[PDFBOX-5540] export:text creates jibberish / malformed output

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.0.16, 2.0.27, 3.0.0 PDFBox
• Fix Version/s: 2.0.28, 3.0.0 PDFBox
• Component/s: Text extraction
• Labels: regression
• Environment: Same on Windows, Linux and macOS
• Assignee: Tilman Hausherr
• Reporter: Alfons
• Created: 06/Nov/22 20:19
• Updated: 13/Apr/23 14:54
• Resolved: 12/Nov/22 03:53

Description

Using PDFBox as part of Tika and having issues with some PDFs outputting unreadable content. Copying text from Adobe / macOS Preview / Browsers works as expected.

I have also tried "re-encoding" the PDF by editing and saving it with Acrobat, thinking it could be an issue with their original PDF creator and using pdfbox with different encodings, but output mostly remained unchanged.

I attached the PDF and text it produces. Running it PDFBox via CLI as follows:

root % java -jar pdfbox-app-3.0.0-alpha3.jar export:text -i test.pdf          
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Invalid ToUnicode CMap in font 
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Using predefined identity CMap instead
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Invalid ToUnicode CMap in font 
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Using predefined identity CMap instead
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Invalid ToUnicode CMap in font 
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Using predefined identity CMap instead
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Invalid ToUnicode CMap in font 
Nov 06, 2022 9:12:47 PM org.apache.pdfbox.pdmodel.font.PDFont loadUnicodeCmap
WARNUNG: Using predefined identity CMap instead