Author: nick
Date: Wed Nov 29 06:49:40 2006
New Revision: 480585
URL: http://svn.apache.org/viewvc?view=rev&rev=480585
Log:
Support compressed pictures properly, from bug #41032
Added:
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.doc
(with props)
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.emf
(with props)
Modified:
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/FIBFieldHandler.java
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/UnhandledDataStructure.java
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/usermodel/Picture.java
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/TestHWPFPictures.java
Modified:
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/FIBFieldHandler.java
URL:
http://svn.apache.org/viewvc/jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/FIBFieldHandler.java?view=diff&rev=480585&r1=480584&r2=480585
==============================================================================
---
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/FIBFieldHandler.java
(original)
+++
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/FIBFieldHandler.java
Wed Nov 29 06:49:40 2006
@@ -25,6 +25,8 @@
import org.apache.poi.hwpf.model.io.HWPFOutputStream;
import org.apache.poi.util.LittleEndian;
+import org.apache.poi.util.POILogFactory;
+import org.apache.poi.util.POILogger;
public class FIBFieldHandler
{
@@ -122,6 +124,8 @@
public static final int STTBLISTNAMES = 91;
public static final int STTBFUSSR = 92;
+ private static POILogger log =
POILogFactory.getLogger(FIBFieldHandler.class);
+
private static final int FIELD_SIZE = LittleEndian.INT_SIZE * 2;
private HashMap _unknownMap = new HashMap();
@@ -146,9 +150,18 @@
{
if (dsSize > 0)
{
- UnhandledDataStructure unhandled = new UnhandledDataStructure(
- tableStream, dsOffset, dsSize);
- _unknownMap.put(new Integer(x), unhandled);
+ if (dsOffset + dsSize > tableStream.length)
+ {
+ log.log(POILogger.WARN, "Unhandled data structure points to
outside the buffer. " +
+ "offset = " + dsOffset + ", length = " +
dsSize +
+ ", buffer length = " + tableStream.length);
+ }
+ else
+ {
+ UnhandledDataStructure unhandled = new UnhandledDataStructure(
+ tableStream, dsOffset, dsSize);
+ _unknownMap.put(new Integer(x), unhandled);
+ }
}
}
_fields[x*2] = dsOffset;
Modified:
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/UnhandledDataStructure.java
URL:
http://svn.apache.org/viewvc/jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/UnhandledDataStructure.java?view=diff&rev=480585&r1=480584&r2=480585
==============================================================================
---
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/UnhandledDataStructure.java
(original)
+++
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/model/UnhandledDataStructure.java
Wed Nov 29 06:49:40 2006
@@ -23,7 +23,13 @@
public UnhandledDataStructure(byte[] buf, int offset, int length)
{
+// System.out.println("Yes, using my code");
_buf = new byte[length];
+ if (offset + length > buf.length)
+ {
+ throw new IndexOutOfBoundsException("buffer length is " + buf.length +
+ "but code is trying to read " +
length + " from offset " + offset);
+ }
System.arraycopy(buf, offset, _buf, 0, length);
}
Modified:
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/usermodel/Picture.java
URL:
http://svn.apache.org/viewvc/jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/usermodel/Picture.java?view=diff&rev=480585&r1=480584&r2=480585
==============================================================================
---
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/usermodel/Picture.java
(original)
+++
jakarta/poi/trunk/src/scratchpad/src/org/apache/poi/hwpf/usermodel/Picture.java
Wed Nov 29 06:49:40 2006
@@ -18,9 +18,14 @@
package org.apache.poi.hwpf.usermodel;
import org.apache.poi.util.LittleEndian;
+import org.apache.poi.util.POILogger;
+import org.apache.poi.util.POILogFactory;
import java.io.OutputStream;
import java.io.IOException;
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
+import java.util.zip.InflaterInputStream;
/**
* Represents embedded picture extracted from Word Document
@@ -28,8 +33,11 @@
*/
public class Picture
{
+ private static final POILogger log = POILogFactory.getLogger(Picture.class);
+
// public static final int FILENAME_OFFSET = 0x7C;
// public static final int FILENAME_SIZE_OFFSET = 0x6C;
+ static final int MFPMM_OFFSET = 0x6;
