IMPORTANT: The facility being described here is an extension to the VM mailer software. The extension happens to have release 1.7b of LISTSERV as a pre-requisite, but it is NOT an enhancement of the LISTSERV software: you cannot assume that this facility is available on all sites running version 1.7b or higher. ABSTRACT: A modification to the 'LOCAL' delivery exit (requiring version 2.08 of the VM mailer) makes it possible for VM sites running LISTSERV version 1.7b or higher to support the generic local addresses 'owner-xyz' and 'xyz-request' without requiring any configuration/maintenance of the mailer's SPECIAL table. The mailer passes on requests for such addresses to LISTSERV, which examines the header of the corresponding list to extract the relevant addresses and delivers the message. JUSTIFICATION: This function both makes LISTSERV and BITNET a more friendly environment to Internet-literate users and provides convenient access to list owners/maintainers for BITNET users with RFC822 mail. Most importantly, this service comes at no manpower cost to the LISTSERV site, beyond initial installation: there is no need to define mailer aliases and keep track of changes made by list owners so as to make corresponding changes to the mailer tables. DESCRIPTION: The change modifies the behaviour of the 'LOCAL' delivery exit, causing it to look for addresses of the form 'owner-xyz' or 'xyz-request' if it failed to find a match in the SPECIAL table (an address of the form 'owner-xyz' which has an entry in the SPECIAL table is processed according to that entry, and not to the new logic). Case is ignored when performing that comparison. If the address matches either form, the mailer will attempt to (silently) pass on the message to the local user 'LISTSERV'; if there is no such local user, normal processing continues. No message is issued as the message is forwarded to LISTSERV. When LISTSERV receives the request, it checks for a local list by the specified name; if no list by that name is found, it delivers the message to the LISTSERV maintainers (who can then look for misspelled list names) and exits. If the list exists, the message is delivered to the appropriate recipients, based on the mailbox used: owner-xyz, which is normally used for delivery errors, is mapped to the contents of the "Errors-To=" keyword. xyz-request, normally used for administrative requests, is mapped to the people listed in the various "Owner=" keywords. INSTALLATION: >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> > WARNING - LISTSERV release 1.7b and mailer release 2.08 are required! < <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< Note that release 1.7b of LISTSERV is still being beta-tested as this note is written; please do not report the lack of 1.7b update for your site unless you had requested to beta-test the code. Sites with a source distribution: order LOCAL ERIC01 from LISTSERV@SEARN, add it to your LOCAL AUXLCL file, re-assemble the LOCAL exit, and rebuild MAILER MODULE. Sites with an object distribution only: due to the licensing conditions for the Waterloo C library code included in MAILER MODULE, the updated MODULE file may not be made available from SEARN. Instead, you should order LOKAL TEXT from LISTSERV@SEARN, place it on your mailer's A-disk, and use the LOKAL exit as a replacement for LOCAL in your MAILER MTPLATE/PROFILE files (WARNING: make sure to check you are running release 2.08 of MAILER!). DISCLAIMER: This code was written in a couple hours, using unearthed memories about the internals of MAILER V1 gathered during an assignment in 1987. It works, but is not satisfactory: no tracing, waste of addressability space through not calling new powerful routines the author was not familiar with and, most importantly, the function cannot be disabled through a configuration variable (this is a serious shortcoming for sites not running release 1.7b). Thus, this code cannot be included in standard mailer distributions as is. Still, one can hope that John Wagner will be kind enough to consider the possibility of including a better version of this function in a future release. Eric