Remove the temporary Url list, and remove the disambiguities:
- An existing files named "file:..." or "inode:..." was handled as documentID/inode in the loop.
- An inode number matching an existing file name was interpreted as file name, even when the "-i" option was specified.
Also, parsing errors of inode/documentId numbers where ignored, stating
"No index information found" instead.
Add the currently processed Url to the error messages, to make it
more obvious which one the error refers to.