🚀 go-pugleaf

From: "Barton E. Schaefer" <island!ucbcad!cse.ogc.edu!schaefer>

On Jun 21,  4:03pm, Larry W. Virden wrote:
} Subject: strange mush problem
} 
} Mush running on VAX ultrix 2.3 - if you need my .mushrc I can send it.
} 
} When I get an address wrong, my dead.letter file ends up with TWO copies
} of the failed article.  Here is what I did and after that what was in
} dead.letter:
}
} UNIXEX (1051) $ mush whodidit
} To: whodidit
} Subject: this is a test
}  
} Lets see what we get in dead.letter when this happens
} send, continue editing, discard [s,c,d]? s
} Sending letter ... whodidit... User unknown
} Saving message in /v/lwv27/dead.letter
} /v/lwv27/dead.letter... Sent
} Saved unfinished letter in ~/dead.letter.
} sent.

[ dead.letter with two copies of the message omitted ]

That's not a bug, it's a feature. :-)

In spite of my :-) I'm serious.  One copy is being saved by your MTA
(presumably smail or sendmail), and the other copy is being saved by
mush.  Mush detects that the MTA exited with a failure code and saves a
dead.letter to prevent you from losing it; it has no idea that the MTA
has already saved the same letter.

You can see this by changing mush's dead-letter file with "set dead":

    set dead = ~/LETTER.DEAD

and then sending another message to a bogus address.  You should see

    Sending letter ... whodidit... User unknown
    Saving message in /v/lwv27/dead.letter
    /v/lwv27/dead.letter... Sent
    Saved unfinished letter in ~/LETTER.DEAD.
    sent.

Mush saves the dead.letter in case the MTA gets completely blown away
(root does a kill -9 or something :-).  This also means you will get
a dead.letter from mush even when the MTA does not write one -- for
example, sendmail normally does not write dead.letter if only one address
in a long list turns out to be wrong, but it will exit with an error
status that mush can detect.

-- 
Bart Schaefer           "And if you believe that, you'll believe anything."
                                                            -- DangerMouse
CSNET / Internet                schaefer@cse.ogc.edu
UUCP                            ...{sequent,tektronix,verdix}!ogccse!schaefer