public long CheckXRefOffset(long startXRefOffset, bool isLenientParsing)
{
// repair mode isn't available in non-lenient mode
if (!isLenientParsing)
{
return(startXRefOffset);
}
source.Seek(startXRefOffset);
ReadHelper.SkipSpaces(source);
if (source.Peek() == 'x' && ReadHelper.IsString(source, "xref"))
{
return(startXRefOffset);
}
if (startXRefOffset > 0)
{
if (CheckXRefStreamOffset(source, startXRefOffset, true, pool))
{
return(startXRefOffset);
}
return(CalculateXRefFixedOffset(startXRefOffset));
}
// can't find a valid offset
return(-1);
}
private void BruteForceSearchForTables(IInputBytes bytes)
{
if (bfSearchXRefTablesOffsets != null)
{
return;
}
// a pdf may contain more than one xref entry
bfSearchXRefTablesOffsets = new List <long>();
var startOffset = bytes.CurrentOffset;
bytes.Seek(MinimumSearchOffset);
// search for xref tables
while (bytes.MoveNext() && !bytes.IsAtEnd())
{
if (ReadHelper.IsString(bytes, "xref"))
{
var newOffset = bytes.CurrentOffset;
bytes.Seek(newOffset - 1);
// ensure that we don't read "startxref" instead of "xref"
if (ReadHelper.IsWhitespace(bytes.CurrentByte))
{
bfSearchXRefTablesOffsets.Add(newOffset);
}
bytes.Seek(newOffset + 4);
}
}
bytes.Seek(startOffset);
}
private void BfSearchForXRefTables()
{
if (bfSearchXRefTablesOffsets == null)
{
// a pdf may contain more than one xref entry
bfSearchXRefTablesOffsets = new List <long>();
long originOffset = source.GetPosition();
source.Seek(MinimumSearchOffset);
// search for xref tables
while (!source.IsEof())
{
if (ReadHelper.IsString(source, "xref"))
{
long newOffset = source.GetPosition();
source.Seek(newOffset - 1);
// ensure that we don't read "startxref" instead of "xref"
if (ReadHelper.IsWhitespace(source))
{
bfSearchXRefTablesOffsets.Add(newOffset);
}
source.Seek(newOffset + 4);
}
source.Read();
}
source.Seek(originOffset);
}
}
private bool checkObjectKeys(IRandomAccessRead source, CosObjectKey objectKey, long offset)
{
// there can't be any object at the very beginning of a pdf
if (offset < MINIMUM_SEARCH_OFFSET)
{
return(false);
}
long objectNr = objectKey.Number;
long objectGen = objectKey.Generation;
long originOffset = source.GetPosition();
string objectString = ObjectHelper.createObjectString(objectNr, objectGen);
try
{
source.Seek(offset);
if (ReadHelper.IsString(source, OtherEncodings.StringAsLatin1Bytes(objectString)))
{
// everything is ok, return origin object key
source.Seek(originOffset);
return(true);
}
}
catch (InvalidOperationException exception)
{
// Swallow the exception, obviously there isn't any valid object number
}
finally
{
source.Seek(originOffset);
}
// no valid object number found
return(false);
}
private void BruteForceSearchForEndOfFileMarker(IInputBytes source)
{
if (lastEndOfFileMarker != null)
{
return;
}
long startOffset = source.CurrentOffset;
source.Seek(MINIMUM_SEARCH_OFFSET);
while (!source.IsAtEnd())
{
// search for EOF marker
if (ReadHelper.IsString(source, "%%EOF"))
{
long tempMarker = source.CurrentOffset;
if (tempMarker >= source.Length)
{
lastEndOfFileMarker = tempMarker;
break;
}
try
{
source.Seek(tempMarker + 5);
// check if the following data is some valid pdf content
// which most likely indicates that the pdf is linearized,
// updated or just cut off somewhere in the middle
