PDF/A-1b bug with local characters

[miha-blurp]

Expected Behavior

The PDF/A document should be generated successfully with embedded font and Unicode text (šššččč) and validate as PDF/A-1b compliant.

Actual Behavior

When I add letters ččč the pdf fails to pass https://www.pdf-online.com/osa/validate.aspx (I also tried other validators ...) with the issue reported:

Validating file "minimal_not_working.pdf" for conformance level pdfa-1aThe key CIDSet is required but missing.The document does not conform to the requested standard.The document contains fonts without embedded font programs or encoding information (CMAPs).The document does not conform to the PDF/A-1a standard.Done.

If I remove "ččč" it passes no problem.

Steps to Reproduce the Behavior

Sub minimal_not_working()

    Dim Document As New PdfDocument()

    Dim font As XFont = New XFont("Arial Unicode MS", 12, XFontStyleEx.Regular, New XPdfFontOptions(PdfFontEmbedding.EmbedCompleteFontFile))

    Document.Info.Title = "PDF/A Minimal Test"
    Document.Info.Author = "Tester"
    Document.Info.Subject = "PDF/A-1b Compliance"
    Document.Info.Creator = "PDFSharp VB.NET"
    Document.Version = 14
    Document.SetPdfA()

    Dim UAManager As UAManager = UAManager.ForDocument(Document)
    Dim PdfPage As PdfPage = Document.AddPage()
    Dim gfx As XGraphics = XGraphics.FromPdfPage(PdfPage)

    Dim sb As StructureBuilder = UAManager.StructureBuilder
    sb.BeginElement(PdfGroupingElementTag.Article)
    sb.BeginElement(PdfBlockLevelElementTag.Paragraph)

    gfx.DrawString("šššččč", font, New XSolidBrush(XColor.FromArgb(0, 0, 0)), New XRect(50, 100, 500, 30), XStringFormats.TopLeft)
    sb.End()
    sb.End()
    Document.Save("minimal_not_working.pdf")
End Sub

The following variation works correctly:

gfx.DrawString("ššš", font, New XSolidBrush(XColor.FromArgb(0, 0, 0)), New XRect(50, 100, 500, 30), XStringFormats.TopLeft)

Environment

• PDFsharp version: 6.2.0 (PDFsharp-GDI)
• .NET version: .NET Framework 4.7.2
• Font: Arial Unicode MS

Additional Context

I tried many different fonts and no luck. Thank You.

minimal_working.pdf
minimal_not_working.pdf

[miha-blurp]

Here is validator from Vera validator reporting the issue, I am adding this, since it has more details:

Specification: ISO 19005-1:2005, Clause: 6.3.5, Test number: 3
For all CIDFont subsets referenced within a conforming file, the font descriptor dictionary shall include a CIDSet stream identifying which CIDs are present in the embedded CIDFont file, as described in PDF Reference Table 5.20 Failed
1 occurrences Hide
PDCIDFont
fontName.search(/[A-Z]{6}+/) != 0 || (containsCIDSet == true && cidSetListsAllGlyphs == true)
root/document[0]/pages[0](9 0 obj PDPage)/contentStream[0](10 0 obj PDContentStream)/operators[10]/font[0](DQMZHK+Arial Unicode MS)/DescendantFonts[0](DQMZHK+Arial Unicode MS)

[ThomasHoevel]

It turned out that the way how PDFsharp add Unicode font subsets to the PDF file is not compatible with PDF/A-1A or PDF/A-1B.
The fonts are correct when you validate against PDF/A-2x or PDF/A-3x.

So if PDF/A-2x or PDF/A-3x is an option for you:
After creating the PDF file, you can search for "pdfaid:part1</pdfaid:part>" and replace it with "pdfaid:part2</pdfaid:part>" or "pdfaid:part3</pdfaid:part>".

Or download the PDFsharp source and make the change in the source before you compile your own DLLs.
I'm sorry for the inconvenience.

Files are valid PDF/A-1A files as long as only ANSI characters are used. That's why it works with "š" or "äöü" and other characters.
The "č" is not ANSI.