[TIKA-3783] Filename detection misses when a # and several . are in a filename

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 2.4.1
• Component/s: detector

Environment

java 8
TIKA-1928

• Assignee: Unassigned
• Reporter: Alexander
• Created: 02/Jun/22 06:05
• Updated: 02/Jun/22 12:10
• Resolved: 02/Jun/22 09:56

Description

NameDetector.detect() strip any fragments after #, but not only after extension, when filename contains dots and hashes.

Example: 'ABC#192.168.0.1#2.xxx' will be stripped to 'ABC#192.168.0.1'.

There are filenames that contains dots not only in extension.

Should change line:

int dot = name.indexOf('.')

https://github.com/apache/tika/blob/a52e4d153e950077f7fdedadcc5d75604fe2563d/tika-core/src/main/java/org/apache/tika/detect/NameDetector.java#L119

to:

int dot = name.lastIndexOf('.');