static final int BLOCK_TYPE_OFFSET = 0xE;
static final int PICT_HEADER_OFFSET = 0x4;
static final int UNKNOWN_HEADER_SIZE = 0x49;
@@ -41,13 +49,22 @@
public static final byte[] TIFF = new byte[]{0x49, 0x49, 0x2A, 0x00};
public static final byte[] TIFF1 = new byte[]{0x4D, 0x4D, 0x00, 0x2A};
+ public static final byte[] EMF = { 0x01, 0x00, 0x00, 0x00 };
+ public static final byte[] WMF1 = { (byte)0xD7, (byte)0xCD, (byte)0xC6,
(byte)0x9A, 0x00, 0x00 };
+ public static final byte[] WMF2 = { 0x01, 0x00, 0x09, 0x00, 0x00, 0x03 }; //
Windows 3.x
+ // TODO: DIB, PICT
+
public static final byte[] IHDR = new byte[]{'I', 'H', 'D', 'R'};
+ public static final byte[] COMPRESSED1 = { (byte)0xFE, 0x78, (byte)0xDA };
+ public static final byte[] COMPRESSED2 = { (byte)0xFE, 0x78, (byte)0x9C };
+
private int dataBlockStartOfsset;
private int pictureBytesStartOffset;
private int dataBlockSize;
private int size;
// private String fileName;
+ private byte[] rawContent;
private byte[] content;
private byte[] _dataStream;
private int aspectRatioX;
@@ -77,9 +94,12 @@
if (fillBytes)
{
- fillImageContent(_dataStream);
+ fillImageContent();
}
+ }
+ private void fillWidthHeight()
+ {
String ext = suggestFileExtension();
// trying to extract width and height from pictures content:
if ("jpg".equalsIgnoreCase(ext)) {
@@ -121,8 +141,8 @@
*/
public void writeImageContent(OutputStream out) throws IOException
{
- if (content!=null && content.length>0) {
- out.write(content, 0, size);
+ if (rawContent!=null && rawContent.length>0) {
+ out.write(rawContent, 0, size);
} else {
out.write(_dataStream, pictureBytesStartOffset, size);
}
@@ -135,11 +155,20 @@
{
if (content == null || content.length<=0)
{
- fillImageContent(this._dataStream);
+ fillImageContent();
}
return content;
}
+ public byte[] getRawContent()
+ {
+ if (rawContent == null || rawContent.length <= 0)
+ {
+ fillRawImageContent();
+ }
+ return rawContent;
+ }
+
/**
*
* @return size in bytes of the picture
@@ -171,10 +200,12 @@
*/
public String suggestFileExtension()
{
- if (content!=null && content.length>0) {
- return suggestFileExtension(content, 0);
+ String extension = suggestFileExtension(_dataStream,
pictureBytesStartOffset);
+ if ("".equals(extension)) {
+ // May be compressed. Get the uncompressed content and inspect that.
+ extension = suggestFileExtension(getContent(), 0);
}
- return suggestFileExtension(_dataStream, pictureBytesStartOffset);
+ return extension;
}
@@ -188,11 +219,16 @@
return "gif";
} else if (matchSignature(_dataStream, BMP, pictureBytesStartOffset)) {
return "bmp";
- } else if (matchSignature(_dataStream, TIFF, pictureBytesStartOffset)) {
- return "tiff";
- } else if (matchSignature(_dataStream, TIFF1, pictureBytesStartOffset)) {
+ } else if (matchSignature(_dataStream, TIFF, pictureBytesStartOffset) ||
+ matchSignature(_dataStream, TIFF1, pictureBytesStartOffset)) {
return "tiff";
+ } else if (matchSignature(content, WMF1, 0) ||
+ matchSignature(content, WMF2, 0)) {
+ return "wmf";
+ } else if (matchSignature(content, EMF, 0)) {
+ return "emf";
}
+ // TODO: DIB, PICT
return "";
}
@@ -233,10 +269,44 @@
// return fileName.trim();
// }
- private void fillImageContent(byte[] dataStream)
+ private void fillRawImageContent()
{
- this.content = new byte[size];
- System.arraycopy(dataStream, pictureBytesStartOffset, content, 0, size);
+ this.rawContent = new byte[size];
+ System.arraycopy(_dataStream, pictureBytesStartOffset, rawContent, 0,
size);
+ }
+
+ private void fillImageContent()
+ {
+ byte[] rawContent = getRawContent();
+
+ // HACK: Detect compressed images. In reality there should be some way to
determine
+ // this from the first 32 bytes, but I can't see any similarity
between all the
+ // samples I have obtained, nor any similarity in the data block
contents.
+ if (matchSignature(rawContent, COMPRESSED1, 32) ||
matchSignature(rawContent, COMPRESSED2, 32))
+ {
+ try
+ {
+ InflaterInputStream in = new InflaterInputStream(
+ new ByteArrayInputStream(rawContent, 33, rawContent.length - 33));
+ ByteArrayOutputStream out = new ByteArrayOutputStream();
+ byte[] buf = new byte[4096];
+ int readBytes;
+ while ((readBytes = in.read(buf)) > 0)
+ {
+ out.write(buf, 0, readBytes);
+ }
+ content = out.toByteArray();
+ }
+ catch (IOException e)
+ {
+ // Problems reading from the actual ByteArrayInputStream should never
happen
+ // so this will only ever be a ZipException.
