Message Text

For outbound messages in MIME mode, the content-type depends on whether there are attachments and what the message text looks like. If there are attachments, the Content-type is multipart/mixed; the message text and each attachment become a separate part of the message content, each with its own content-type. If there are no attachments, the content-type of the message is text/plain and there is only one part.

The message text is not line-wrapped unless some line exceeds 140 characters in length. If one does, the entire text is wrapped to 76 columns and the quoted-printable encoding is used to preserve line breaks. The content-type depends on what characters are found in the message text, as follows:

• If only 7-bit characters are found and no line exceeds 140 characters in length, the message is ASCII text.

  Content-type: text/plain; charset=us-ascii
  (Content-Transfer-Encoding=7bit is assumed.)

• If long lines or 8-bit characters are found, the message is text and the character set is determined by the locale. It should be chosen from the character sets defined by ISO standard 8859.

  Content-type: text/plain; charset=iso-8859-1 (or another valid charset)
  Content-Transfer-Encoding: quoted-printable

For inbound MIME messages, if the first message content part has Content-type: text/* (that is, any text type) and its character set is recognized, it is mapped to PR_BODY. A first message content part not meeting this criterion becomes an attachment. Any subsequent parts also become attachments.

In uuencode mode, message text in outbound messages is line-wrapped to 78 columns, as for MS Mail 3.x. The content-type is "text/plain." To preserve the original message's paragraph breaks under these circumstances, observe the following conventions in the wrapped text. There are three possible reasons for ending a line of text, each with its own character sequence:

• Line-break. The original text contained a newline entered by the user (paragraph mark). In the transport, this maps to a newline with no preceding blanks. If the user enters a newline preceded by blanks, the blanks should be stripped out.
• Line-nobreak. The original text contained a word too long to fit on a single line of the message. In the transport, this maps to a newline preceded by two blanks.
• Line-wrap. The original text contained no newline, the text is too long to fit on a single line of the message, but it can be broken between two words. In the transport, this maps to a newline preceded by a single blank.