.sh 1 "Format of messages"
This section describes the format of messages.
line, which consists of the word
followed by a user name, followed by anything, followed by
a date in the format returned by the
library routine described in section 3 of the Unix Programmer's
date may be optionally followed by a single space and a
time zone indication, which
should be three capital letters, such as PDT.
Each header field line is of the form:
can be anything, but only certain header fields are recognized as
having any meaning. The recognized header fields are:
Other header fields are also significant to other systems; see,
for example, the current Arpanet message standard for much more
A header field can be continued onto following lines by making the
first character on the following line a space or tab character.
If any headers are present, they must be followed by a blank line.
The part that follows is called the
of the message, and must be ASCII text, not containing null characters.
Each line in the message body must be terminated with an ASCII newline
character and no line may be longer than 512 characters.
If binary data must be passed through the mail system, it is suggested
that this data be encoded in a system which encodes six bits into
For example, one could use the upper and lower case letters, the digits,
and the characters comma and period to make up the 64 characters.
Then, one can send a 16-bit binary number
as three characters. These characters should be packed into lines,
preferably lines about 70 characters long as long lines are transmitted
The message delivery system always adds a blank line to the end of
each message. This blank line must not be deleted.
The UUCP message delivery system sometimes adds a blank line to
the end of a message each time it is forwarded through a machine.
It should be noted that some network transport protocols enforce
limits to the lengths of messages.