public void TestUnicodeParts()
{
HWPFDocument doc = HWPFTestDataSamples.OpenSampleFile("HeaderFooterUnicode.doc");
TextPieceTable tbl = doc.TextTable;
// In three bits, split every 512 bytes
Assert.AreEqual(3, tbl.TextPieces.Count);
TextPiece tpA = (TextPiece)tbl.TextPieces[0];
TextPiece tpB = (TextPiece)tbl.TextPieces[1];
TextPiece tpC = (TextPiece)tbl.TextPieces[2];
Assert.IsTrue(tpA.IsUnicode);
Assert.IsTrue(tpB.IsUnicode);
Assert.IsTrue(tpC.IsUnicode);
Assert.AreEqual(256, tpA.CharacterLength);
Assert.AreEqual(256, tpB.CharacterLength);
Assert.AreEqual(19, tpC.CharacterLength);
Assert.AreEqual(512, tpA.BytesLength);
Assert.AreEqual(512, tpB.BytesLength);
Assert.AreEqual(38, tpC.BytesLength);
Assert.AreEqual(0, tpA.Start);
Assert.AreEqual(256, tpA.End);
Assert.AreEqual(256, tpB.Start);
Assert.AreEqual(512, tpB.End);
Assert.AreEqual(512, tpC.Start);
Assert.AreEqual(531, tpC.End);
// Save and re-load
HWPFDocument docB = SaveAndReload(doc);
tbl = docB.TextTable;
Assert.AreEqual(3, tbl.TextPieces.Count);
tpA = (TextPiece)tbl.TextPieces[0];
tpB = (TextPiece)tbl.TextPieces[1];
tpC = (TextPiece)tbl.TextPieces[2];
Assert.IsTrue(tpA.IsUnicode);
Assert.IsTrue(tpB.IsUnicode);
Assert.IsTrue(tpC.IsUnicode);
Assert.AreEqual(256, tpA.CharacterLength);
Assert.AreEqual(256, tpB.CharacterLength);
Assert.AreEqual(19, tpC.CharacterLength);
Assert.AreEqual(512, tpA.BytesLength);
Assert.AreEqual(512, tpB.BytesLength);
Assert.AreEqual(38, tpC.BytesLength);
Assert.AreEqual(0, tpA.Start);
Assert.AreEqual(256, tpA.End);
Assert.AreEqual(256, tpB.Start);
Assert.AreEqual(512, tpB.End);
Assert.AreEqual(512, tpC.Start);
Assert.AreEqual(531, tpC.End);
}
public void TestReadWrite()
{
FileInformationBlock fib = _hWPFDocFixture._fib;
byte[] mainStream = _hWPFDocFixture._mainStream;
byte[] tableStream = _hWPFDocFixture._tableStream;
int fcMin = fib.GetFcMin();
ComplexFileTable cft = new ComplexFileTable(mainStream, tableStream, fib.GetFcClx(), fcMin);
HWPFFileSystem fileSys = new HWPFFileSystem();
cft.WriteTo(fileSys);
MemoryStream tableOut = fileSys.GetStream("1Table");
MemoryStream mainOut = fileSys.GetStream("WordDocument");
byte[] newTableStream = tableOut.ToArray();
byte[] newMainStream = mainOut.ToArray();
ComplexFileTable newCft = new ComplexFileTable(newMainStream, newTableStream, 0, 0);
TextPieceTable oldTextPieceTable = cft.GetTextPieceTable();
TextPieceTable newTextPieceTable = newCft.GetTextPieceTable();
Assert.AreEqual(oldTextPieceTable.Text.ToString(), newTextPieceTable.Text.ToString());
}
public void TestAsciiParts()
{
HWPFDocument doc = HWPFTestDataSamples.OpenSampleFile("ThreeColHeadFoot.doc");
TextPieceTable tbl = doc.TextTable;
// All ascii, so stored in one big lump
Assert.AreEqual(1, tbl.TextPieces.Count);
TextPiece tp = (TextPiece)tbl.TextPieces[0];
Assert.AreEqual(0, tp.Start);
Assert.AreEqual(339, tp.End);
Assert.AreEqual(339, tp.CharacterLength);
Assert.AreEqual(339, tp.BytesLength);
Assert.IsTrue(tp.GetStringBuilder().ToString().StartsWith("This is a sample word document"));
// Save and re-load
HWPFDocument docB = SaveAndReload(doc);
tbl = docB.TextTable;
Assert.AreEqual(1, tbl.TextPieces.Count);
tp = (TextPiece)tbl.TextPieces[0];
Assert.AreEqual(0, tp.Start);
Assert.AreEqual(339, tp.End);
