Changelog for "service-qmail-smtpd-run" script http://qmail.jms1.net/scripts/ John Simpson 2005-10-17 jms1 - changed all "==" tests to "=". the script should now work under the original bourne shell as well as bash. 2005-11-24 jms1 - (no code changed.) changed the filename of the script as stored on the http://qmail.jms1.net/ web site. 2005-12-17 jms1 - (no code changed.) added the comment (above) about the web page with documentation about the "normal" sets of options. 2005-12-23 jms1 - added SMTP_CDB and SMTPSSL_CDB variables to set the locations of the tcpserver access control files 2005-12-26 jms1 - added MFCHECK, SPFBEHAVIOR and the six AUTH_SET_* variables available with the 6c2 patch. added a check for RECORDIO which allows it to be commented out or deleted in the top of the file. also added the QUSER variable so that people who need to run qmail-smtpd as a user other than the standard "qmaild" user can do so more easily. added comments for the first set of variables defined. 2006-01-02 jms1 - for 127.0.0.1 services, RBLSMTPD is now disabled along with the tcpserver authentication stuff. 2006-02-06 jms1 - adding support for "greylisting". if you want to use this, you can use the "qgreylist" program, available here: http://www.jonatkins.com/page/software/qgreylist or my own "jgreylist" program, available here: http://qmail.jms1.net/scripts/jgreylist.shtml 2006-03-21 jms1 - simplifying SMTP_CDB process, removed SMTPSSL_CDB (just use SMTP_CDB for the one file which affects this one service.) also made a few error messages more descriptive. 2006-04-12 jms1 - adding support for the AUTH_CDB patch 2006-04-21 jms1 - adding support for SPF_LOG and SPF_BLOCK_PLUS_ALL 2006-07-21 jms1 - fixing some typos, adding the AUTH_CDB stuff again (which seems to have not made its way into the file for some reason) 2006-07-25 jms1 - commenting out the AUTH_CDB line (which should have been commented out by default.) you should UN-comment this line if you're going to use the AUTH_CDB functionality. 2006-08-17 jms1 - adding "-t0" to rblsmtpd command line. if a client connects and they're blacklisted, this makes rblsmtpd just show them the error message and immediately hang up. why waste the time on them? 2006-08-21 jms1 - finished the web page for my own greylisting program. re-arranging the final command line so that the greylisting program is called before rblsmtpd. why do RBL checks (which cost time and network traffic) for an IP that you're only going to block anyway? (the greylist check takes a lot less time.) 2006-10-19 jms1 - changing the order in the final command line so that rblsmtpd (compiled C) runs before jgreylist (interpreted perl.) thanks to "Lance" for the suggestion. 2006-12-19 jms1 - removing "relays.ordb.org" from the default list of RBLs, since they have stopped providing their list. 2007-01-03 jms1 - adding variables to support newer features. 2007-01-16 jms1 - changing the script variable RBLSMTPD to RBLSMTPD_PROG to avoid confusion about an "RBLSMTPD" environment variable (which this script doesn't get involved with at all.) also added code to send a copy of the actual service command line to the log just before starting. 2007-03-04 jms1 - adding comments for the MAXLOAD, MAXCONNIP, MAXCONNC, and DIEMSG variables, which control how tcpserver works if you are using the "tcpserver limits" patch, documented here: http://linux.voyager.hr/ucspi-tcp/ 2007-03-04 jms1 - moving LOCAL to the normal configuration section so its value can be used for SMTPGREETING, DIEMSG, etc. 2007-04-08 jms1 - adding checks to not export empty variables. thanks to John Johnstone for the suggestion. 2007-05-10 jms1 - adding a generic place to specify an SSL/TLS key file, which will be used instead of control/servercert.pem. Thanks to Bob Walter for indirectly suggesting this. 2007-09-27 jms1 - moved the string of change notifications out to a separate changelog file. adding variables to support newer features, rearranging things so that options which logically belong together, ARE together. 2011-06-30 jms1 - adding RBLSMTPD_TIMEOUT variable, rather than using "-t0". Thanks to Niamh Holding for the suggestion.