Bugzilla 53205 - Fix some parsing errors and encoding issues in HDGF

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1365638 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Yegor Kozlov 2012-07-25 15:45:09 +00:00
parent f8988b11af
commit 302a48a25b
8 changed files with 132 additions and 70 deletions

View File

@ -34,6 +34,7 @@
<changes>
<release version="3.9-beta1" date="2012-??-??">
<action dev="poi-developers" type="fix">53205 - Fix some parsing errors and encoding issues in HDGF </action>
<action dev="poi-developers" type="add">53204 - Improved performanceof PageSettingsBlock in HSSF </action>
<action dev="poi-developers" type="add">53500 - Getter for repeating rows and columns</action>
<action dev="poi-developers" type="fix">53369 - Fixed tests failing on JDK 1.7</action>

View File

@ -161,70 +161,76 @@ public final class Chunk {
continue;
}
// Process
switch(type) {
// Types 0->7 = a flat at bit 0->7
case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
int val = contents[offset] & (1<<type);
command.value = Boolean.valueOf(val > 0);
break;
case 8:
command.value = Byte.valueOf(contents[offset]);
break;
case 9:
command.value = new Double(
LittleEndian.getDouble(contents, offset)
);
break;
case 12:
// A Little Endian String
// Starts 8 bytes into the data segment
// Ends at end of data, or 00 00
// Ensure we have enough data
if(contents.length < 8) {
command.value = "";
try {
// Process
switch(type) {
// Types 0->7 = a flat at bit 0->7
case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
int val = contents[offset] & (1<<type);
command.value = Boolean.valueOf(val > 0);
break;
}
// Find the end point
int startsAt = 8;
int endsAt = startsAt;
for(int j=startsAt; j<contents.length-1 && endsAt == startsAt; j++) {
if(contents[j] == 0 && contents[j+1] == 0) {
endsAt = j;
case 8:
command.value = Byte.valueOf(contents[offset]);
break;
case 9:
command.value = new Double(
LittleEndian.getDouble(contents, offset)
);
break;
case 12:
// A Little Endian String
// Starts 8 bytes into the data segment
// Ends at end of data, or 00 00
// Ensure we have enough data
if(contents.length < 8) {
command.value = "";
break;
}
}
if(endsAt == startsAt) {
endsAt = contents.length;
}
int strLen = (endsAt-startsAt) / 2;
command.value = StringUtil.getFromUnicodeLE(contents, startsAt, strLen);
break;
case 25:
command.value = Short.valueOf(
LittleEndian.getShort(contents, offset)
);
break;
case 26:
command.value = Integer.valueOf(
LittleEndian.getInt(contents, offset)
);
break;
// Types 11 and 21 hold the offset to the blocks
case 11: case 21:
if(offset < contents.length - 3) {
int bOffset = (int)LittleEndian.getUInt(contents, offset);
BlockOffsetCommand bcmd = (BlockOffsetCommand)command;
bcmd.setOffset(bOffset);
}
break;
// Find the end point
int startsAt = 8;
int endsAt = startsAt;
for(int j=startsAt; j<contents.length-1 && endsAt == startsAt; j++) {
if(contents[j] == 0 && contents[j+1] == 0) {
endsAt = j;
}
}
if(endsAt == startsAt) {
endsAt = contents.length;
}
default:
logger.log(POILogger.INFO,
"Command of type " + type + " not processed!");
int strLen = endsAt - startsAt;
command.value = new String(contents, startsAt, strLen, header.getChunkCharset().name());
break;
case 25:
command.value = Short.valueOf(
LittleEndian.getShort(contents, offset)
);
break;
case 26:
command.value = Integer.valueOf(
LittleEndian.getInt(contents, offset)
);
break;
// Types 11 and 21 hold the offset to the blocks
case 11: case 21:
if(offset < contents.length - 3) {
int bOffset = (int)LittleEndian.getUInt(contents, offset);
BlockOffsetCommand bcmd = (BlockOffsetCommand)command;
bcmd.setOffset(bOffset);
}
break;
default:
logger.log(POILogger.INFO,
"Command of type " + type + " not processed!");
}
}
catch (Exception e) {
logger.log(POILogger.ERROR, "Unexpected error processing command, ignoring and continuing. Command: " +
command, e);
}
// Add to the array

