Mailman User Guide

Mailman is a microframework for processing incoming email.

Here is an example Mailman app that takes incoming messages to a support email account, and adds them to a database.

# mailman_app.rb
require 'mailman'

Mailman.config.maildir = '~/Maildir'

Mailman::Application.run do
  to 'support@example.org' do
    Ticket.new_from_message(message)
  end
end

The Mailman app could then be started by running ruby mailman_app.rb.

Installation

Installation is as simple as gem install mailman.

Routes & Conditions

A Condition specifies the part of the message to match against. to, from, and subject are some valid conditions. A Matcher is used by a condition to determine whether it matches the message. Matchers can be strings or regular expressions. One or more Condition/Matcher pairs are combined with a block of code to form a Route.

Matchers

There are string, regular expression and header matchers. All can perform captures.

String

String matchers are very simple. They search through a whole field for a specific substring. For instance: 'ID' would match Ticket ID, User ID, etc.

They can also perform named captures. '%username%@example.org' will match any email address that ends with @example.org, and store the user part of the address in a capture called username. Captures can be accessed by using the params helper inside of blocks, or with block arguments (see below for details).

The capture names may only contain letters and underscores. Behind the scenes they are compiled to regular expressions, and each capture is the equivalent to .*. There is currently no way to escape % characters. If a literal % is required, and Mailman thinks it is a named capture, use a regular expression matcher instead.

Regular expression

Regular expressions may be used as matchers. All captures will be available from the params helper (params[:captures]) as an Array, and as block arguments.

Headers

You can match against headers using strings or regular expressions. Multiple headers can be listed and all of them must be matched in order for the route to progress.

Capture groups are available for headers matched with regular expressions, although they are provided in a slightly more complex form, as some headers can appear multiple times.

For an email that looks like this:

To: hello@mailman.example.com
X-Forwarded-To: someone@example.com
X-Forwarded-To: other@example.in

You can expect the captures to work like this:

Mailman::Application.run do
  header to: /h(ell)o/, x_forwarded_to: /(.+)@example.(.+)/ do
    p params[:captures]
    # => {
    #   to: [
    #     [ 'ell' ]
    #   ],
    #   x_forwarded_to: [
    #     [ 'someone', 'com' ],
    #     [ 'other', 'in' ]
    #   ]
    # }
  end
end

Using block arguments isn't advised for header matching as the exact number of arguments is limited only by the inbound email.

Routes

Routes are defined within a Mailman application block:

Mailman::Application.run do
  # routes here
end

Messages are passed through routes in the order they are defined in the application from top to bottom. The first matching route's block will be called.

Condition Chaining

Conditions can be chained so that the route will only be executed if all conditions pass:

to('support@example.org').subject(/urgent/) do
  # process urgent message here
end

Special routes

The default route is a catch-all that is run if no other routes match:

default do
  # process non-matching messages
end

Block Arguments

All captures from matchers are available as block arguments:

from('%user%@example.org').subject(/Ticket (\d+)/) do |username, ticket_id|
  puts "Got message from #{username} about Ticket #{ticket_id}"
end

Class Routing

Messages can also be routed to methods. For instance, to route to an Object with a receive instance method defined, this will work:

from '%user%@example.org', Sample

Messages can also be routed to arbitrary instance methods:

from '%user%@example.org', 'ExampleClass#new_message'

The method should accept two arguments, the message object, and the params:

def receive(message, params)
  # process message here
end

Route Helpers

There are two helpers available inside of route blocks:

The params hash holds all captures from matchers:

from('%user%@example.org').subject(/RE: (.*)/) do
  params[:user] #=> 'chunkybacon'
  # it is an indifferent hash, so you can use strings and symbols
  # interchangeably as keys
  params['captures'][0] #=> 'A very important message about pigs'
end

The message helper is a Mail::Message object that contains the entire message. See the mail docs for information on the properties available.

Conditions

Currently there are five conditions available: to, from, cc, subject, body

More can be added easily (see lib/mailman/route/conditions.rb).

Receivers

There are currently three types of receivers in Mailman: Standard Input, Maildir, and POP3. If IMAP or any complex setups are required, use a mail retriever like getmail with the Maildir receiver.

Standard Input

If a message is piped to a Mailman app, this receiver will override any configured receivers. The app will process the message, and then quit. This receiver is useful for testing and debugging. This feature can be disabled with the Mailman.config.ignore_stdin option.

Example: cat plain_message.eml | ruby mailman_app.rb

Note that the standard input receiver is not supported on Windows platforms.

POP3

The POP3 receiver is enabled when the Mailman.config.pop3 hash is set. It will poll every minute by default (this can be changed with Mailman.config.poll_interval). After new messages are processed, they will be deleted from the server. No copy of messages will be saved anywhere after processing. If you want to keep a copy of messages, it is recommended that you use a mail retriever with the Maildir receiver. You could also use Gmail and set it to keep messages after they have been retrieved with POP3.

You can pass a Hash to ssl with SSL context options. For example, when you have a self-signed certificate: ssl: { ca_file: '/etc/pki/my_ca.pem' }.

IMAP

The IMAP receiver is enabled when the Mailman.config.imap hash is set. Polling can be set with Mailman.config.poll_interval. This will read all unread messages in the INBOX by default. Here are example settings for gmail.

