");
- b.append(ple.offset);
- b.append(" bytes.");
- throw new IllegalPropertySetDataException(b.toString());
- }
}
/* Look for the codepage. */
}
case Variant.VT_CF:
{
+ if(l1 < 0) {
+ /**
+ * YK: reading the ClipboardData packet (VT_CF) is not quite correct.
+ * The size of the data is determined by the first four bytes of the packet
+ * while the current implementation calculates it in the Section constructor.
+ * Test files in Bugzilla 42726 and 45583 clearly show that this approach does not always work.
+ * The workaround below attempts to gracefully handle such cases instead of throwing exceptions.
+ *
+ * August 20, 2009
+ */
+ l1 = LittleEndian.getInt(src, o1); o1 += LittleEndian.INT_SIZE;
+ }
final byte[] v = new byte[l1];
- for (int i = 0; i < l1; i++)
- v[i] = src[(o1 + i)];
+ System.arraycopy(src, o1, v, 0, v.length);
value = v;
break;
}
}
case Variant.VT_CF:
{
- final byte[] b = (byte[]) value;
+ final byte[] b = (byte[]) value;
out.write(b);
length = b.length;
break;
text.append( getPropertiesText(dsi) );
// Now custom ones
- CustomProperties cps = dsi.getCustomProperties();
+ CustomProperties cps = dsi == null ? null : dsi.getCustomProperties();
if(cps != null) {
Iterator keys = cps.keySet().iterator();
while(keys.hasNext()) {
_docstream = new byte[docProps.getSize()];
dir.createDocumentInputStream("VisioDocument").read(_docstream);
- // Read in the common POI streams
- readProperties();
-
// Check it's really visio
String typeString = new String(_docstream, 0, 20);
if(! typeString.equals(VISIO_HEADER)) {
// Now, build records based on the PowerPoint stream
buildRecords();
- // Look for Property Streams:
- readProperties();
-
// Look for any other streams
readOtherStreams();
{
// Sort out the hpsf properties
super(directory, pfilesystem);
- readProperties();
// read in the main stream.
DocumentEntry documentProps = (DocumentEntry)
import org.apache.poi.hssf.extractor.ExcelExtractor;
import org.apache.poi.hssf.usermodel.HSSFWorkbook;
+import org.apache.poi.hssf.HSSFTestDataSamples;
import org.apache.poi.poifs.filesystem.POIFSFileSystem;
public final class TestHPSFPropertiesExtractor extends TestCase {
assertTrue(fsText.indexOf("AUTHOR = marshall") > -1);
assertTrue(fsText.indexOf("TITLE = Titel: \u00c4h") > -1);
}
+
+ public void test42726() throws Exception {
+ HPSFPropertiesExtractor ex = new HPSFPropertiesExtractor(HSSFTestDataSamples.openSampleWorkbook("42726.xls"));
+ String txt = ex.getText();
+ assertTrue(txt.indexOf("PID_AUTHOR") != -1);
+ assertTrue(txt.indexOf("PID_EDITTIME") != -1);
+ assertTrue(txt.indexOf("PID_REVNUMBER") != -1);
+ assertTrue(txt.indexOf("PID_THUMBNAIL") != -1);
+ }
}