Assert.AreEqual(339, tp.CharacterLength);
Assert.AreEqual(339, tp.BytesLength);
Assert.IsTrue(tp.GetStringBuilder().ToString().StartsWith("This is a sample word document"));
}
public ComplexFileTable(byte[] documentStream, byte[] tableStream, int offset, int fcMin)
{
//skips through the prms before we reach the piece table. These contain data
//for actual fast saved files
List <SprmBuffer> sprmBuffers = new List <SprmBuffer>();
//skips through the prms before we reach the piece table. These contain data
//for actual fast saved files
while (tableStream[offset] == GRPPRL_TYPE)
{
offset++;
int size = LittleEndian.GetShort(tableStream, offset);
offset += LittleEndianConsts.SHORT_SIZE;
byte[] bs = LittleEndian.GetByteArray(tableStream, offset, size);
offset += size;
SprmBuffer sprmBuffer = new SprmBuffer(bs, false, 0);
sprmBuffers.Add(sprmBuffer);
}
this._grpprls = sprmBuffers.ToArray();
if (tableStream[offset] != TEXT_PIECE_TABLE_TYPE)
{
throw new IOException("The text piece table is corrupted");
}
int pieceTableSize = LittleEndian.GetInt(tableStream, ++offset);
offset += LittleEndianConsts.INT_SIZE;
_tpt = new TextPieceTable(documentStream, tableStream, offset, pieceTableSize, fcMin);
}
/**
* Constructor used to read an old-style binTable
* in from a Word document.
*
* @param documentStream
* @param offset
* @param size
* @param fcMin
*/
public OldCHPBinTable(byte[] documentStream, int OffSet,
int size, int fcMin, TextPieceTable tpt)
{
PlexOfCps binTable = new PlexOfCps(documentStream, OffSet, size, 2);
int length = binTable.Length;
for (int x = 0; x < length; x++)
{
GenericPropertyNode node = binTable.GetProperty(x);
int pageNum = LittleEndian.GetShort(node.Bytes);
int pageOffset = POIFSConstants.SMALLER_BIG_BLOCK_SIZE * pageNum;
CHPFormattedDiskPage cfkp = new CHPFormattedDiskPage(documentStream,
pageOffset, fcMin, tpt);
int fkpSize = cfkp.Size();
for (int y = 0; y < fkpSize; y++)
{
_textRuns.Add(cfkp.GetCHPX(y));
}
}
}
public OldPAPBinTable(byte[] documentStream, int OffSet,
int size, int fcMin, TextPieceTable tpt)
{
PlexOfCps binTable = new PlexOfCps(documentStream, OffSet, size, 2);
int length = binTable.Length;
for (int x = 0; x < length; x++)
{
GenericPropertyNode node = binTable.GetProperty(x);
int pageNum = LittleEndian.GetShort(node.Bytes);
int pageOffset = POIFSConstants.SMALLER_BIG_BLOCK_SIZE * pageNum;
PAPFormattedDiskPage pfkp = new PAPFormattedDiskPage(documentStream,
documentStream, pageOffset, tpt);
int fkpSize = pfkp.Size();
for (int y = 0; y < fkpSize; y++)
{
PAPX papx = pfkp.GetPAPX(y);
_paragraphs.Add(papx);
}
}
_paragraphs.Sort((IComparer <PAPX>)PropertyNode.PAPXComparator.instance);
}
public override bool Equals(Object o)
{
TextPieceTable tpt = (TextPieceTable)o;
int size = tpt._textPieces.Count;
if (size == _textPieces.Count)
{
for (int x = 0; x < size; x++)
{
if (!tpt._textPieces[x].Equals(_textPieces[x]))
{
return false;
}
}
return true;
}
return false;
}
public void TestReadWrite()
{
TextPieceTable fakeTPT = new TextPieceTable();
FileInformationBlock fib = _hWPFDocFixture._fib;
byte[] mainStream = _hWPFDocFixture._mainStream;
byte[] tableStream = _hWPFDocFixture._tableStream;
_pAPBinTable = new PAPBinTable(mainStream, tableStream, null, fib.GetFcPlcfbtePapx(), fib.GetLcbPlcfbtePapx(), fakeTPT);
