concurrency/locking about ocfl-py HOT 1 OPEN

Certainly nothing in the code yet. I'd be interested in talking about this as a possibility. Perhaps a connector to a lock service that could have some simple default implementation (maybe using unix directory creation which would obviously only work for clients using the same lock filesystem). I can imagine a first stab for be object-level locks/checks could be:

• write lock - get a lock for an object (based on storage root plus object id, or full path/URL if object not in a storage root)
• update procedure - think about changes to object and version writes to account for locking per https://ocfl.io/1.0/implementation-notes/#an-example-approach-to-updating-ocfl-object-versions
• read checks:
  • blocking writes - check/get same lock as for write
  • non-blocking write: assumes good update procedure, read root inventory, check version directories (extra version directory indicates write in progress which might then not be an error)

from ocfl-py.