src/mailman/mta/bulk.py

   1 # Copyright (C) 2009 by the Free Software Foundation, Inc.
   2 #
   3 # This file is part of GNU Mailman.
   4 #
   5 # GNU Mailman is free software: you can redistribute it and/or modify it under
   6 # the terms of the GNU General Public License as published by the Free
   7 # Software Foundation, either version 3 of the License, or (at your option)
   8 # any later version.
   9 #
  10 # GNU Mailman is distributed in the hope that it will be useful, but WITHOUT
  11 # ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  12 # FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
  13 # more details.
  14 #
  15 # You should have received a copy of the GNU General Public License along with
  16 # GNU Mailman.  If not, see <http://www.gnu.org/licenses/>.
  17
  18 """Module stuff."""
  19
  20 from __future__ import absolute_import, unicode_literals
  21
  22 __metaclass__ = type
  23 __all__ = [
  24     'BulkDelivery',
  25     ]
  26
  27
  28 import logging
  29 import smtplib
  30
  31 from itertools import chain
  32
  33 from zope.interface import implements
  34
  35 from mailman.config import config
  36 from mailman.interfaces.mta import IMailTransportAgentDelivery
  37 from mailman.mta.connection import Connection
  38
  39
  40 log = logging.getLogger('mailman.smtp')
  41
  42 # A mapping of top-level domains to bucket numbers.  The zeroth bucket is
  43 # reserved for everything else.  At one time, these were the most common
  44 # domains.
  45 CHUNKMAP = dict(
  46     com=1,
  47     net=2,
  48     org=2,
  49     edu=3,
  50     us=3,
  51     ca=3,
  52     )
  53
  54
  55 \f
  56 class BulkDelivery:
  57     """Deliver messages to the MTA in as few sessions as possible."""
  58
  59     implements(IMailTransportAgentDelivery)
  60
  61     def __init__(self, max_recipients=None):
  62         """Create a bulk deliverer.
  63
  64         :param max_recipients: The maximum number of recipients per delivery
  65             chunk.  None, zero or less means to group all recipients into one
  66             big chunk.
  67         :type max_recipients: integer
  68         """
  69         self._max_recipients = (max_recipients
  70                                 if max_recipients is not None
  71                                 else 0)
  72         self._connection = Connection(
  73             config.mta.smtp_host, int(config.mta.smtp_port),
  74             self._max_recipients)
  75
  76     def chunkify(self, recipients):
  77         """Split a set of recipients into chunks.
  78
  79         The `max_recipients` argument given to the constructor specifies the
  80         maximum number of recipients in each chunk.
  81
  82         :param recipients: The set of recipient email addresses
  83         :type recipients: sequence of email address strings
  84         :return: A list of chunks, where each chunk is a set containing no
  85             more than `max_recipients` number of addresses.  The chunk can
  86             contain fewer, and no packing is guaranteed.
  87         :rtype: list of sets of strings
  88         """
  89         if self._max_recipients <= 0:
  90             yield set(recipients)
  91             return
  92         # This algorithm was originally suggested by Chuq Von Rospach.  Start
  93         # by splitting the recipient addresses into top-level domain buckets,
  94         # using the "most common" domains.  Everything else ends up in the
  95         # zeroth bucket.
  96         by_bucket = {}
  97         for address in recipients:
  98             localpart, at, domain = address.partition('@')
  99             domain_parts = domain.split('.')
 100             bucket_number = CHUNKMAP.get(domain_parts[-1], 0)
 101             by_bucket.setdefault(bucket_number, set()).add(address)
 102         # Fill chunks by sorting the tld values by length.
 103         chunk = set()
 104         for tld_chunk in sorted(by_bucket.values(), key=len, reverse=True):
 105             while tld_chunk:
 106                 chunk.add(tld_chunk.pop())
 107                 if len(chunk) == self._max_recipients:
 108                     yield chunk
 109                     chunk = set()
 110             # Every tld bucket starts a new chunk, but only if non-empty
 111             if len(chunk) > 0:
 112                 yield chunk
 113                 chunk = set()
 114         # Be sure to include the last chunk, but only if it's non-empty.
 115         if len(chunk) > 0:
 116             yield chunk
 117
 118     def deliver(self, mlist, msg, msgdata):
 119         """See `IMailTransportAgentDelivery`."""
 120         recipients = msgdata.get('recipients')
 121         if recipients is None:
 122             return
 123         # Blow away any existing Sender and Errors-To headers and substitute
 124         # our own.  Our interpretation of RFC 5322 $3.6.2 is that Mailman is
 125         # the "agent responsible for actual transmission of the message"
 126         # because what we send to list members is different than what the
 127         # original author sent.  RFC 2076 says Errors-To is "non-standard,
 128         # discouraged" but we include it for historical purposes.
 129         del msg['sender']
 130         del msg['errors-to']
 131         # The message metadata can override the calculation of the sender, but
 132         # otherwise it falls to the list's -bounces robot.  If this message is
 133         # not intended for any specific mailing list, the site owner's address
 134         # is used.
 135         sender = msgdata.get('sender')
 136         if sender is None:
 137             sender = (config.mailman.site_owner
 138                       if mlist is None
 139                       else mlist.bounces_address)
 140         msg['Sender'] = sender
 141         msg['Errors-To'] = sender
 142         message_id = msg['message-id']
 143         for recipients in self.chunkify(msgdata['recipients']):
 144             try:
 145                 refused = self._connection.sendmail(
 146                     sender, recipients, msg.as_string())
 147             except smtplib.SMTPRecipientsRefused as error:
 148                 log.error('%s recipients refused: %s', message_id, error)
 149                 refused = error.recipients
 150         return refused