search

openpreserve / jhove

File validation and characterisation.
http://jhove.openpreservation.org
Other
171 stars 79 forks source link

False positive PDF-HUL-94 java.lang.NullPointerException #809

Closed asciim0 closed 1 year ago

asciim0 commented 1 year ago

The attached file throws a PDF-HUL-94 NullPointerException but seems to be syntactially ok. Cross-checked the file with Preflight, qpdf and pdfcpu. Pdfcpu did report an error for the /Tabs entry in the PagesDictionary as /Tabs is not supported with the claimed PDF version 1.4 of the file ... however, since JHOVE should only check agianst the 1.7 ISO spec, and is not version-feature-exact for anything y 1.7 this shouldn't make a difference,

Due to no futher info on the validation error and due to no negative results with other tools I consider this currently a false positive.

algenreportNordsee_06082021.pdf

asciim0 commented 1 year ago

There's apparently no relation between the /Tabs entry and the NullPointer. If someone can let me know what check triggers the NullPointer exception, I can maybe pinpoint the error a little bit from there.

samalloing commented 1 year ago

Hi @asciim0

This problem is solved in the next relase of jHove 1.28. The problem is that Subject and Keyword are empty (null). That nullPointerException is handled now (see https://github.com/openpreserve/jhove/pull/782)

Sam

asciim0 commented 1 year ago

Awesome, thanks Sam! Will crosscheck with 1.28.

carlwilson commented 1 year ago

Hi both, I've created a new version of this PR #820, which I think is safer but achieves its aims in less code. @asciim0 I've added your test file to the regression corpus and can confirm that the fix means it's now Well-formed and valid. @samalloing could you take a look at #820 and let me know what you think. Once it's merged I'll put a link to a test build here.