[TIKA-2862] Make PDF Parser GraalVM native mode ready

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Not A Problem
• Affects Version/s: 1.20
• Component/s: parser
• Assignee: Sergey Beryozkin
• Reporter: Sergey Beryozkin
• Created: 02/May/19 17:03
• Updated: 22/May/19 12:54
• Resolved: 22/May/19 12:54

Description

PDF Parser is not Graal native mode ready yet, the following is reported when it is processed as part of Quarkus native mode build:

Error: Detected a FileDescriptor in the image heap. ....You can manually delay class initialization to image run time by using the option --delay-class-initialization-to-runtime=<class-name>. ...
Detailed message:
 Trace:     object org.apache.fontbox.ttf.BufferedRandomAccessFile
     object org.apache.fontbox.ttf.RAFDataStream
     object org.apache.fontbox.ttf.TrueTypeFont
     object org.apache.pdfbox.pdmodel.font.PDType1Font
     method org.apache.pdfbox.pdmodel.interactive.form.PDAcroForm.verifyOrCreateDefaults()
 Call path from entry point to org.apache.pdfbox.pdmodel.interactive.form.PDAcroForm.verifyOrCreateDefaults(): 
     at org.apache.pdfbox.pdmodel.interactive.form.PDAcroForm.verifyOrCreateDefaults(PDAcroForm.java:106)
     at org.apache.pdfbox.pdmodel.interactive.form.PDAcroForm.<init>(PDAcroForm.java:93)
     at org.apache.pdfbox.pdmodel.PDDocumentCatalog.getAcroForm(PDDocumentCatalog.java:108)
     at org.apache.tika.parser.pdf.PDFParser.handleXFAOnly(PDFParser.java:534)
     at org.apache.tika.parser.pdf.PDFParser.parse(PDFParser.java:164)

 

See also https://medium.com/graalvm/understanding-class-initialization-in-graalvm-native-image-generation-d765b7e4d6ed

Issue Links

• is blocked by: PDFBOX-4548 Update PDType1Font to make PDFBox GraalVM native mode ready