Hello,

I found a bug in Tika Batch mode when it attempts to process any filename
near the linux file limit of 255 bytes. This causes the main program to
crash and fails to process any more files.

Note that the original filename is 255 bytes. Tika Batch appends .xml in
the output file by default.

java -jar tika-app-2.7.0.jar -i /home/ubuntu/longfilename-test -o
/home/ubuntu/longfilename-output/


BatchProcess:Caused by: java.nio.file.FileSystemException:
/home/ubuntu/longfilename-output/longfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelongfilenamelon.xml:
File name too long

....
BatchProcess:INFO  [pool-2-thread-1] 14:43:20,587
org.apache.tika.batch.BatchProcess MAIN_LOOP_EXCEPTION_NO_RESTART

Thanks!

Reply via email to