PDFBox usage in native image

[GodierLuna]

Hello,

I am building a Quarkus app in which I need a text extractor to get the contents of a pdf document.

I have first tried to use apache pdfbox and then went with tika but always get the same two errors, no matter what I do /@RegisteForReflection/serialization...)

Caused by: java.lang.UnsatisfiedLinkError: No awt in java.library.path
Caused by: java.lang.NoClassDefFoundError: Could not initialize class org.apache.pdfbox.pdmodel.PDDocument

My infra is on quarkus 3.27

Thank youu

[holly-cummins]

To confirm, are you using the raw libraries, or the Quarkiverse extensions? There's both a Quarkiverse Tika extension and a PDFBox extension, which I assume you discovered.

What operating system are you using? AWT on MacOS is tricky (see discussion), but on Linux it works will if the AWT extension is being used. I don't know what the status is on Windows.

[GodierLuna]

No I'm not on MacOS, I'm coding on windows but then I deploy my app on some pods that run the native image

[holly-cummins]

No I'm not on MacOS, I'm coding on windows but then I deploy my app on some pods that run the native image

AWT on pods should work, as long as the AWT extension is present. Can you confirm it is, in the Dev UI?