+ log.log(POILogger.INFO, "Possibly corrupt compression or
non-compressed data", e);
+ }
+ } else {
+ // Raw data is not compressed.
+ content = rawContent;
+ }
}
private static int getPictureBytesStartOffset(int dataBlockStartOffset,
byte[] _dataStream, int dataBlockSize)
@@ -322,18 +392,28 @@
this.height = getBigEndianInt(_dataStream, IHDR_CHUNK_WIDTH + 4);
}
}
+
/**
* returns pixel width of the picture or -1 if dimensions determining was
failed
*/
public int getWidth()
{
+ if (width == -1)
+ {
+ fillWidthHeight();
+ }
return width;
}
+
/**
* returns pixel height of the picture or -1 if dimensions determining was
failed
*/
public int getHeight()
{
+ if (height == -1)
+ {
+ fillWidthHeight();
+ }
return height;
}
Modified:
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/TestHWPFPictures.java
URL:
http://svn.apache.org/viewvc/jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/TestHWPFPictures.java?view=diff&rev=480585&r1=480584&r2=480585
==============================================================================
---
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/TestHWPFPictures.java
(original)
+++
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/TestHWPFPictures.java
Wed Nov 29 06:49:40 2006
@@ -31,38 +31,40 @@
* @author nick
*/
public class TestHWPFPictures extends TestCase {
- private HWPFDocument docA;
- private HWPFDocument docB;
private String docAFile;
private String docBFile;
+ private String docCFile;
private String imgAFile;
private String imgBFile;
+ private String imgCFile;
protected void setUp() throws Exception {
String dirname = System.getProperty("HWPF.testdata.path");
docAFile = dirname + "/testPictures.doc";
docBFile = dirname + "/two_images.doc";
+ docCFile = dirname + "/vector_image.doc";
imgAFile = dirname + "/simple_image.jpg";
imgBFile = dirname + "/simple_image.png";
+ imgCFile = dirname + "/vector_image.emf";
}
/**
* Test just opening the files
*/
public void testOpen() throws Exception {
- docA = new HWPFDocument(new FileInputStream(docAFile));
- docB = new HWPFDocument(new FileInputStream(docBFile));
+ HWPFDocument docA = new HWPFDocument(new
FileInputStream(docAFile));
+ HWPFDocument docB = new HWPFDocument(new
FileInputStream(docBFile));
}
/**
* Test that we have the right numbers of images in each file
*/
public void testImageCount() throws Exception {
- docA = new HWPFDocument(new FileInputStream(docAFile));
- docB = new HWPFDocument(new FileInputStream(docBFile));
+ HWPFDocument docA = new HWPFDocument(new
FileInputStream(docAFile));
+ HWPFDocument docB = new HWPFDocument(new
FileInputStream(docBFile));
assertNotNull(docA.getPicturesTable());
assertNotNull(docB.getPicturesTable());
@@ -81,7 +83,7 @@
* Test that we have the right images in at least one file
*/
public void testImageData() throws Exception {
- docB = new HWPFDocument(new FileInputStream(docBFile));
+ HWPFDocument docB = new HWPFDocument(new
FileInputStream(docBFile));
PicturesTable picB = docB.getPicturesTable();
List picturesB = picB.getAllPictures();
@@ -102,6 +104,26 @@
assertBytesSame(pic1B, pic1.getContent());
assertBytesSame(pic2B, pic2.getContent());
+ }
+
+ /**
+ * Test that compressed image data is correctly returned.
+ */
+ public void testCompressedImageData() throws Exception {
+ HWPFDocument docC = new HWPFDocument(new
FileInputStream(docCFile));
+ PicturesTable picC = docC.getPicturesTable();
+ List picturesC = picC.getAllPictures();
+
+ assertEquals(1, picturesC.size());
+
+ Picture pic = (Picture)picturesC.get(0);
+ assertNotNull(pic);
+
+ // Check the same
+ byte[] picBytes = readFile(imgCFile);
+
+ assertEquals(picBytes.length, pic.getContent().length);
+ assertBytesSame(picBytes, pic.getContent());
}
Added:
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.doc
URL:
http://svn.apache.org/viewvc/jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.doc?view=auto&rev=480585
==============================================================================
Binary file - no diff available.
Propchange:
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.doc
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
Added:
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.emf
URL:
http://svn.apache.org/viewvc/jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.emf?view=auto&rev=480585
==============================================================================
Binary file - no diff available.
Propchange:
jakarta/poi/trunk/src/scratchpad/testcases/org/apache/poi/hwpf/data/vector_image.emf
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
Mailing List: http://jakarta.apache.org/site/mail2.html#poi
The Apache Jakarta POI Project: http://jakarta.apache.org/poi/
