[[!img notmuch-logo.png alt="Notmuch logo" class="left"]] Notmuch is a great mail indexing tool that can also be used *in conjunction* with existing Mail User Agents (MUA) instead of replacing them. The advantage of such mixed solutions is that users can benefit from notmuch features (such as full-text search and thread reconstruction) without *having to* change MUA. A popular geek MUA is [the Mutt e-mail client](http://www.mutt.org); integrating notmuch with Mutt is not seamless, but fairly straightforward. There are two principal possibilities, each with its own ups and downs: # Using Notmuch with mutt-kz ## Install: git clone https://github.com/karelzak/mutt-kz.git cd mutt-kz ./configure && make && make install ## Configuration: # notmuch set nm_default_uri="notmuch:///PATH/TO/MY/Maildir" # path to the maildir set virtual_spoolfile=yes # enable virtual folders set sendmail="/PATH/TO/bin/nm_sendmail" # enables parsing of outgoing mail virtual-mailboxes \ "INBOX" "notmuch://?query=tag:INBOX and NOT tag:archive"\ "Unread" "notmuch://?query=tag:unread"\ "Starred" "notmuch://?query=tag:*"\ "Sent" "notmuch://?query=tag:sent" # sets up queries for virtual folders # notmuch bindings macro index \\ "" # looks up a hand mane query macro index A "+archive -unread -inbox\n" # tag as Archived macro index I "-inbox -unread\n" # removed from inbox macro index S "-inbox -unread +junk\n" # tag as Junk mail macro index + "+*\n" # tag as starred macro index - "-*\n" # tag as unstarred # sidebar set sidebar_width = 20 set sidebar_visible = yes # set to "no" to disable sidebar view at startup color sidebar_new yellow default # sidebar bindings bind index sidebar-prev # got to previous folder in sidebar bind index sidebar-next # got to next folder in sidebar bind index sidebar-open # open selected folder from sidebar # sidebar toggle macro index ,@) " set sidebar_visible=no; macro index ~ ,@( 'Toggle sidebar'" macro index ,@( " set sidebar_visible=yes; macro index ~ ,@) 'Toggle sidebar'" macro index ~ ,@( 'Toggle sidebar' # toggle the sidebar * Using: when you open `mutt` you get the INBOX opened. There you can crawl through your mails, and tag them as appropriate, either manually using the " ` " command, or using the bindings defined in configuration (such as A/I/S/+/-). * Mail tagging on sending You may have noticed in `mutt-kz`'s configuration that I set the `sendmail` variable of mutt to a `nm_sendmail` script. This is for tagging outgoing mail each time I send a mail. Here is the content of the script (which may be used directly in mutt's variable, I did not try). Source of nm_sendmail: #!/bin/bash tee >(notmuch-deliver -t sent -f Sent) | sendmail $* ## Mail filtering/tagging For mail tagging on arrival, I prefer to use a simple procmail delivery along with notmuch-delivery (which can be compiled in the `contrib/` directory of notmuch's sources). Of course, you could use formail or maildrop, instead of procmail, but it is flexible enough for my needs, and here is an example of configuration that can be useful: PATH=/bin:/usr/bin:/usr/local/bin # ensure each mail is unique :0 Wh: msgid.lock | formail -D 8192 msgid.cache # update addressbook with current mail :0 Wh | /usr/local/bin/notmuch_abook update NOINBOX="-r inbox" TAGS="" # manage dynamic tagging, using the ' + ' token in mail addresses # e.g.: user+TAG@fqdn.tld will generate the tag TAG :0:notmuch.lock * ^TO\/user\+[a-z0-9]+@fqdn\.tld * MATCH ?? ^user\+\/[a-z0-9]+ { TAGS="-t ${MATCH}" } # match all mails from mailing-lists, don't let them go to inbox, but tag them with ml :0:notmuch.lock * ^List-[Ii][dD]:.* { TAGS="${TAGS} -t ml -r inbox" } # tag all mails coming from mutt-kz mailing list :0:notmuch.lock * .*mutt-kz\.lists\.fedoraproject\.org.* | notmuch-deliver $TAGS -t mutt -t notmuch # tag all mails coming from notmuch mailing list :0:notmuch.lock * .*notmuch\.notmuchmail\.org.