public CrossReferenceParser(ILog log, XrefOffsetValidator offsetValidator,
XrefCosOffsetChecker xrefCosChecker,
CrossReferenceStreamParser crossReferenceStreamParser,
CrossReferenceTableParser crossReferenceTableParser)
{
this.log = log;
this.offsetValidator = offsetValidator;
this.crossReferenceStreamParser = crossReferenceStreamParser;
this.crossReferenceTableParser = crossReferenceTableParser;
this.xrefCosChecker = xrefCosChecker;
}
public CrossReferenceTable Parse(IInputBytes bytes, bool isLenientParsing, long crossReferenceLocation,
long offsetCorrection,
IPdfTokenScanner pdfScanner,
ISeekableTokenScanner tokenScanner)
{
long fixedOffset = offsetValidator.CheckXRefOffset(crossReferenceLocation, tokenScanner, bytes, isLenientParsing);
if (fixedOffset > -1)
{
crossReferenceLocation = fixedOffset;
log.Debug($"Found the first cross reference table or stream at {fixedOffset}.");
}
var table = new CrossReferenceTableBuilder();
var prevSet = new HashSet <long>();
long previousCrossReferenceLocation = crossReferenceLocation;
var missedAttempts = 0;
// Parse all cross reference tables and streams.
while (previousCrossReferenceLocation > 0 && missedAttempts < 100)
{
log.Debug($"Reading cross reference table or stream at {previousCrossReferenceLocation}.");
if (previousCrossReferenceLocation >= bytes.Length)
{
break;
}
// seek to xref table
tokenScanner.Seek(previousCrossReferenceLocation);
tokenScanner.MoveNext();
if (tokenScanner.CurrentToken is OperatorToken tableToken && tableToken.Data == "xref")
{
missedAttempts = 0;
log.Debug("Element was cross reference table.");
CrossReferenceTablePart tablePart = CrossReferenceTableParser.Parse(tokenScanner,
previousCrossReferenceLocation, isLenientParsing);
var nextOffset = tablePart.GetPreviousOffset();
if (nextOffset >= 0)
{
nextOffset += offsetCorrection;
}
previousCrossReferenceLocation = nextOffset;
DictionaryToken tableDictionary = tablePart.Dictionary;
CrossReferenceTablePart streamPart = null;
// check for a XRef stream, it may contain some object ids of compressed objects
if (tableDictionary.ContainsKey(NameToken.XrefStm))
{
log.Debug("Cross reference table contained referenced to stream. Reading the stream.");
int streamOffset = ((NumericToken)tableDictionary.Data[NameToken.XrefStm]).Int;
// check the xref stream reference
fixedOffset = offsetValidator.CheckXRefOffset(streamOffset, tokenScanner, bytes, isLenientParsing);
if (fixedOffset > -1 && fixedOffset != streamOffset)
{
log.Warn($"/XRefStm offset {streamOffset} is incorrect, corrected to {fixedOffset}");
streamOffset = (int)fixedOffset;
// Update the cross reference table to be a stream instead.
tableDictionary = tableDictionary.With(NameToken.XrefStm, new NumericToken(streamOffset));
tablePart = new CrossReferenceTablePart(tablePart.ObjectOffsets, streamOffset,
tablePart.Previous, tableDictionary, tablePart.Type);
}
// Read the stream from the table.
if (streamOffset > 0)
{
try
{
TryParseCrossReferenceStream(streamOffset, pdfScanner, out streamPart);
}
catch (InvalidOperationException ex)
{
if (isLenientParsing)
{
log.Error("Failed to parse /XRefStm at offset " + streamOffset, ex);
}
else
{
throw;
}
}
}
else
{
if (isLenientParsing)
{
log.Error("Skipped XRef stream due to a corrupt offset:" + streamOffset);
}
else
{
throw new PdfDocumentFormatException("Skipped XRef stream due to a corrupt offset:" + streamOffset);
}
}
}
table.Add(tablePart);
if (streamPart != null)
{
table.Add(streamPart);
}
}