Job outputs should only contain regular files
The job outputs browsable and downloadable by the user should only contain regular files, i.e. they must be
os.path.isfile() (because the UI and API do support subdirectories) and must not be
os.path.islink() (for security reason, see #269 (closed)).
This applies to:
- the list of files in JobDetail
- the generation of the .zip archive