document that ignoring a sender's "other" filesystem does not affect the
operation of --delete-excluded at the matching receiver's patch (/foo/bar).
I ran into this under a pile of scripts where the backups first backed up /foo,
then /foo/bar, then /foo/baz, and the underlying volume's CoW snapshots were
quickly gaining size because /foo was listed in the excludes (first deleted,
The current documentation isn't wrong - just being explicit would help
eliminate this as a gotcha.