fix(storage): reset size_written only after successful clear

[kevetka]

Problem:
In both clear_database and clear_database_full methods of ChunkStorage, the size_written counter was being reset to 0 before calling the fallible database.clear(). If clear() returns an Err, the method propagates the error to the caller, but size_written has already been zeroed. This leaves the storage in an inconsistent state: the database still contains all data, yet the byte counter indicates that nothing has been written.

Fix:
Reordered the operations so that size_written is updated only after all fallible operations have succeeded:

• clear_database: database.clear()? → size_written = 0 → Ok(())
• clear_database_full: database.clear()? → target_map.clear()? → size_written = 0 → Ok(())
  With this fix, if clear() fails, the state remains consistent: the database is untouched and size_written still reflects the actual amount of stored data.

[Vlm326]

Ok👌