PDFSharp fails to open a document, created by PDFSharp when document contains a JPEG image made from a MemoryStream

[Volkmire]

Description

PDFSharp fails to open a document, created by PDFSharp when document contains a JPEG image made from a MemoryStream. I've tested bmp and png files and the bug seems to not affect them, but don't quote me on that.

I've reproduced this bug both on Linux and Windows.
The bug is not reproduced with FileStream and w/e stream ASP.NET Core wraps Request's body with.

The document can be opened by means of other PDF viewers (Firefox, Adobe Acrobat).

Steps to reproduce

var path = @""; // your path to jepg
using var fs = File.OpenRead(path);
using var ms = new MemoryStream();
fs.CopyTo(ms);
ms.Seek(0, SeekOrigin.Begin); // not required
using var image = XImage.FromStream(ms); // if you pass fs stream directly without copying to Memory - no error raised on open

using var document = new PdfDocument();
var page = document.AddPage();

double width = Math.Min(595, image.PixelHeight);
double height = width * image.PixelHeight / image.PixelWidth;
page.Width = XUnit.FromPoint(width);
page.Height = XUnit.FromPoint(height);

using var gfx = XGraphics.FromPdfPage(page);
gfx.DrawImage(image, 0, 0, width, height);

// Doesn't matter if you save to file or stream. For this example I'm saving to file
string outPath = $"broken.pdf";
await document.SaveAsync(outPath);

using var savedDocumentFs = File.OpenRead(outPath);
using var pdf = PdfReader.Open(savedDocumentFs); // this line throws if JPEG in memory stream was passed to original document

Expected result

PDF file opens

Actual result

Exception ObjectNotAvailableException is being thrown with message Cannot retrieve stream length. PDFsharp cannot read this PDF file. If you think your file is a valid PDF file please send it to us so that we can fix this bug in the PDF parser.

Stack trace:

   at PdfSharp.Pdf.IO.Lexer.<>c.<DetermineStreamLength>b__17_0() in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\Lexer.cs:line 980
   at PdfSharp.Internal.SuppressExceptions.HandleError(SuppressExceptions suppressExceptions, Action throwException) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Internal\SuppressExceptions.cs:line 25
   at PdfSharp.Pdf.IO.Lexer.DetermineStreamLength(Int64 start, Int32 searchLength, SuppressExceptions suppressObjectOrderExceptions) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\Lexer.cs:line 980
   at PdfSharp.Pdf.IO.Parser.ReadDictionaryStream(PdfDictionary dict, SuppressExceptions suppressObjectOrderExceptions) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\Parser.cs:line 394
   at PdfSharp.Pdf.IO.Parser.ReadObjectInternal(PdfObject pdfObject, PdfObjectID objectID, Boolean includeReferences, Boolean fromObjectStream, SuppressExceptions suppressObjectOrderExceptions) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\Parser.cs:line 329
   at PdfSharp.Pdf.IO.Parser.ReadIndirectObject(PdfReference pdfReference, SuppressExceptions suppressObjectOrderExceptions, Boolean withoutDecrypting) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\Parser.cs:line 946
   at PdfSharp.Pdf.IO.Parser.ReadAllIndirectObjects() in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\Parser.cs:line 1062
   at PdfSharp.Pdf.IO.PdfReader.OpenFromStream(Stream stream, String password, PdfDocumentOpenMode openMode, PdfPasswordProvider passwordProvider, PdfReaderOptions options) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\PdfReader.cs:line 392
   at PdfSharp.Pdf.IO.PdfReader.Open(Stream stream, String password, PdfDocumentOpenMode openMode, PdfPasswordProvider passwordProvider, PdfReaderOptions options) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\PdfReader.cs:line 238
   at PdfSharp.Pdf.IO.PdfReader.Open(Stream stream, PdfDocumentOpenMode openMode, PdfReaderOptions options) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\PdfReader.cs:line 215
   at PdfSharp.Pdf.IO.PdfReader.Open(Stream stream) in D:\THHO\Repos\PDFsharp\src\foundation\src\PDFsharp\src\PdfSharp\Pdf.IO\PdfReader.cs:line 209

(!works) jesterrace.jpeg
(works) jesterrace.png
(works) jesterrace.bmp

[Phil-oops]

Ich weiß nicht ob es dir bewusst ist, aber auf dem oberen Bild ist ein Effekt drauf. Wenn du es heruntergeladen hast, halte mal den Daumen auf der "Burg?" gedrückt, diese wird dann mit einem 3D Effekt hervorgehobenen. Ich kann mir sehr gut vorstellen, dass es damit zu tun hat. Wenn du das Bild dringen benötigst, dann ist die vermutlich schnellste Lösung es ohne Effekte zu speichern oder ein Screenshot zu machen. 🤷‍♂️

[Volkmire]

Ich weiß nicht ob es dir bewusst ist, aber auf dem oberen Bild ist ein Effekt drauf. Wenn du es heruntergeladen hast, halte mal den Daumen auf der "Burg?" gedrückt, diese wird dann mit einem 3D Effekt hervorgehobenen. Ich kann mir sehr gut vorstellen, dass es damit zu tun hat. Wenn du das Bild dringen benötigst, dann ist die vermutlich schnellste Lösung es ohne Effekte zu speichern oder ein Screenshot zu machen. 🤷‍♂️

I'm not aware that jpegs can do that, it's probably your gallery app or something. The image is just a reference image, you can replace it with another and still get the same effect. For example, I've reproduced the bug with the image you've attach to your reply. :D

[ThomasHoevel]

There is a bug in PDFsharp that will be fixed with a future release.

In your current code, replace using var ms = new MemoryStream(); with using var ms = new MemoryStream((int)fs.Length); and it will work with the current version of PDFsharp.

We are sorry for the inconvenience.

[Volkmire]

Nice. Will try to adapt the fix to our realities. Thanks @ThomasHoevel.