Document expected behavior when using --replace-ttl/--replace-ttl-with when not using --force-updates

What's happening?
When using the --replace-ttl/--replace-ttl-with and restoring a backup containing documents which have expired on the target cluster, 'cbbackupmgr' will report that these documents have been skipped.

Is this the expected behavior?
Yes, because during a restore we use seqno/lww conflict resolution which will detect that the cluster already contains this document (or a later version) and will therefore return a 'AlreadyExits' error.

Steps to reproduce?
1) Create a bucket
2) Load 'n' documents which have a short ttl
3) Perform a backup of the bucket
4) Wait for the documents to expire
5a) Restore the backup to a new cluster (documents should be created)
5b) Restore the backup to a new bucket (documents should be created)
5c) Restore the backup to the same cluster/bucket (documents should be skipped)
5d) Restore the backup to the same cluster/bucket where the purge seqno has advanced to a point where the document tombstones have been purged (documents should be created)
5e) Restore the backup to the same cluster/bucket whilst using --force-updates (documents should be updated with the new ttl)

What's the issue?
This is a bit of an edge case and the behavior is not exactly immediately obvious so we should extend the documentation to state that this is the expected behavior (perhaps we should also explain why).