How to grab the oldest unread message from `mail` in a script?

Question

In a script I am trying to finish writing now, I need to get the oldest unread message from the current user's mailbox (specifically, I need the subject line for this project).

After extracting the oldest message, I also need to forward it to a different user for archive purposes.

How can I extract just that information in bash or Python?

Would it be better to use procmail to extract the subject line into a file and then forward the message, requiring only file processing by the script? If so, how could I do that instead?

Answer 1

In conventional mbox systems, the first message in /var/mail/you is -- by definition -- unread and the oldest available.

formail -zxSubject: -1 -s </var/mail/you

To extract the message to a file (so you can conveniently forward just that message) and then extract the Subject: header, try something like

tmp=$(maketemp -t findoldestunread.XXXXXXXX) || exit 127
trap 'rm -f $tmp' EXIT   # remove temp file when done
trap 'exit 1' 1 2 3 5 15 # remove also in case of error / interruption
formail -1 -s </var/mail/you >"$tmp"
formail -zxSubject: <"$tmp"
sendmail -oi uncannyvictim@example.com <"$tmp"

In Maildir, the situation is somewhat more complex. New messages are temporarily stored in tmp and then moved to cur. The file name indicates flags; in so many words, look for a file name containing a comma near the end where S is not among the characters after the comma.

find path/to/Maildir/cur -maxdepth 1 -type f \
    -regex '.*,[^S,]*$' -printf '%C@ %p\n' |
sort | sed '1s/^[^ ]* //;q' | xargs formail -zxSubject:

The printf format specifier prints the timestamp first for easy sorting. We sort, grab the first line, trim the timestamp, and feed it to formail for extracting the Subject: header. (This will obviously break if you have funky file names with newlines in them or something; there are ways around that, but I'm lazy and practical here.)

Refactoring to just print the file name should be trivial in this case -- just remove the pipe to xargs and capture the output into a variable.

Some Maildir implementations will also have an index of some sort which might make this task a lot easier and faster than traversing the entire mailbox in the file system, but again, without more knowledge about which implementation you are using, this is just a speculative note at this point.

If your mailbox is not in either of these two formats, (you are weird and) you will need to update your question with more details.

Your reference to mail implies that you are probably using a traditional Berkeley mbox system, but there are many versions, some of which are somewhat esoteric.

Answer 2

The mail command can be used non-interactively to print out the list of emails in the inbox:

mail -H | head -1
>U     1  jenny@mybox.example.com Tue Apr 29 14 08:13 18/774    "Subject of mail"

Answer 3

There are at least three mail delivery formats. It would help to know which one you are using.

formail will read a mail file and can output headers from particular message via skip and total options. This assumes the messages have been deliver into a known file. If you want to archive all the messages this would be a suitable tool.

procmail is more suitable to archiving the messages as they arrive. It can easily be configured to send a copy to another address while delivering. There are standard methods to avoid mail locks, and skip certain types of messages.