GitLab

  - try introducing some invariants more precisely, to make the
    explanation of the code easier
  - address JH’s remark about order of writes

  - "offset" --> "slot"
  - "race" --> "competition" (there are no data races per se, as implied
    locations are atomic)
  - clarify thread interference
  - clarify explanation of try_enqueue

WIP paper on queue: remove distinction between \atloc and \naloc, as access modes are now ascribed to individual cells