Mailman.config.imap = {
  server: 'imap.gmail.com',
  port: 993, # you usually don't need to set this, but it's there if you need to
  ssl: true,
  # Use starttls instead of ssl (do not specify both)
  #starttls: true,
  username: 'foo@somedomain.com',
  password: 'totallyunsecuredpassword'
}

When using gmail, remember to enable IMAP
You can pass a Hash to ssl, just like with POP3.

HTTP

If Mailman.config.http is set then the HTTP receiver will be used. This will set up an HTTP server which expects emails to be delivered to http://0.0.0.0:6245/ by default. You can alter the listening endpoint with the options below and altering the parser will allow different SMTP-HTTP gateways to be used.

NB. You will need to make sure rack (and optionally thin) are included in your gemset in order to use the HTTP receiver.

The default options are:

Mailman.config.http = {
  host: '0.0.0.0',
  port: 6245,
  path: '/',
  parser: :raw_post,
  parser_opts: {
    part_name: 'message'
  }
}

Raw Post

The Raw Post Parser (:raw_post) expects the raw contents of emails to be delivered via multipart POST request to the specified endpoint. You can specify the name of the part which will contain the email data with the :part_name parser option ('message' by default).

The default HTTP configuration is perfect for Cloudmailin. If you set your target as http://yourpublicserver:6245/ then mailman will work with the most basic config:

Mailman.config.http = {}

Mailman::Application.run do
  # ... etc
end

Sendgrid

The parser for Sendgrid (:sendgrid) expects requests to conform to the Inbound Parse Webhook specification. Once you have pointed your domain's MX record at mx.sendgrid.net and configured an inbound address that points to your mailman server and the path you specified, all requests should flow as expected.

Mailman.config.http = {
  parser: :sendgrid,
  path: "/emails"
}

Mailman::Application.run do
  # ... etc
end

Maildir

The Maildir receiver is enabled when Mailman.config.maildir is set to a directory. If the cur, new, and tmp folders do not already exist in the folder, they will be created. All messages in new folder will be processed when the application launches, then moved to the cur folder, and marked as seen. After processing these messages, Mailman will use the fssm gem to monitor the new folder, and process messages as they are created.

Configuration

Configuration is stored in the Mailman.config object. All paths are relative to the process's working directory or absolute if starting with a /.

Logging

Mailman.config.logger can be set to a Logger instance. You should change this if you want to log to a file in production.

Example: Mailman.config.logger = Logger.new('logs/mailman.log')

Default: Logger.new(STDOUT)

POP3 Receiver

Mailman.config.pop3 stores an optional POP3 configuration hash. If it is set, Mailman will use POP3 polling as the receiver.

Example:

Mailman.config.pop3 = {
  :username => 'chunkybacon@gmail.com',
  :password => 'foobar',
  :server   => 'pop.gmail.com',
  :port     => 995, # you can usually omit this, but it's there
  :ssl      => true # defaults to false
}

Polling

Mailman.config.poll_interval is the duration in seconds to wait between checking for new messages on the server. It is currently only used by the POP3 receiver. If it is set to 0, Mailman will do a one-time retrieval and then exit.

Default: 60

Maildir

Mailman.config.maildir is the location of a Maildir folder to watch. If it is set, Mailman will use Maildir watching as the receiver.

Example: Mailman.config.maildir = '~/Maildir'

Rails

Mailman.config.rails_root is the location of the root of a Rails app to load the environment from. If this option is set to nil, Rails environment loading will be disabled.

Default: '.'

Standard input receiver

Mailman.config.ignore_stdin disables the STDIN receiver, which can interfere with running Mailman with cron or as a daemon.

Default: false

Graceful death

Mailman.config.graceful_death, if set, will catch SIGINTs (Control-C) and allow the mail receiver to finish its current iteration before exiting. Note that this currently only works with POP3 receivers.

Middleware

Mailman.config.middleware gives you access to the Mailman middleware stack. Middleware allows you to execute code before and after each message is processed. Middleware is super useful for things like error handling and any other actions you need to do before/after each message is processed.

Here's an example of some simple error logging middleware:

# Define the middleware
class ErrorLoggingMiddleware
  def call(mail)
    begin
      yield
    rescue
      puts "There was an error processing this message! #{mail.subject}"
      raise
    end
  end
end

# Add it to the Mailman middleware stack
Mailman.config.middleware.add ErrorLoggingMiddleware

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

USER_GUIDE.md

USER_GUIDE.md

Mailman User Guide

Installation

Routes & Conditions

Matchers

String

Regular expression

Headers

Routes

Condition Chaining

Special routes

Block Arguments

Class Routing

Route Helpers

Conditions

Receivers

Standard Input

POP3

IMAP

HTTP

Raw Post

Sendgrid

Maildir

Configuration

Logging

POP3 Receiver

Polling

Maildir

Rails

Standard input receiver

Graceful death

Middleware

Files

USER_GUIDE.md

Latest commit

History

USER_GUIDE.md

File metadata and controls

Mailman User Guide

Installation

Routes & Conditions

Matchers

String

Regular expression

Headers

Routes

Condition Chaining

Special routes

Block Arguments

Class Routing

Route Helpers

Conditions

Receivers

Standard Input

POP3

IMAP

HTTP

Raw Post

Sendgrid

Maildir

Configuration

Logging

POP3 Receiver

Polling

Maildir

Rails

Standard input receiver

Graceful death

Middleware