github.com/mailhog/smtp implements an SMTP server state machine.

It attempts to encapsulate as much of the SMTP protocol (plus its extensions) as possible without compromising configurability or requiring specific backend implementations.

• ESMTP server implementing RFC5321
• Support for:
  • AUTH RFC4954
  • PIPELINING RFC2920
  • STARTTLS RFC3207

proto := NewProtocol()
reply := proto.Start()
reply = proto.ProcessCommand("EHLO localhost")
// ...

See MailHog-Server and MailHog-MTA for example implementations.

Interaction with the state machine is via:

• the Parse function
• the ProcessCommand and ProcessData functions

You can mix the use of all three functions as necessary.

Parse should be used on a raw text stream. It looks for an end of line (\r\n), and if found, processes a single command. Any unprocessed data is returned.

If any unprocessed data is returned, Parse should be called again to process then next command.

text := "EHLO localhost\r\nMAIL FROM:<test>\r\nDATA\r\nTest\r\n.\r\n"
var reply *smtp.Reply
for {
  text, reply = proto.Parse(text)
  if len(text) == 0 {
    break
  }
}

ProcessCommand should be used for an already parsed command (i.e., a complete SMTP "line" excluding the line ending).

ProcessData should be used if the protocol is in DATA state.

reply = proto.ProcessCommand("EHLO localhost")
reply = proto.ProcessCommand("MAIL FROM:<test>")
reply = proto.ProcessCommand("DATA")
reply = proto.ProcessData("Test\r\n.\r\n")

The state machine provides hooks to manipulate its behaviour.

See for more information.

The state machine also exports variables to control its behaviour:

See for more information.

Copyright ©‎ 2014-2015, Ian Kent (https://iankent.uk)

Released under MIT license, see LICENSE for details.