[PATCH 6/6] man: document notmuch count --output=files

[David Bremner]

Other than the bikeshed about starting from zero instead of one, and not
being completely happy with the option name --duplicate (but not being
inspired to suggest a better one), this series looks OK to me.

It seems like it would be quite useful to query based on the number of
duplicates, but that is a much more ambitious task, I think.

d