ReadHelper.SkipSpaces(source);
ObjectHelper.ReadObjectNumber(source);
ObjectHelper.ReadGenerationNumber(source);
}
catch (Exception)
{
// save the EOF marker as the following data is most likely some garbage
lastEndOfFileMarker = tempMarker;
}
}
source.MoveNext();
}
source.Seek(startOffset);
// no EOF marker found
if (lastEndOfFileMarker == null)
{
lastEndOfFileMarker = long.MaxValue;
}
}
private static bool CheckObjectKeys(IInputBytes bytes, IndirectReference objectKey, long offset)
{
// there can't be any object at the very beginning of a pdf
if (offset < MinimumSearchOffset)
{
return(false);
}
var objectNr = objectKey.ObjectNumber;
long objectGen = objectKey.Generation;
var originOffset = bytes.CurrentOffset;
var objectString = ObjectHelper.CreateObjectString(objectNr, objectGen);
try
{
if (offset >= bytes.Length)
{
bytes.Seek(originOffset);
return(false);
}
bytes.Seek(offset);
if (ReadHelper.IsWhitespace(bytes.CurrentByte))
{
bytes.MoveNext();
}
if (ReadHelper.IsString(bytes, objectString))
{
// everything is ok, return origin object key
bytes.Seek(originOffset);
return(true);
}
}
catch (Exception)
{
// Swallow the exception, obviously there isn't any valid object number
}
finally
{
bytes.Seek(originOffset);
}
// no valid object number found
return(false);
}
private void bfSearchForLastEOFMarker(IRandomAccessRead source)
{
if (lastEOFMarker == null)
{
long originOffset = source.GetPosition();
source.Seek(MINIMUM_SEARCH_OFFSET);
while (!source.IsEof())
{
// search for EOF marker
if (ReadHelper.IsString(source, "%%EOF"))
{
long tempMarker = source.GetPosition();
source.Seek(tempMarker + 5);
try
{
// check if the following data is some valid pdf content
// which most likely indicates that the pdf is linearized,
// updated or just cut off somewhere in the middle
ReadHelper.SkipSpaces(source);
ObjectHelper.ReadObjectNumber(source);
ObjectHelper.ReadGenerationNumber(source);
}
catch (InvalidOperationException exception)
{
// save the EOF marker as the following data is most likely some garbage
lastEOFMarker = tempMarker;
}
}
source.Read();
}
source.Seek(originOffset);
// no EOF marker found
if (lastEOFMarker == null)
{
lastEOFMarker = long.MaxValue;
}
}
}
private void BfSearchForXRefStreams(IInputBytes bytes)
{
if (bfSearchXRefStreamsOffsets != null)
{
return;
}
// a pdf may contain more than one /XRef entry
bfSearchXRefStreamsOffsets = new List <long>();
var startOffset = bytes.CurrentOffset;
bytes.Seek(MinimumSearchOffset);
// search for XRef streams
var objString = " obj";
while (bytes.MoveNext() && !bytes.IsAtEnd())
{
if (!ReadHelper.IsString(bytes, "xref"))
{
continue;
}
// search backwards for the beginning of the stream
long newOffset = -1;
long xrefOffset = bytes.CurrentOffset;
bool objFound = false;
for (var i = 1; i < 40; i++)
{
if (objFound)
{
break;
}
long currentOffset = xrefOffset - (i * 10);
if (currentOffset > 0)
{
bytes.Seek(currentOffset);
for (int j = 0; j < 10; j++)
{
if (ReadHelper.IsString(bytes, objString))
{
long tempOffset = currentOffset - 1;
bytes.Seek(tempOffset);
var generationNumber = bytes.Peek();
// is the next char a digit?
if (generationNumber.HasValue && ReadHelper.IsDigit(generationNumber.Value))
{
tempOffset--;
bytes.Seek(tempOffset);
// is the digit preceded by a space?
