Re: PdfSmartCopy generating bad PDFs

> -----Original Message-----
> From: Gylfi Ingvason [mailto:Gylfi.Ingvason@...]
> Sent: Thursday, June 18, 2009 6:11 PM
> To: itextsharp-questions@...
> Subject: [itextsharp-questions] PdfSmartCopy generating bad PDFs
>
>
> The PDF file below, when modified and concatenated with a similar
> file, will generate bad PDF output using PdfSmartCopy. This can be
> replicated doing the
> following:
>
> 1. Cut and paste the PDF file below into a separate .pdf file. Use an
> editor that does not molest line terminators and make sure no line
> wrapping is taking place - all dictionary entries should be contained
> within a single line and not broken up.
>
> 2. After you have verified that the file opens and displays in Acrobat
> (if it asks to save changes when you close it, it means that line
> terminators are not correct), make another copy of the file, and
> change "(Blue Content)"
> to "(Red  Content)" and in the line above  change "0 0 1 rg"
> to "1 0 0 rg".
> Make sure to add an extra space after "Red" to keep the byte count the
> same.
>
> 3. Concatenate the two files using the following code or something
> similar:
>
> PdfReader reader = new PdfReader("File1.pdf"); Document document = new
> Document(reader.GetPageSizeWithRotation(1));
> PdfSmartCopy writer = new PdfSmartCopy(document, new
> FileStream("Concatenated.pdf", FileMode.Create, FileAccess.Write));
> document.Open();
>
> writer.AddPage(writer.GetImportedPage(reader, 1)); reader.Close();
>
> reader = new PdfReader("File2.pdf");
> writer.AddPage(writer.GetImportedPage(reader, 1)); reader.Close();
> writer.Close();
>
> 5. The "Concatenated.pdf" will contain two pages that both have the
> "Blue Content" displayed. Upon examining the generated file, you will
> find that both pages reference the same content and the "Red  Content"
> is missing. A
> slightly different composition of the input file will generate output
> where an XObject Form will reference itself which causes the Foxit
> reader to immediately kill itself and cause Acrobat to complain about
> a damaged file.
>
> The pdf file below was hand created whittling down concatenated
> customer files where single complex invoices and checks were being
> prepared for printing. Personally, I find the re-use of named
> resources like this somewhat questionable, but the PDF specification
> seems to indicate that this is valid since the name re-use is within
> the scope of the Form only.
>
> At any rate, I would like the group to know that using PdfSmartCopy
> can produce erroneous, and under some circumstances, deadly (for the
> Reader) results. I don't trust myself to debug the source for this,
> but perhaps Paulo or somebody else with more knowledge of the code can
> take a closer look and patch this.
>
> Thanks - Gylfi
>
> %PDF-1.4
> %âãÏÓ
> 1 0 obj<</Pages 2 0 R/Type/Catalog>>
> endobj
> 2 0 obj<</Kids[3 0 R]/Count 1/Type/Pages>> endobj
> 3 0 obj<</Resources<</XObject<</Xf1 5 0 R>>/ProcSet [/PDF
> /Text]>>/Parent 2 0 R/MediaBox[0 0 612 792]/Contents 4 0 R/Type/Page>>
> endobj
> 4 0 obj <</Length 26>>stream
> q 1 0 0 1 0 0 cm /Xf1 Do Q
> endstream
> endobj
> 5 0 obj <</BBox[0 0 612 792]/Matrix[1 0 0 1 0 0]/FormType
> 1/Subtype/Form/Length 26/Resources<</XObject<</Xf1 6 0
> R>>/ProcSet[/PDF/Text]>>/Type/XObject>>stream
> q 1 0 0 1 0 0 cm /Xf1 Do Q
> endstream
> endobj
> 6 0 obj <</BBox[0 0 612 792]/Matrix[1 0 0 1 0 0]/FormType
> 1/Subtype/Form/Length 26/Resources<</XObject<</Xf1 7 0
> R>>/ProcSet[/PDF/Text]>>/Type/XObject>>stream
> q 1 0 0 1 0 0 cm /Xf1 Do Q
> endstream
> endobj
> 7 0 obj <</BBox[0 0 612 792]/Matrix[1 0 0 1 0 0]/FormType
> 1/Subtype/Form/Length 26/Resources<</XObject<</Xf1 8 0
> R>>/ProcSet[/PDF/Text]>>/Type/XObject>>stream
> q 1 0 0 1 0 0 cm /Xf1 Do Q
> endstream
> endobj
> 8 0 obj <</BBox[0 0 612 792]/Matrix[1 0 0 1 0 0]/FormType
> 1/Subtype/Form/Length 26/Resources<</XObject<</Xf1 9 0
> R>>/ProcSet[/PDF/Text]>>/Type/XObject>>stream
> q 1 0 0 1 0 0 cm /Xf1 Do Q
> endstream
> endobj
> 9 0 obj <</BBox[0 0 612 792]/Matrix[1 0 0 1 0 0]/FormType
> 1/Subtype/Form/Length 67/Resources<</Font<</F1 10 0
> R>>/ProcSet[/PDF/Text]>>/Type/XObject>>stream
> q
> BT
> 1 0 0 1 10 700 Tm
> /F1 15 Tf
> 0 0 1 rg
> (Blue Content)Tj
> 0 g
> ET
> Q
> endstream
> endobj
> 10 0
> obj<</Subtype/Type1/BaseFont/Courier/Encoding/WinAnsiEncoding/
> Type/Font>>
> endobj
> xref
> 0 11
> 0000000000 65535 f
> 0000000015 00000 n
> 0000000059 00000 n
> 0000000109 00000 n
> 0000000245 00000 n
> 0000000318 00000 n
> 0000000521 00000 n
> 0000000724 00000 n
> 0000000927 00000 n
> 0000001130 00000 n
> 0000001371 00000 n
> trailer
> <</Root 1 0 R/Size 11>>
> startxref
> 1457
> %%EOF