* | notmuch-deliver $TAGS -t notmuch # Mark all spams as junk mail :0:notmuch.lock * ^X-Spam-Status: Yes | notmuch-deliver -t junk :0:notmuch.lock * ^Subject: .*SPAM.* | notmuch-deliver -t junk ### All unmatched mails :0:notmuch.lock * .* | notmuch-deliver -v $TAGS ## Addressbook management and vim There are some emacs tips over [here](http://notmuchmail.org/emacstips/#index15h2) that explains how to configure emacs with an addressbook, but the few solutions were not fast enough for me. And I never could test the vala-based code. So I updated the notmuch_addresses code to use a cache to be able to make better matches and still be lightning fast. And finally, I needed to have a way to autocomplete my recipients' addresses automagically from within `vim`. So that's why I created the `notmuch_abook` code. You can either install it as a plugin in vim, using vundle: Vundle 'guyzmo/notmuch-abook' Or if you're not using vim (which is objectively the best editor ever), you may want to use it as standalone: % pip install notmuch_abook To configure the address book, append to `~/.notmuch-config` the following: % cat > ~/.notmuch-config << EOF [addressbook] path=/home/YOURUSERNAME/.notmuch-abook.db backend=sqlite3 EOF where YOURUSERNAME is your home directory. Then, you can sync the addressbook cache by running. It will create the database file specified in configuration, and go through all notmuch's indexed mails to get all addresses headers and cache them. It takes around 20 seconds on my 10000 mails index: % notmuch_abook create And you can query for addresses using the lookup command, which will match the beginning of each word in the name and address, as follows: % notmuch_abook lookup Foo Foobar Bar Bar Foobar Bar Bar Bar Bar Bar Bar If you're using vim, when you edit a mail having a filetype set to mail, you will be able to open a completion menu when you do at every address header. # Using Notmuch with bare Mutt, the old fashioned way * Using [notmuch-mutt](http://git.notmuchmail.org/git/notmuch/tree/HEAD:/contrib/notmuch-mutt), which will create a "virtual" maildir folder with search results whenever a search is made. The upside is that you can search all your folders simultanously; the downside is that your modifications in the results listing do not carry over, also having to switch folders comes with some more annoyances. A [how to use Notmuch with Mutt](http://upsilon.cc/~zack/blog/posts/2011/01/how_to_use_Notmuch_with_Mutt/) has been written by Stefano Zacchiroli. (Note by the howto author: I've linked the howto from this wiki rather than splicing it in, in order to avoid duplication of information. If you think it would be better to have it here, feel free to copy the text here. The howto is [available](http://git.upsilon.cc/cgi-bin/gitweb.cgi?p=zack-homepage.git;a=history;f=blog/posts/2011/01/how_to_use_Notmuch_with_Mutt.mdwn) in markdown syntax from the Git repository of my homepage.) * Using a simple macro that will emulate the "limit" mutt functionality using notmuch. See the [alternative notmuch integration](http://log.or.cz/?p=228) blog post for instructions and details, or simply put these two macros to your muttrc: # 'L' performs a notmuch query, showing only the results macro index L "unset wait_keyread -p 'notmuch query: ' x; echo \$x >~/.cache/mutt_terms~i \"\`notmuch search --output=messages \$(cat ~/.cache/mutt_terms) | head -n 600 | perl -le '@a=<>;chomp@a;s/\^id:// for@a;$,=\"|\";print@a'\`\"" "show only messages matching a notmuch pattern" # 'a' shows all messages again (supersedes default binding) macro index a "all\n" "show all messages (undo limit)" The upside (if you are used to working in the context of a single folder) is that this really does use the limit functionality of mutt and you are still in your original folder. The downside is that this approach has scaling problems and works well only for reasonably specific queries.