Rev 153 |
Compare with Previous |
Last modification |
| RSS feed
Changelog for relaydelay.pl and associated files:
2005-05-28 - Added a config option to force checking for all of sendmail's
mailers rather than just ones that look smtp-like.
2004-05-28 - Added a config option to check for stale db handles.
2004-05-22 - Included two little utilities to do simple queries on the db
and easily add white/blacklist entries.
2004-05-14 - A few corrections for the use of the $verbose flag. Added a
suggested modification to make db_connect discover db problems faster, but
left it commented out because it causes alot of latency if the db isn't
2004-05-13 - Integrated a patch from Martin A. Walker <firstname.lastname@example.org>
to allow logging to a logfile and running as a daemon.
Changed the accessdb checks to use BerkeleyDB module instead of DB_File,
as it appears to be more compatible with the db versions sendmail uses on
more distributions. Left lines in source but commented out for using
DB_File. In case anyone needs to, they can revert to using DB_File easily.
Added a check to verify that the dns_name table exists if name updates are
2004-05-12 - Rewrote the db_maintenance.pl script so that it now segments
the work into sizeable chunks that keep it from keeping the db tables
locked for excessive periods of time.
2003-10-28 - Fixed cosmetic problem.
2003-08-11 - Attempted to fix a reported milter issue with privdata.
2003-08-11 - Fixed some non-numeric warnings.
2003-07-23 - Added checks to allow manual blacklisting within the greylist
db based on mail_from address/domain. Also added some extra info to
a few debugging messages.
2003-07-22 - Added functionality to check the sendmail access db and
bypass the greylist checks if an appropriate entry is found.
Also rearranged the checks so that rows in the greylist database will
have their message counts updated even if there is another test that
causes the greylist checks to be skipped.
2003-07-22 - Cleaned up a small mite in the reverse_track function. It
probably won't cause anyone trouble, but fixed anyway.
0.04 - Release
2003-07-21 - Fixed a pretty serious bug if the config option
$tempfail_messages_after_data_phase was enabled. Please do not use
version 0.03 for this reason.
0.03 - Release
2003-07-19 - A few cosmetic changes and comment updates.
2003-07-15 - Added an update whereby manual whitelist entries for localhost
and other local interface addresses are no longer necessary, as the
milter now considers any mail that is "looped back" on a local interface
as being local and exempt from the greylist checks.
2003-07-15 - Fixed a bug where mail that looks as if it is spoofed from
a local domain was not being checked, due to sendmail lying about it
coming from the "local" mailer when it's really not local.
2003-07-13 - CONFIG CHANGE - Changed the milter to have config parameters
for the milter/sendmail interface. Now the only option to the milter
script is an optional config file if different from the default location.
This also removes the requirement of modifying the milter script if the
config is stored in a different location.
2003-07-13 - NEW OPTION - Added a config option to limit the number of
threads the milter will spawn. Once a thread is spawned it will continue
to exist until the milter is shutdown. Since each thread increases the
memory footprint, and consumes a db connection, this is a good way to
2003-07-12 - Fixed a small problem where bounce messages were not properly
expired as they should be. Also fixed some minor warnings.
2003-07-11 - Backed out changes for trying to fix the accounting error of
a few days ago. The "solution" caused more trouble than the problem.
2003-07-10 - NEW OPTION - Added a new feature to automatically
pre-whitelist recipients of locally generated outbound mail. Mail that
is relayed for another IP-wildcard whitelisted host, or sent from an
authenticated sender is also considered local for the purposes of this
2003-07-10 - Removed the table locking from the milter. Made some changes
to a few queries to make the ordering of returned rows deterministic.
As a result of the removing the DB locks, there is now a very small
(miniscule) race condition that may allow duplicate rows to be
inserted into the database, but the changes to the queries makes it so
that duplicates will not cause any noticable harm.
2003-07-09 - Uncovered a slight accounting problem in the milter that
caused some rows not to be expired as they should, as well as
incorrectly updating the passed_count. The problem only shows itself
if there is another MX host forwarding for a domain. Changed the milter
so that we do not increment the passed count if the outmailer is not
local (if the mail will be forwarded to another host). This will keep
the counts of individual emails correct (rather than relay hops), and
will keep records stemming from mails to invalid users from being
upgraded to whitelisted rather than immediately expired.
2003-07-07 - Discovered postfix normally uses <email@example.com> for
its envelope from address when doing SMTP callbacks. Changed milter to
treat any email from postmaster@ as a callback and delay the tempfail
until after the DATA phase.
0.02 - Release
2003-06-26 - Allowed for setting the initial delay to 0 seconds, which
allows new users to run in a "learning only" mode where no emails
will be tempfailed. Also, if the initial delay is set to a negative
number, the first delivery attempt of unknown triplets will be failed,
but any subsequent attempts will be allowed (within other time limits).
2003-06-26 - Added a new table to the database definition for reporting,
and a new script to move expired entries to the reporting table to help
keep the active table to a minimum size (for speed).
2003-06-26 - Added a check to avoid greylisting checks if the mail client
is authenticated with sendmail. Suggested by Borja Perez.
2003-06-26 - Minor comment changes. Update to whitelist_ip.txt.
2003-06-25 - Changed defaults in relaydelay.pl and relaydelay.conf to be
58 minutes for initial timeout, with a 5 hour window.
2003-06-25 - Added whitelist_ip.txt file to the sources, describing IPs
and networks that most users of a Greylisting implementation will want.
0.01 - Initial Release