Usage:

1. Save the server.py script on the server machine, make it executable
2. Set the HOST parameter on line 14 of list_client_programs.py to the ip address of the server
3. Save the list_client_programs.py script on the client machine, make it executable
4. Run ./server.py on the server machine
5. Run ./list_client_programs.py on the client machine
6. Running program data will be saved to the same repository as the script

Notes / Improvements:

• Only supports IPv4
• Should be capturing any errors opening/closing files/sockets
• Should consider rotation/cleanup of output files
• Size of data sent / read - is 1024 optimal?
• Clearer usage information and addition of --help output
• Messy to use temporary file for processing data on client side
• Also slow, would be better to use in-memory buffer
• Also the temp file never gets cleaned up
• No failure tolerance, if the script breaks, no more data until we run list_client_programs and server again
• Would be better to leave server open and run client process as a cronjob
• Needs linting
• Should abstract repeated code - (e.g. logic to create the unique filenames)
• Not much logging output for user
• No requirements.txt or package structure (though no requirements beyond python3)
• Question mark over portability to other operating systems
• Re the "EOF" marker, server.py should deal with normal case first, exceptional case after
• What if a process has the string "EOF" in its name?
• 5 seconds is a magic value, should be at least a constant, better a command line option
• Server IP address is a magic value, should be a command line option

Tests to consider:

• Test that all of mock ps output data turns up in output file as expected (i.e. output file is complete)
• Test output file contains no duplicate information
• Test output file contains no information that is not in mock ps output data (i.e. output file is accurate)
• Test what would happen if process interrupted on either machine (currently would fail)