дано mdaemon не сильно свежий (6.5.1)
демон помимо своего домена держит gateway на нутряной почтовик (типа пока так надо, скажем domain1.ru)
соот-но проверить валидность адресов из domain1.ru он не может и жрёт всё подряд Ж-(
Но ходют слухи что в более свежих версиях есть такая фишка нюхать по ldap кто существует в domain1.
Соот-но вопрос к тем у кого есть распоследний демон Ж-) есть ли у Вас такая настройка для gateway ?
вопрос больше не актуален.
демон найден. тестовая проверка по ldap прокатила. бум апгрейдить %-)
з.ы. что ещё понравилось - расширили антиспамовские фишки %-)
Spam Filtering Using Spam-Assassin(tm)
Spam filtering using SpamAssassin technology has been integrated into MDaemon. SpamAssassin is a heuristic mail filter that examines email messages using numerous rules in order to compute a 'score'. This score is then used to determine the likelihood of a message being spam. By examining many thousands of known spam messages the rules have been optimized over time and are very reliable in detecting the fingerprint of a spam message.
You can optimize, customize, or add new rules by editing the configuration files which are part of the SpamAssassin software package. For complete information on this and on all that SpamAssassin brand software can do visit http://www.SpamAssassin.org.
A new 'Security' top-level menu contains the "Spam filtering" option which loads the configuration GUI for the SpamAssassin functionality. Also, a SPAM tab has been added to the main GUI which reports all the spam filtering activity. A new log file called MDaemon-SPAM.log will track spam filtering activity. You can elect not to use this new log via a new setting in the logging options dialog. Also, the Misc. Options dialog has a new option to add Spam logging to the Composite log.
SpamAssassin is a trademark of Deersoft, Inc.
и это
Bayesian Engine for Spam / Non-Spam Detection
Bayesian support within the new Spam Filtering system has been completed. Bayesian classification is a mathematical process whereby machines can actually learn. Through repeated exposure to known samples of spam (and non-spam) Bayesian enabled software can learn to recognize patterns. Over time, Bayesian classification techniques grow more and more reliable.
MDaemon supports Bayesian classification through its incorporation of the SpamFilter heuristic software and through use of a "learning" program that is used to rapidly process and index spam and non-spam sample messages. The learning tool, SA-Learn, will be executed by MDaemon at midnight each night. The tool will examine the content of a "known-spam" folder and a "known-non-spam" folder. You provide the folders and the content. Since the Bayesian results improve over time the scoring tends to get more and more accurate.
SA-Learn trusts that the messages you feed it really and truly are spam and/or non-spam. It takes your word that the input data is accurate. Here's what you should do:
Go to the Spam Filtering|Bayesian tab and use the "Pub Folder" buttons to create two public folders - one for false negatives (messages that should have been flagged as spam but weren't) and another folder for false positives (messages that were incorrectly flagged as spam). Make these folders accessible to all your IMAP users. Now, instruct your users to place copies of any false positive or false negative messages that they receive into these public folders. MDaemon will process them through the learning tool at midnight each night. After processing, these messages are automatically deleted.
