[TIKA-544] AutoDetectParser ignores charset in Content-Type metadata

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Invalid
• Affects Version/s: 0.7
• Fix Version/s: 0.8
• Assignee: Kenneth William Krugler
• Reporter: Kenneth William Krugler
• Created: 05/Nov/10 20:14
• Updated: 05/Nov/10 21:30
• Resolved: 05/Nov/10 21:30

Description

AutoDetectParser.parse() does this:

MediaType type = detector.detect(stream, metadata);
metadata.set(Metadata.CONTENT_TYPE, type.toString());

So the charset in the Metadata.CONTENT_TYPE is blown away. On entry to the parser, it should pick this apart and set the Metadata.CONTENT_ENCODING if it's null, and the content-type has charset info.