[TIKA-1157] Mp3 file won't convert => 100% CPU

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.4
• Fix Version/s: 1.5
• Component/s: general

Environment

AMD64 Linux (kernel 3.8.0)
OpenJDK 1.7.0_25 64 bits

• Assignee: Unassigned
• Reporter: Damien Dykman
• Created: 07/Aug/13 19:00
• Updated: 08/Aug/13 16:28
• Resolved: 08/Aug/13 16:28

Description

With some Mp3 files, Tika does not return but will use 100% CPU. Memory consumption remains constant.

More details about the file (returned by /usr/bin/file command):
MPEG ADTS, layer III, v2, 32 kbps, 22.05 kHz, Monaural

The problem can be reproduce by simply running:

java -jar tika-app-1.4.jar https://dl.dropboxusercontent.com/u/4508166/tika/taliban.mp3

DISCLAIMER: this mp3 file is a real life example extracted from the Enron email corpus (public domain?). The humoristic content or the name of this file are not meant to propagate any political views.