Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

NPE caused by undefined object #177

Open
wants to merge 2 commits into
base: trunk
Choose a base branch
from
Open

NPE caused by undefined object #177

wants to merge 2 commits into from

Conversation

Trisia
Copy link

@Trisia Trisia commented Feb 1, 2024

example:

125 0 obj
<</Tabs/S/Group<</S/Transparency/Type/Group/CS/DeviceRGB>>/Contents[69 0 R 3646 0 R 70 0 R]/Type/Page/QITE_pageid<</UF 3619 0 R/P 5/D(AA\r)/F 3620 0 R/I 3621 0 R>>/Resources<</ExtGState<</Xi10 1 0 R/GS7 3640 0 R/GS8 3641 0 R>>/ProcSet[/PDF/Text/ImageB/ImageC/ImageI]/Font<</F7 3633 0 R/F8 3624 0 R/F9 3642 0 R/F1 3627 0 R/F2 3628 0 R/F3 3629 0 R/Xi11 2 0 R>>>>/Parent 82 0 R/StructParents 5/MediaBox[0 0 595.2 841.92]>>

see

QITE_pageid<</UF 3619 0 R/P 5/D(AA\r)/F 3620 0 R/I 3621 0 R>>

object 3621 0 R is not defined

When I use PDDocument.saveIncremental to save the document it causes the following error:

java.lang.NullPointerException
	at java.util.Hashtable.computeIfAbsent(Hashtable.java:1004)
	at org.apache.pdfbox.pdfwriter.COSWriter.getObjectKey(COSWriter.java:1089)
	at org.apache.pdfbox.pdfwriter.COSWriter.writeReference(COSWriter.java:1367)
	at org.apache.pdfbox.pdfwriter.COSWriter.visitFromDictionary(COSWriter.java:1207)
	at org.apache.pdfbox.pdfwriter.COSWriter.writeDictionary(COSWriter.java:1155)
	at org.apache.pdfbox.pdfwriter.COSWriter.visitFromDictionary(COSWriter.java:1202)
	at org.apache.pdfbox.cos.COSDictionary.accept(COSDictionary.java:1265)
	at org.apache.pdfbox.pdfwriter.COSWriter.doWriteObject(COSWriter.java:610)
	at org.apache.pdfbox.pdfwriter.COSWriter.doWriteObject(COSWriter.java:643)
	at org.apache.pdfbox.pdfwriter.COSWriter.doWriteObjects(COSWriter.java:540)
	at org.apache.pdfbox.pdfwriter.COSWriter.doWriteBody(COSWriter.java:450)
	at org.apache.pdfbox.pdfwriter.COSWriter.visitFromDocument(COSWriter.java:1299)
	at org.apache.pdfbox.cos.COSDocument.accept(COSDocument.java:413)
	at org.apache.pdfbox.pdfwriter.COSWriter.write(COSWriter.java:1568)
	at org.apache.pdfbox.pdmodel.PDDocument.saveIncremental(PDDocument.java:1078)

so we can skip those not define object, make it work.

@THausherr
Copy link
Contributor

Does this problem also occur with 2.0? Does it also occur in ordinary saving? Can you share the file?

@Trisia
Copy link
Author

Trisia commented Feb 2, 2024

The issue occurred with PDFBox 3.0.0, and I haven't tested it with 2.X.X. In fact, the PDF is encrypted and doesn't allow editing. I'm sorry, but for certain reasons, I can't provide the PDF file.

@THausherr
Copy link
Contributor

I had a look at 2.0, one of the changes isn't needed (2.0 doesn't use computeIfAbsent because it doesn't exist in the jdk, and also avoids using a null, so maybe this was introduced in refactoring), the other ones are. Your changes look useful so I'll commit them next week to give time for other opinions (COSWriter is a difficult class). I'll also add some logging.

@lehmi
Copy link
Contributor

lehmi commented Feb 3, 2024

We have to ensure that the resulting pdf isn't (more) corrupt that the origin one if indirect object references are omitted. To remove a reference from a COSArray shouldn't be a big problem (the object reference is simply missing) but to remove a reference from a COSDictionary without removing the key will lead to a corrupt pdf In such cases the key should be removed as well or the corrupt reference should be replaced by a COSNull object

@THausherr
Copy link
Contributor

In that case I'd really like to test this with a file. I forgot to mention yesterday that I tried to modify a PDF by "blanking" an object and then load and call saveIncremental() but no problem occured.

@Trisia
Copy link
Author

Trisia commented Feb 4, 2024

I thought the issue was caused by document encryption, but after a simple test, I discovered it was not the case. The document seems to have been generated using iText 5.5.8.

0003719771 00000 n 
0003759924 00000 n 
trailer
<</Info 5 0 R/Encrypt 7894 0 R/ID [<e3d8be7479575ba4b5622d7b671d2255><64acf33dec0b965797b470cbf8346962>]/Root 75 0 R/Size 7895>>
%iText-5.5.8
startxref
3760055
%%EOF

perhaps the original document had already lost the obj references.

maybe remove lost obj reference from a COSArray is good idea.

@THausherr
Copy link
Contributor

https://issues.apache.org/jira/browse/PDFBOX-5717 is the same problem

@THausherr
Copy link
Contributor

Please test whether your code works with the recent changes by @lehmi with the latest snapshot build:
https://repository.apache.org/content/groups/snapshots/org/apache/pdfbox/pdfbox-app/3.0.2-SNAPSHOT/

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants