XFA form support

[mgnslndh]

It would be awesome if this library would support reading XFA forms. Is this something that you have thought about adding or need help with?

[InusualZ]

Any help is appreciated, but yeah, It has been discussed earlier #160 and #25

[mgnslndh]

I will take a stab at it. First step will be to retrieve the embedded XML for XPD/XFA. Beyond that I will probably need guidance in how the API should look like but I will look at PDFBox for reference. Do you have access to any related specifications?

[EliotJones]

Hi @mgnslndh currently there's no support as @InusualZ says. Unfortunately I think XFA is still a proprietary Adobe internal specification that the few applications supporting it have to license so I don't know if there's any accessible specification.

As far as the PDF specification goes you need section 8.6.7 here https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdf_reference_archive/pdf_reference_1-7.pdf

Additionally because it's deprecated in the PDF 2.0 specification I'm not sure what the future support will look like.

As a starter the form dictionary would be accessed something like:

var xfaDictionary = (document.Structure.Catalog
    .CatalogDictionary
    .Data["AcroForm"] as DictionaryToken)?.Data["XFA"] as DictionaryToken;

But with the internal PdfPig types you'd need to use the various utility methods that actually follow indirect references TryGet<T>(this DictionaryToken dictionary, NameToken name, IPdfTokenScanner tokenScanner, out T token) where T : IToken. A good starting reference is the AcroFormFactory:

https://github.com/UglyToad/PdfPig/blob/master/src/UglyToad.PdfPig/AcroForms/AcroFormFactory.cs#L50

[mgnslndh]

Thanks for the direction! I've started implementing simple support. Will push PR for you to look at when I've come a bit further.