Upgrade to python3

[imerr]

PR to upgrade to python3 - companion to ArchiveTeam/archiveteam-megawarc-factory#20, fixes #3

Changes:

• Ports the megawarc script to python3
  • Initial conversion with 2to3
  • Manual fixes of mostly encoding errors, everything relevant should be handled as byte strings now
  • Removes ordereddict.py shim for old python versions (it is included in all modern ones)
  • megawarc-fix script has not been modified past 2to3 and been deprecated & moved to old/ - please do let me know if a port is required and I can get on that - I don't think our tooling uses it currently though.
• Adds a simple test suite based on real world data with result verification based on the current python2 version.
  • Test runs with code coverage. Currently 83% are covered (only leaving out some utilities and error handling)
  • Test data comparison notes:
    • The tar header format has changed and there is a few other fields I am ignoring in the meta json comparison (since e.g. the user/group ids are based on what is running the test)
    • gzip includes timestamps in headers and is not reproducible, so I am extracting and checksumming the extracted files instead
• fixes the previously broken convert command (bad unpack handling, likely happened when adding zst support - did not look into it further though)

Todo:

• Complete pass over megawarc to catch any string encoding issues
• Test functionality other than packing a megawarc
• Add test suite data after review
• Run tests on CI (?)