sendmail interfaces

There are three ways sendmail communicates, both in receiving and in sending mail. These are:

• using the conventional UNIX system argument vector/return status

• speaking SMTP over a pair of UNIX system pipes

• speaking SMTP over a socket

For simplicity, the following descriptions of these three methods assume sendmail is sending to a mailer.

Argument vector/exit status

This technique is the standard UNIX system method for communicating with a process. A list of recipients is sent in the argument vector, and the message is sent on the standard input. This is the method by which sendmail communicates with execmail, rmail, lmail, and uux. Anything that the recipient mailer prints is simply collected and sent back to the mail sender if there were any problems. The exit status from the mailer is collected after the message is sent, and a diagnostic is printed if appropriate.

SMTP over pipes

The SMTP protocol can be used to run an interactive lock-step interface with a mailer over a pair of named pipes. A subprocess is still created, but no recipient addresses are passed to the mailer via the argument list. Instead, they are passed one at a time in commands sent to the mailer's standard input. Anything appearing on the mailer's standard output must be an SMTP reply code. This method is not used in the standard configuration. This method is useful when the mailer is a program, such as lmail, but the number of recipients is larger than the allowable number of arguments that can be passed in the argument vector.

SMTP over a socket

This method is similar to SMTP over pipes, except that it uses a socket. This method is exceptionally flexible in that the mailer need not reside on the same machine. It is normally used to connect to a sendmail process on another machine. This is the method by which sendmail routes and receives mail across the Internet via TCP/IP.