[TIKA-1024] An MP3 with an UTF-16 ID3 tag containing only the BOM should produce empty string value for that tag

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.3
• Component/s: parser
• Assignee: Michael McCandless
• Reporter: Michael McCandless
• Created: 13/Nov/12 14:41
• Updated: 18/Nov/12 15:53
• Resolved: 18/Nov/12 15:53

Description

This seems to be a difference between JVMs: on IBM's JVM I incorrectly see the BOM as the value of the tag, while on Oracle's JVM I correctly get the empty string.

I'm not sure if this is a bug in IBM's JVM ... the javadocs are not totally clear how a UTF-16 string containing only the BOM should be decoded by new String(...) ... to fix this I think we should just detect this case and short-circuit empty string return.