Read PDF Node Error after updating to 0.114.0

Hi,

This looks like a strange issue, but I’m reporting this here in case if anyone else is facing a similar issue. I recently updated the n8n version to 0.114.0 and one of my workflows which was reading a PDF started failing with “bad XRef entry” error. Note that it happens only with a few PDFs and not all. The moment I reverted back to the older version (v0.110.0), this started working fine.

The problem may not really be with n8n per se, but one of the libs used by Read PDF node. Was there anything changed recently with the Read PDF node?

Regards,
Santhosh

Very strange. No there have been no changes to the node for a very long time. Actually, really want to totally replace the node in the future as it is really not that helpful right now. I only other thing that came to mind is that the underlying library pdf-parse changed but also the have not been any changes for 2 years.

So can currently sadly not explain why it would break for some of your PDFs right now. Sorry.

Both 0.114.0 and 0.110.0 are on pdf-parse 1.1.1.

If you can share a sample PDF that triggers the error, I can look into it.