if (ReadHelper.IsSpace(bytes.CurrentByte))
{
int length = 0;
bytes.Seek(--tempOffset);
while (tempOffset > MinimumSearchOffset && ReadHelper.IsDigit(bytes.CurrentByte))
{
bytes.Seek(--tempOffset);
length++;
}
if (length > 0)
{
bytes.MoveNext();
newOffset = bytes.CurrentOffset;
}
}
}
objFound = true;
break;
}
currentOffset++;
bytes.MoveNext();
}
}
}
if (newOffset > -1)
{
bfSearchXRefStreamsOffsets.Add(newOffset);
}
bytes.Seek(xrefOffset + 5);
}
bytes.Seek(startOffset);
}
private void BfSearchForXRefStreams()
{
if (bfSearchXRefStreamsOffsets == null)
{
// a pdf may contain more than one /XRef entry
bfSearchXRefStreamsOffsets = new List <long>();
long originOffset = source.GetPosition();
source.Seek(MinimumSearchOffset);
// search for XRef streams
var objString = " obj";
while (!source.IsEof())
{
if (ReadHelper.IsString(source, "xref"))
{
// search backwards for the beginning of the stream
long newOffset = -1;
long xrefOffset = source.GetPosition();
bool objFound = false;
for (int i = 1; i < 40 && !objFound; i++)
{
long currentOffset = xrefOffset - (i * 10);
if (currentOffset > 0)
{
source.Seek(currentOffset);
for (int j = 0; j < 10; j++)
{
if (ReadHelper.IsString(source, objString))
{
long tempOffset = currentOffset - 1;
source.Seek(tempOffset);
int genId = source.Peek();
// is the next char a digit?
if (ReadHelper.IsDigit(genId))
{
tempOffset--;
source.Seek(tempOffset);
if (ReadHelper.IsSpace(source))
{
int length = 0;
source.Seek(--tempOffset);
while (tempOffset > MinimumSearchOffset && ReadHelper.IsDigit(source))
{
source.Seek(--tempOffset);
length++;
}
if (length > 0)
{
source.Read();
newOffset = source.GetPosition();
}
}
}
objFound = true;
break;
}
else
{
currentOffset++;
source.Read();
}
}
}
}
if (newOffset > -1)
{
bfSearchXRefStreamsOffsets.Add(newOffset);
}
source.Seek(xrefOffset + 5);
}
source.Read();
}
source.Seek(originOffset);
}
}
private void bfSearchForObjects(IRandomAccessRead source)
{
bfSearchForLastEOFMarker(source);
bfSearchCOSObjectKeyOffsets = new Dictionary <CosObjectKey, long>();
long originOffset = source.GetPosition();
long currentOffset = MINIMUM_SEARCH_OFFSET;
long lastObjectId = long.MinValue;
int lastGenID = int.MinValue;
long lastObjOffset = long.MinValue;
char[] objString = " obj".ToCharArray();
char[] endobjString = "endobj".ToCharArray();
bool endobjFound = false;
do
{
source.Seek(currentOffset);
if (ReadHelper.IsString(source, "obj"))
{
long tempOffset = currentOffset - 1;
source.Seek(tempOffset);
int genID = source.Peek();
// is the next char a digit?
if (ReadHelper.IsDigit(genID))
{
genID -= 48;
tempOffset--;
source.Seek(tempOffset);
if (ReadHelper.IsSpace(source))
{
while (tempOffset > MINIMUM_SEARCH_OFFSET && ReadHelper.IsSpace(source))
{
source.Seek(--tempOffset);
}
bool objectIDFound = false;
while (tempOffset > MINIMUM_SEARCH_OFFSET && ReadHelper.IsDigit(source))
{
source.Seek(--tempOffset);
objectIDFound = true;
}
if (objectIDFound)
{
source.Read();
long objectId = ObjectHelper.ReadObjectNumber(source);
if (lastObjOffset > 0)
{
// add the former object ID only if there was a subsequent object ID
bfSearchCOSObjectKeyOffsets[new CosObjectKey(lastObjectId, lastGenID)] = lastObjOffset;
}
lastObjectId = objectId;
lastGenID = genID;
lastObjOffset = tempOffset + 1;
currentOffset += objString.Length - 1;
endobjFound = false;
}
}
}
}
else if (ReadHelper.IsString(source, "endobj"))
{
endobjFound = true;
currentOffset += endobjString.Length - 1;
}
currentOffset++;
} while (currentOffset < lastEOFMarker && !source.IsEof());
if ((lastEOFMarker < long.MaxValue || endobjFound) && lastObjOffset > 0)
{
// if the pdf wasn't cut off in the middle or if the last object ends with a "endobj" marker
// the last object id has to be added here so that it can't get lost as there isn't any subsequent object id
bfSearchCOSObjectKeyOffsets[new CosObjectKey(lastObjectId, lastGenID)] = lastObjOffset;
}
// reestablish origin position
source.Seek(originOffset);
}