HWPFFileSystem fileSys = new HWPFFileSystem();
_pAPBinTable.WriteTo(fileSys, fakeTPT);
MemoryStream tableOut = fileSys.GetStream("1Table");
MemoryStream mainOut = fileSys.GetStream("WordDocument");
byte[] newTableStream = tableOut.ToArray();
byte[] newMainStream = mainOut.ToArray();
PAPBinTable newBinTable = new PAPBinTable(newMainStream, newTableStream, null, 0, newTableStream.Length, fakeTPT);
List <PAPX> oldTextRuns = _pAPBinTable.GetParagraphs();
List <PAPX> newTextRuns = newBinTable.GetParagraphs();
Assert.AreEqual(oldTextRuns.Count, newTextRuns.Count);
int size = oldTextRuns.Count;
for (int x = 0; x < size; x++)
{
PropertyNode oldNode = (PropertyNode)oldTextRuns[x];
PropertyNode newNode = (PropertyNode)newTextRuns[x];
Assert.IsTrue(oldNode.Equals(newNode));
}
}
/**
* This constructs a CHPFormattedDiskPage from a raw fkp (512 byte array
* read from a Word file).
*/
public CHPFormattedDiskPage(byte[] documentStream, int offset, int fcMin, TextPieceTable tpt)
: this(documentStream, offset, tpt)
{
}
public CHPBinTable(byte[] documentStream, byte[] tableStream, int offset,
int size, int fcMin, TextPieceTable tpt) : this(documentStream, tableStream, offset, size, tpt)
{
}
public SectionTable(byte[] documentStream, byte[] tableStream, int OffSet,
int size, int fcMin,
TextPieceTable tpt, int mainLength)
{
PlexOfCps sedPlex = new PlexOfCps(tableStream, OffSet, size, SED_SIZE);
this.tpt = tpt;
this._text = tpt.TextPieces;
int length = sedPlex.Length;
for (int x = 0; x < length; x++)
{
GenericPropertyNode node = sedPlex.GetProperty(x);
SectionDescriptor sed = new SectionDescriptor(node.Bytes, 0);
int fileOffset = sed.GetFc();
//int startAt = CPtoFC(node.Start);
//int endAt = CPtoFC(node.End);
int startAt = node.Start;
int endAt = node.End;
// check for the optimization
if (fileOffset == unchecked ((int)0xffffffff))
{
_sections.Add(new SEPX(sed, startAt, endAt, new byte[0]));
}
else
{
// The first short at the offset is the size of the grpprl.
int sepxSize = LittleEndian.GetShort(documentStream, fileOffset);
byte[] buf = new byte[sepxSize];
fileOffset += LittleEndianConsts.SHORT_SIZE;
Array.Copy(documentStream, fileOffset, buf, 0, buf.Length);
_sections.Add(new SEPX(sed, startAt, endAt, buf));
}
}
// Some files seem to lie about their unicode status, which
// is very very pesky. Try to work around these, but this
// is Getting on for black magic...
int mainEndsAt = mainLength;
bool matchAt = false;
bool matchHalf = false;
for (int i = 0; i < _sections.Count; i++)
{
SEPX s = _sections[i];
if (s.End == mainEndsAt)
{
matchAt = true;
}
else if (s.End == mainEndsAt || s.End == mainEndsAt - 1)
{
matchHalf = true;
}
}
if (!matchAt && matchHalf)
{
//System.err.println("Your document seemed to be mostly unicode, but the section defInition was in bytes! Trying anyway, but things may well go wrong!");
for (int i = 0; i < _sections.Count; i++)
{
SEPX s = _sections[i];
GenericPropertyNode node = sedPlex.GetProperty(i);
int startAt = node.Start;
int endAt = node.End;
s.Start = (startAt);
s.End = (endAt);
}
}
}
public OldSectionTable(byte[] documentStream, int offset,
int size, int fcMin,
TextPieceTable tpt) : this(documentStream, offset, size)
{
}
public ComplexFileTable()
{
_tpt = new TextPieceTable();
}
