From: Dominik Stadler Date: Sun, 5 Feb 2017 22:34:40 +0000 (+0000) Subject: Revert previous changes enough to restore the case in the commoncrawl-corpus which... X-Git-Tag: REL_3_16_FINAL~122 X-Git-Url: https://source.dussan.org/?a=commitdiff_plain;h=c7cf41f988fbaf029cde1e7b2a4f88a022709dc1;p=poi.git Revert previous changes enough to restore the case in the commoncrawl-corpus which is failing now, add a unit-test for this case, enable one disabled unit-test git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1781822 13f79535-47bb-0310-9956-ffa450edef68 --- diff --git a/src/java/org/apache/poi/ddf/EscherArrayProperty.java b/src/java/org/apache/poi/ddf/EscherArrayProperty.java index 0dff27a954..21a6db1601 100644 --- a/src/java/org/apache/poi/ddf/EscherArrayProperty.java +++ b/src/java/org/apache/poi/ddf/EscherArrayProperty.java @@ -24,7 +24,7 @@ import org.apache.poi.util.HexDump; import org.apache.poi.util.LittleEndian; /** - * Escher array properties are the most wierd construction ever invented + * Escher array properties are the most weird construction ever invented * with all sorts of special cases. I'm hopeful I've got them all. */ public final class EscherArrayProperty extends EscherComplexProperty implements Iterable { @@ -43,7 +43,7 @@ public final class EscherArrayProperty extends EscherComplexProperty implements /** * When reading a property from data stream remember if the complex part is empty and set this flag. */ - private boolean emptyComplexPart = false; + private boolean emptyComplexPart; public EscherArrayProperty(short id, byte[] complexData) { super(id, checkComplexData(complexData)); @@ -67,9 +67,13 @@ public final class EscherArrayProperty extends EscherComplexProperty implements } public void setNumberOfElementsInArray(int numberOfElements) { - int expectedArraySize = getArraySizeInBytes(numberOfElements, getSizeOfElements()); - resizeComplexData(expectedArraySize, getComplexData().length); - LittleEndian.putShort(getComplexData(), 0, (short)numberOfElements); + int expectedArraySize = numberOfElements * getActualSizeOfElements(getSizeOfElements()) + FIXED_SIZE; + if (expectedArraySize != getComplexData().length) { + byte[] newArray = new byte[expectedArraySize]; + System.arraycopy(getComplexData(), 0, newArray, 0, getComplexData().length); + setComplexData(newArray); + } + LittleEndian.putShort(getComplexData(), 0, (short) numberOfElements); } public int getNumberOfElementsInMemory() { @@ -77,8 +81,12 @@ public final class EscherArrayProperty extends EscherComplexProperty implements } public void setNumberOfElementsInMemory(int numberOfElements) { - int expectedArraySize = getArraySizeInBytes(numberOfElements, getSizeOfElements()); - resizeComplexData(expectedArraySize, expectedArraySize); + int expectedArraySize = numberOfElements * getActualSizeOfElements(getSizeOfElements()) + FIXED_SIZE; + if (expectedArraySize != getComplexData().length) { + byte[] newArray = new byte[expectedArraySize]; + System.arraycopy(getComplexData(), 0, newArray, 0, expectedArraySize); + setComplexData(newArray); + } LittleEndian.putShort(getComplexData(), 2, (short) numberOfElements); } @@ -89,9 +97,13 @@ public final class EscherArrayProperty extends EscherComplexProperty implements public void setSizeOfElements(int sizeOfElements) { LittleEndian.putShort( getComplexData(), 4, (short) sizeOfElements ); - int expectedArraySize = getArraySizeInBytes(getNumberOfElementsInArray(), sizeOfElements); - // Keep just the first 6 bytes. The rest is no good to us anyway. - resizeComplexData(expectedArraySize, 6); + int expectedArraySize = getNumberOfElementsInArray() * getActualSizeOfElements(getSizeOfElements()) + FIXED_SIZE; + if (expectedArraySize != getComplexData().length) { + // Keep just the first 6 bytes. The rest is no good to us anyway. + byte[] newArray = new byte[expectedArraySize]; + System.arraycopy( getComplexData(), 0, newArray, 0, 6 ); + setComplexData(newArray); + } } public byte[] getElement(int index) { @@ -108,13 +120,13 @@ public final class EscherArrayProperty extends EscherComplexProperty implements @Override public String toString() { - StringBuffer results = new StringBuffer(); + StringBuilder results = new StringBuilder(); results.append(" {EscherArrayProperty:" + '\n'); - results.append(" Num Elements: " + getNumberOfElementsInArray() + '\n'); - results.append(" Num Elements In Memory: " + getNumberOfElementsInMemory() + '\n'); - results.append(" Size of elements: " + getSizeOfElements() + '\n'); + results.append(" Num Elements: ").append(getNumberOfElementsInArray()).append('\n'); + results.append(" Num Elements In Memory: ").append(getNumberOfElementsInMemory()).append('\n'); + results.append(" Size of elements: ").append(getSizeOfElements()).append('\n'); for (int i = 0; i < getNumberOfElementsInArray(); i++) { - results.append(" Element " + i + ": " + HexDump.toHex(getElement(i)) + '\n'); + results.append(" Element ").append(i).append(": ").append(HexDump.toHex(getElement(i))).append('\n'); } results.append("}" + '\n'); @@ -150,25 +162,23 @@ public final class EscherArrayProperty extends EscherComplexProperty implements */ public int setArrayData(byte[] data, int offset) { if (emptyComplexPart){ - resizeComplexData(0, 0); - return 0; - } - - short numElements = LittleEndian.getShort(data, offset); - // LittleEndian.getShort(data, offset + 2); // numReserved - short sizeOfElements = LittleEndian.getShort(data, offset + 4); - - // TODO: this part is strange - it doesn't make sense to compare - // the size of the existing data when setting a new data array ... - int arraySize = getArraySizeInBytes(numElements, sizeOfElements); - if (arraySize - FIXED_SIZE == getComplexData().length) { - // The stored data size in the simple block excludes the header size - sizeIncludesHeaderSize = false; + setComplexData(new byte[0]); + } else { + short numElements = LittleEndian.getShort(data, offset); + // LittleEndian.getShort(data, offset + 2); // numReserved + short sizeOfElements = LittleEndian.getShort(data, offset + 4); + + // the code here seems to depend on complexData already being + // sized correctly via the constructor + int arraySize = getActualSizeOfElements(sizeOfElements) * numElements; + if (arraySize == getComplexData().length) { + // The stored data size in the simple block excludes the header size + setComplexData(new byte[arraySize + 6]); + sizeIncludesHeaderSize = false; + } + System.arraycopy(data, offset, getComplexData(), 0, getComplexData().length ); } - int cpySize = Math.min(arraySize, data.length-offset); - resizeComplexData(cpySize, 0); - System.arraycopy(data, offset, getComplexData(), 0, cpySize); - return cpySize; + return getComplexData().length; } /** @@ -199,15 +209,10 @@ public final class EscherArrayProperty extends EscherComplexProperty implements return sizeOfElements; } - private static int getArraySizeInBytes(int numberOfElements, int sizeOfElements) { - return numberOfElements * getActualSizeOfElements((short)(sizeOfElements & 0xFFFF)) + FIXED_SIZE; - } - - @Override public Iterator iterator() { return new Iterator(){ - private int idx = 0; + int idx; @Override public boolean hasNext() { return (idx < getNumberOfElementsInArray()); diff --git a/src/java/org/apache/poi/ddf/EscherComplexProperty.java b/src/java/org/apache/poi/ddf/EscherComplexProperty.java index 04712be4df..06bdaf5c37 100644 --- a/src/java/org/apache/poi/ddf/EscherComplexProperty.java +++ b/src/java/org/apache/poi/ddf/EscherComplexProperty.java @@ -28,7 +28,6 @@ import org.apache.poi.util.LittleEndian; * stored here. */ public class EscherComplexProperty extends EscherProperty { - // TODO - make private and final private byte[] _complexData; /** @@ -88,23 +87,17 @@ public class EscherComplexProperty extends EscherProperty { /** * Get the complex data value. - * + * * @return the complex bytes */ public byte[] getComplexData() { return _complexData; } - protected void resizeComplexData(int newSize, int bytesToKeep) { - if (newSize == _complexData.length) { - return; - } - byte[] newArray = new byte[newSize]; - System.arraycopy(_complexData, 0, newArray, 0, Math.min(bytesToKeep, newSize)); - _complexData = newArray; + protected void setComplexData(byte[] _complexData) { + this._complexData = _complexData; } - - + /** * Determine whether this property is equal to another property. * diff --git a/src/scratchpad/testcases/org/apache/poi/hwpf/usermodel/TestBugs.java b/src/scratchpad/testcases/org/apache/poi/hwpf/usermodel/TestBugs.java index 7a721657df..b1e02f35c9 100644 --- a/src/scratchpad/testcases/org/apache/poi/hwpf/usermodel/TestBugs.java +++ b/src/scratchpad/testcases/org/apache/poi/hwpf/usermodel/TestBugs.java @@ -903,6 +903,7 @@ public class TestBugs{ assertEquals(section2NumColumns, section.getNumColumns()); } + @Test public void testRegressionIn315beta2() { HWPFDocument hwpfDocument = HWPFTestDataSamples.openSampleFile("cap.stanford.edu_profiles_viewbiosketch_facultyid=4009&name=m_maciver.doc"); assertNotNull(hwpfDocument); @@ -929,4 +930,10 @@ public class TestBugs{ fs.close(); } } + + @Test + public void testCommonCrawlRegression() throws IOException { + HWPFDocument document = HWPFTestDataSamples.openSampleFile("ca.kwsymphony.www_education_School_Concert_Seat_Booking_Form_2011-12.doc"); + document.close(); + } } diff --git a/test-data/document/ca.kwsymphony.www_education_School_Concert_Seat_Booking_Form_2011-12.doc b/test-data/document/ca.kwsymphony.www_education_School_Concert_Seat_Booking_Form_2011-12.doc new file mode 100755 index 0000000000..97edf51167 Binary files /dev/null and b/test-data/document/ca.kwsymphony.www_education_School_Concert_Seat_Booking_Form_2011-12.doc differ