[IMAGING-262] TIFF Parser ignores XMP Metadata if TagType is UNDEFINED

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.0-alpha1
• Component/s: Format: TIFF
• Labels: xmp
• Flags: Patch
• Assignee: Unassigned
• Reporter: Matthias Fratz
• Created: 12/Aug/20 14:17
• Updated: 12/Aug/20 14:17

Description

commons-imaging only extracts XMP metadata from TIFFs when the IFD's type field is BYTE. According to the Adobe spec (https://www.adobe.com/devnet/xmp.html, part 3 "Storage in Files" section 1.1.6 "TIFF", table 13) it can be either UNDEFINED or BYTE, and some of the files we are processing have it set to UNDEFINED lately. I cannot share the offending files unfortunately because they're somewhat confidential.

The attached patch changes commons-imaging to write BYTES, but accept either BYTES or UNDEFINED when reading. It passes all existing unit tests and we've had it in production for a while without problems. The diff is against 1.0-alpha1, but applies to master with line offsets.

Please let me know if you need further information.