Adobe PDF SDK (Not Acrobat SDK) - Problems with converting PDFs to PDF/X-1a

[Piyush Popli]

Hi

I understand that this a forum for Acrobat SDK. However, I'm still posting this question here because -

1. I couldn't find a forum specifically for Adobe PDF lib

2. The PDF lib is a subset of the Acrobat SDK (minus the AV layer) so I'm guessing this is "relevant" here too

(Please let me know if I'm mistaken)

That having been said here it is..

We have licensed Adobe PDF SDK that we're using within an in-house service (in a Linux container) to convert PDFs to PDF/X-1a (for reliable printing). This works perfectly most of the time. However, some of the PDFs fail conversion. We get a "[Error 269221888] PDFX Font Error Occured". We're using code which is very similar to that in PDFProcessor.cpp (included in the SDK samples) ie: using the PDFProcessorConvertAndSaveToPDFX. This function is not listed in the docs so I'm guessing that this is an "undocumented  plugin" that comes with the SDK.

We know that the failing PDFs have some "unembedded fonts" and so installed the missing fonts but the conversion still fails. The puzzling this is that the the PDFProcessorConvertAndSaveToPDFA function works on the same file. My understanding is that a PDF-A also has the fonts embedded so I'm guessing that the "font installation" worked. I'm unable to understand why the PDFProcessorConvertAndSaveToPDFX still keeps failing. Another point is that the PDFProcessorConvertAndSaveToPDFX works on the output from PDFProcessorConvertAndSaveToPDFA (ie: the PDF-A file that has the fonts embedded).

All the above means that we have to try the PDFX conversion and if it fails (because of "font issues") convert to PDFA and then convert to PDFX. We'd like to figure out a way to get the PDFX coversion working with the new installed fonts. With PDFProcessor* functions neither being documented nor open sourced, I'm not sure who/where to look to/for a solution.

Any suggestions?

[Test Screen Name]

Hmm, I use Adobe PDF Library and had never observed this function. Since the message is uninformative and specific to the function, I suggest you take this up with PDF Library support. You will have a support contract, presumably with DataLogics, and they have access to the source to look up what this numbered message actually means. They do more than distribute Adobe's code; they also fix bugs.

[Piyush Popli]

We DO have a support license with DataLogics but I though of asking in this forum too thinking someone might have faced a similar issue. (I didn't know that they had access to the lib source and fixed bugs. Thanks!)
That having been said, I think we figured out the issue.

The PDF lib gets initiated with a PDFDataRec struct. It has a member called "dirList". You can use this member to pass a list of directories to search for fonts/charmaps etc. On Linux, this doesn't seem to work as expected and the PDFProcessorConvertAndSaveToPDFX tends to JUST rely on the system installed fonts instead. I installed the fonts in the system dirs and did a fc-cache -r and VOILA it works!

I'll let Datalogics know.

Hope this helps someone else.

Thanks!