FetchFile fails on filenames containing characters with orthographic marks

As reported in the Slack channel:

Hi all! I'm trying to use FetchFile processor to read some files but it's not possible when directories includes special characters. I tried to solve that issue modifying the character set to UTF8. To do so I add "java.arg.8=-Dfile.encoding=UTF8" argument in the boostrap.conf file but it doesn't work properly.

The processor (and any other that interacts with the file system) should be fixed to properly handle filenames such as:

• información.txt
• líneas.csv
• tack_så_mycket.json
• słowo.xml