[SOLR-2460] Some European characters cannot be parsed correctly for some PDFs

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 1.4.1, 3.1
• Fix Version/s: 3.1.1, 3.5
• Component/s: contrib - Solr Cell (Tika extraction)

Environment

Tika, PDFBox

• Assignee: Steven Rowe
• Reporter: Erlend Garåsen
• Created: 07/Apr/11 10:50
• Updated: 27/Nov/11 12:36
• Resolved: 26/Oct/11 00:52

Description

The Norwegian characters (æ, ø and å) in the following PDF document will not display correctly after Solr has indexed it, using Solr Cell:
http://ridder.uio.no/dokument.pdf

If I manually change the version of PDFBox (one of Tika's dependencies) to 1.4.0, the document will parse correctly.

I suggest that the next release of Solr ships with version 0.9 of Tika which also has updated its PDFBox dependencies to 1.4.0