I'd like to suggest an import database feature for generic plaintexts.
Basically the user define a regular expression like "%filename% %size% %YYMMDD%" then the program parses the input file one-line at a time trying to match the regexp.
possible inputs:
- any csv file
- Windows' dir /S /B command output
- Linux' du --all --human-readable --time command output
- rhash -r command output
Scheduled for 2.4 release.
Implemented in svn revision 657.