[PDFBOX-5832] Error when writing a document with OutlineItems containing null SE objects

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.0.2 PDFBox, 4.0.0
• Fix Version/s: 3.0.3 PDFBox, 4.0.0
• Component/s: Writing
• Assignee: Andreas Lehmkühler
• Reporter: Arthur Renard
• Created: 30/May/24 10:08
• Updated: 10/Aug/24 07:18
• Resolved: 04/Jun/24 05:20

Description

Hello,

I'm reaching out to you because we encountered some errors when loading documents after updating PDFBox to v3.0.2.

I cloned the project in local env and tried with v3.0.3-SNAPSHOT but the same error appeared.

When trying to save my document using the PDDocument save() method, the following exception occurs: 
 

java.io.IOException: Error: Unknown type in object stream:COSObject{2240, 0}
	at org.apache.pdfbox.pdfwriter.compress.COSWriterObjectStream.writeObject(COSWriterObjectStream.java:238)	at org.apache.pdfbox.pdfwriter.compress.COSWriterObjectStream.writeCOSDictionary(COSWriterObjectStream.java:341)	at org.apache.pdfbox.pdfwriter.compress.COSWriterObjectStream.writeObject(COSWriterObjectStream.java:230)	at org.apache.pdfbox.pdfwriter.compress.COSWriterObjectStream.writeObjectsToStream(COSWriterObjectStream.java:119)	at org.apache.pdfbox.pdfwriter.COSWriter.doWriteBodyCompressed(COSWriter.java:499)	at org.apache.pdfbox.pdfwriter.COSWriter.visitFromDocument(COSWriter.java:1307) 

I can't share the document used for testing because it contains sensitive information, but after debugging a bit I found that it contains OutlineItems with null SE objects and that is apparently what's causing the error: 


 
The document was produced using Adobe Acrobat Pro 2020 20.5 30636

Unfortunately I don't have access to this software and I coulnd't recreate a similar document to reproduce the issue.
 
I found a user with a similar issue in your mailing lists : https://www.mail-archive.com/[e-mail]/msg13258.html
 
Let me know if you need more details regarding this problem.
 
Also, if you are able to create a test document that would reproduce the issue, would you please mind sharing it? It would be of great help.
Or if you have way to anonymize a document without altering its structure.
 
Many thanks in advance!