PDF generated by SaveToPdfAsync fails validation by mustang

[ThomasSpeedy]

Hello, I generated a zugferd pdf using an existing pdf file by using SaveToPdfAsync(). The generated pdf file can be opened by adobe acrobate and looks like the original pdf but with added xml and is reported to be PDF/A-3A. Then I used mustang's validation tool to validate the generated pdf file. Unfortunately mustang reports some errors:

... ValidationResult [flavour=3a, totalAssertions=17146, assertions=[TestAssertion [ruleId=RuleId [specification=ISO 19005-3:2012, clause=6.1.7.1, testNumber=1], status=failed, message=The value of the Length key specified in the stream dictionary shall match the number of bytes in the file following the LINE FEED (0Ah) character after the stream keyword and preceding the EOL marker before the endstream keyword, location=Location [level=CosDocument, context=root/indirectObjects[18](4 0)/directObject[0]], locationContext=null, errorMessage=Actual length of the stream (8093 byte(s)) does not match the value of the Length key in the Stream dictionary (8094 byte(s))], TestAssertion [ruleId=RuleId [specification=ISO 19005-3:2012, clause=6.2.11.4.1, testNumber=1], status=failed, message=The font programs for all fonts used for rendering within a conforming file shall be embedded within that file, as defined in ISO 32000-1:2008, 9.9, location=Location [level=CosDocument, context=root/document[0]/pages[0](16 0 obj PDPage)/contentStream[0](15 0 obj PDContentStream)/operators[22]/font[0](CenturyGothic)], locationContext=null, errorMessage=The font program is not embedded], TestAssertion [ruleId=RuleId [specification=ISO 19005-3:2012, clause=6.2.11.4.1, testNumber=1], status=failed, message=The font programs for all fonts used for rendering within a conforming file shall be embedded within that file, as defined in ISO 32000-1:2008, 9.9, location=Location [level=CosDocument, context=root/document[0]/pages[0](16 0 obj PDPage)/contentStream[0](15 0 obj PDContentStream)/operators[841]/font[0](CenturyGothic,Bold)], locationContext=null, errorMessage=The font program is not embedded]], isCompliant=false] ...

How can I generate a valid pdf?

Thanks in advance for any advice!
Thomas

[stephanstapel]

Are you using the Alpha version from nuget or the current code from github?

[stephanstapel]

You can run the BasicSaveExampleFile() test and use the resulting file.
I ran the result against VeraPDF already:

but am unsure if font embedding really works like it should. A code review of _LoadFonts() would be great:

ZUGFeRD-csharp/ZUGFeRD.PDF/InvoiceDescriptorPdfSaver.cs

Line 303 in cb46e95

private static List<XFont> _LoadFonts(PdfDocument pdfDocument)

I checked embedding with PDF-XChange Editor:

but is this sufficient?

[ThomasSpeedy]

You could use Acrobat Reader to check if the fonts are embedded correctly. Open generated pdf with it, see Menu -> Dokumenteneigenschaften -> Schriften.
At least my input.pdf without already embedded fonts didn't result in a proper pdf with embedded fonts. Did you use a pdf as input without embedded fonts? At least I couldn't spot one in test folder.

[stephanstapel]

which fonts are used in that pdf?
as far as I understand, the postscript fonts are returned by PDFsharp and need to be translated to the truetype equivalents in order to embed them:

ZUGFeRD-csharp/ZUGFeRD.PDF/InvoiceDescriptorPdfSaver.cs

Line 311 in cb46e95

var postScriptToTrueTypeMap = new Dictionary<string, string>(StringComparer.OrdinalIgnoreCase)

[ThomasSpeedy]

I tried 2 different pdfs, one with font Arial, the other with Century Gothic. Both failed to embed fonts in the output.

[stephanstapel]

can you step through this function:

ZUGFeRD-csharp/ZUGFeRD.PDF/InvoiceDescriptorPdfSaver.cs

Line 305 in cb46e95

List<XFont> retval = new List<XFont>();

and finally check what is returned here:

ZUGFeRD-csharp/ZUGFeRD.PDF/InvoiceDescriptorPdfSaver.cs

Line 368 in cb46e95

return retval;

does it include the desired fonts?

[ThomasSpeedy]

I fixed some issues of your approach and was able to embed the fonts. But somehow the not embedded fonts are still used in the generated pdf:

If you want me to push my improvements to your branch, please grant me access.

[stephanstapel]

@ThomasSpeedy : sounds great. I set up a ruleset, can you try to issue a PR? Otherwise, we can also use the simple method: feel free to just zip the code and upload it here.

[ThomasSpeedy]

Unfortunately:

[stephanstapel]

damn. Could you share your fix as a zip file? I know this is really old school but I guess this will be faster than me figuring out how to adjust github permissions

[ThomasSpeedy]

Because my improvements are still resulting in a pdf that is not considered valid by mustang, it could be worth doing it with github if you want me to further participate

[stephanstapel]

@ThomasSpeedy : I added to the existing test case and accidently dropped the changes into the master but imho that doesn't make the implementation worse.