lsm/PdfPig
1
0
Fork 0
mirror of https://github.com/UglyToad/PdfPig.git synced 2025-11-28 09:28:25 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

#443 make xref parsing more tolerant for near misses

Browse Source 
this whole thing needs a bit of a refactor but this will do for now
This commit is contained in:
Eliot Jones
2022-04-24 13:07:30 -04:00
parent 2a68670896
commit 8ca0ed3842
1 changed files with 25 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
src/UglyToad.PdfPig/Parser/FileStructure/CrossReferenceParser.cs
View File

@@ -42,7 +42,7 @@
long previousCrossReferenceLocation = crossReferenceLocation;
var missedAttempts = 0;
// Parse all cross reference tables and streams.
while (previousCrossReferenceLocation > 0 && missedAttempts < 100)
{
@@ -193,7 +193,30 @@
log.Debug($"The cross reference found at this location ({previousCrossReferenceLocation}) was not a table or stream. " +
$"Found token ({tokenScanner.CurrentToken}) ending at {tokenScanner.CurrentPosition} instead. Seeking next token.");
previousCrossReferenceLocation = tokenScanner.CurrentPosition;
var storedCurrentTokenScannerPosition = tokenScanner.CurrentPosition;
if (missedAttempts == 0)
{
// We might only be a little bit out so let's just check the neighbourhood (for tables only).
const int bufferSize = 128;
var from = Math.Max(0, previousCrossReferenceLocation - bufferSize / 2);
bytes.Seek(from);
var buffer = new byte[bufferSize];
bytes.Read(buffer);
var content = OtherEncodings.BytesAsLatin1String(buffer);
var xrefAt = content.IndexOf("xref", StringComparison.OrdinalIgnoreCase);
if (xrefAt >= 0)
{
previousCrossReferenceLocation = from + xrefAt;
missedAttempts++;
continue;
}
}
previousCrossReferenceLocation = storedCurrentTokenScannerPosition;
missedAttempts++;