This updates and obsoletes the series at [1]. The backend that
constructs queries is unmodified from that series, but the parser now
allows the use of count modifier in several more places. Basically
there is not much more code to maintain to do it for any field based
on terms (notably not date), but the utility is a bit unclear in some
cases.

In (probably) decreasing order of compelling use case

1) (and (from bob) (thread (count 1))) # find messages from bob nobody replied to yet.

2) (and (subject init-systems) (thread (count 200 *))) # find me a mega thread on some topic

3) (and (to bremner) (from (count (2 *))))  # find people that sent me at least 2 messages.

4) (tag (count 1)) # find tags used only once 
   
5) (path (count 1)) # find messages alone in a directory

6) (subject (count 1)) # find words used only once in subjects

[1]: id:20230213122631.2088558-1-david@tethera.net


_______________________________________________
notmuch mailing list -- notmuch@notmuchmail.org
To unsubscribe send an email to notmuch-leave@notmuchmail.org