View File

@ -19,6 +19,8 @@ package org.apache.poi.hdgf.chunks;
import org.apache.poi.util.LittleEndian;
import java.nio.charset.Charset;
/**
* A chunk header
*/
@ -80,6 +82,7 @@ public abstract class ChunkHeader {
public abstract int getSizeInBytes();
public abstract boolean hasTrailer();
public abstract boolean hasSeparator();
public abstract Charset getChunkCharset();
/**
* Returns the ID/IX of the chunk

View File

@ -17,6 +17,8 @@
package org.apache.poi.hdgf.chunks;
import java.nio.charset.Charset;
/**
* A chunk header from v11+
*/
@ -42,4 +44,9 @@ public final class ChunkHeaderV11 extends ChunkHeaderV6 {
return false;
}
@Override
public Charset getChunkCharset() {
return Charset.forName("UTF-16LE");
}
}

View File

@ -17,6 +17,8 @@
package org.apache.poi.hdgf.chunks;
import java.nio.charset.Charset;
/**
* A chunk header from v4 or v5
*/
@ -54,4 +56,9 @@ public final class ChunkHeaderV4V5 extends ChunkHeader {
// V4 and V5 never has separators
return false;
}
@Override
public Charset getChunkCharset() {
return Charset.forName("ASCII");
}
}

View File

@ -17,6 +17,8 @@
package org.apache.poi.hdgf.chunks;
import java.nio.charset.Charset;
/**
* A chunk header from v6
*/
@ -59,4 +61,9 @@ public class ChunkHeaderV6 extends ChunkHeader {
// V6 never has separators
return false;
}
@Override
public Charset getChunkCharset() {
return Charset.forName("ASCII");
}
}

View File

@ -52,19 +52,25 @@ public final class ChunkStream extends Stream {
int pos = 0;
byte[] contents = getStore().getContents();
while(pos < contents.length) {
// Ensure we have enough data to create a chunk from
int headerSize = ChunkHeader.getHeaderSize(chunkFactory.getVersion());
if(pos+headerSize <= contents.length) {
Chunk chunk = chunkFactory.createChunk(contents, pos);
chunksA.add(chunk);
try {
while(pos < contents.length) {
// Ensure we have enough data to create a chunk from
int headerSize = ChunkHeader.getHeaderSize(chunkFactory.getVersion());
if(pos+headerSize <= contents.length) {
Chunk chunk = chunkFactory.createChunk(contents, pos);
chunksA.add(chunk);
pos += chunk.getOnDiskSize();
} else {
System.err.println("Needed " + headerSize + " bytes to create the next chunk header, but only found " + (contents.length-pos) + " bytes, ignoring rest of data");
pos = contents.length;
pos += chunk.getOnDiskSize();
} else {
System.err.println("Needed " + headerSize + " bytes to create the next chunk header, but only found " + (contents.length-pos) + " bytes, ignoring rest of data");
pos = contents.length;
}
}
}
catch (Exception e)
{
System.err.println("Failed to create chunk at " + pos + ", ignoring rest of data." + e);
}
chunks = chunksA.toArray(new Chunk[chunksA.size()]);
}

View File

@ -17,6 +17,7 @@
package org.apache.poi.hdgf;
import org.apache.poi.hdgf.extractor.VisioTextExtractor;
import org.apache.poi.hdgf.streams.PointerContainingStream;
import org.apache.poi.hdgf.streams.TrailerStream;
import org.apache.poi.poifs.filesystem.POIFSFileSystem;
@ -88,4 +89,28 @@ public final class TestHDGFCore extends TestCase {
HDGFDiagram hdgf = new HDGFDiagram(fs);
assertNotNull(hdgf);
}
public void testV6NonUtf16LE() throws Exception {
fs = new POIFSFileSystem(_dgTests.openResourceAsStream("v6-non-utf16le.vsd"));
HDGFDiagram hdgf = new HDGFDiagram(fs);
assertNotNull(hdgf);
VisioTextExtractor textExtractor = new VisioTextExtractor(hdgf);
String text = textExtractor.getText().replace("\u0000", "").trim();
assertEquals("Table\n\n\nPropertySheet\n\n\n\nPropertySheetField", text);
}
public void testUtf16LE() throws Exception {
fs = new POIFSFileSystem(_dgTests.openResourceAsStream("Test_Visio-Some_Random_Text.vsd"));
HDGFDiagram hdgf = new HDGFDiagram(fs);
assertNotNull(hdgf);
VisioTextExtractor textExtractor = new VisioTextExtractor(hdgf);
String text = textExtractor.getText().trim();
assertEquals("text\nView\nTest View\nI am a test view\nSome random text, on a page", text);
}
}