(c) 1997, 1998, 1999, 2004, 2005, 2006, 2008, 2009
  Fred Lindberg, lindberg@id.wustl.edu (code+docs+faq)
  Fred B. Ringel, fredr@rivertown.net (faq)
  Bruce Guenter, bruce@untroubled.org (code+docs)

This software is distributed under the GNU General Public Licence as
published by the Free Software Foundation.  See the file LICENSE for
the conditions under which this software is made available.

Like any other piece of software (and information generally), ezmlm-idx
comes with NO WARRANTY.

This software originated from ezmlm-0.53, written by Dan J. Bernstein,
which has now been placed into the public domain.

ezmlm provided basic message distribution, subscriber address handling,
message archiving (single quoted message retrieval only), and bounce
handling. ezmlm-idx adds multi-message threaded archive retrieval, digests,
remote administration, message and subscription moderation, subscriber-only
restrictions, message trailer, subject prefix, subscriber name storage,
customizable setup, multi-language and MIME support and more. In addition,
improved bounce handling, case-insensitive subscriber address storage,
and support for distributed lists (main + sublist transparent to the
subscriber). Ezmlm-idx now includes optional support for MySQL or
PostgreSQL subscriber databases.

See INSTALL for installation instructions.
See UPGRADE to upgrade from a previous version of ezmlm-idx.
See CHANGES to see what's new in this version.
See FAQ for more info on ezmlm(-idx) functions and setup.
See FILES for a list of files in this package.

NOTE: Some ezmlmrc translations are not up to date.  Please check the
status of the translations at http://www.ezmlm.org/lang/ and if your
language is out of date, consider submitting corrections.  Your
contribution will be acknowledged below in the next ezmlm-idx release.

The RPM will put ezmlm-cgi in the bin directory. To use it, you must manually
move it to a cgi-bin directory, set ownership and SUID (if needed), and
set up /ezmlm/ezcgirc.

See http://cr.yp.to/ezmlm.html for the latest information about ezmlm.

See http://www.ezmlm.org/ for the latest version of ezmlm-idx, as well
as on-line versions of the FAQ and docs.

Mail ``ezmlm-subscribe@lists.untroubled.org'' to join the ezmlm mailing
list (low traffic). This list is run by Bruce Guenter using ezmlm-idx.
The archive is viewable at http://lists.untroubled.org/?list=ezmlm

Send general comments, questions, bug reports, and patches to:
	ezmlm@lists.untroubled.org

ACKNOWLEDGEMENTS

Translations have been provided by:

TRANSLATIONS (from newest to oldest)
nl:	Arthur Muller <Arthur@FutureQuest.net>
it:	Tullio Andreatta <t.andreatta@troppoavanti.it>
nl:	Maurice Snellen <maurice@hypercube.demon.nl>
de:	Lars Kruse <lists@sumpfralle.de>
id:	PakOgah <pakogah@pala.bo-tak.info>
es:	Ruben Cardenal <ruben@ruben.cn>
sv:	Mads E. Eilertsen <mads.e.eilertsen@hist.no>
fr:	Jean-Marc Lasgouttes

Older translation credits:
de:	Frank Tegtmeyer
ja:	Masashi Fujita (objectx@polyphony.scei.co.jp)
da:	Torben Fjerdingstad (unitfj@tfj.uni-c.dk)
fr:	Frank Denis (j@industrie.capgemini.fr)
sv:	Sebastian Andersson (sa@hogia.net)
pl:	Sergiusz Pawlowicz (ser@arch.pwr.wroc.pl)
pt_BR:	Wanderlei Antonio Cavassin (cavassin@connectiva.com.br)
cs:	Jan Kasprzak (kas@informatics.muni.cz)
ru:	Roman V Isaev (rm@techno.ru)
id:	Aria Prima Novianto (aria@isnet.org)
it:	Roberto De Carlo (rodeca@flashnet.it)
es:	Vicent Mas, Francesc Alted, Sonia Lorente, and Cyndy DePoy
cn_GB:	HaiFeng Guo (haifeng@ms.lawton.com.cn)
hu:	Julian Severn-nek (http://js.hu/)
nl:	Willem <willem@king-pin.nl>

ACKNOWLEDGEMENTS (from Fred Lindberg)
- Fred B. Ringel, first and foremost, for a great collaboration, ideas, testing,
  docs clarification. This makes it so much more fun.
- Dan J. Bernstein for qmail, ezmlm, and great libraries.
- Tullio Andreatta, for the confirm (-y option) patches.
- Toshinori Maeno (TM; tmaeno@hpcl.titech.ac.jp) for finding an ezmlm-idx-0.20
  bug, and MIME suggestions, suggestions about hopcount and received headers,
  return-path header in archive, sublist loop detection, ezmlm-limit suggestion,
  many other suggestions and explanations, help to get ezmlm-cgi Japanese
  support to work, as well as pre-release testing.
- Frank Tegtmeyer for inciting digests and format info/suggestions, and for
  ezmlmrc.de.
- Magnus St�l�ker (MAS; stalaker@umc.se) for the PostgresSQL interface.
- Mark Delany for the original ezmlm-issub and for suggesting reordering
  unlink() and doit() in ezmlm-warn.
- Raul Miller and Chris Garrigues for format info/suggestions.
- Thomas Erskine (TEE; tom@crc.doc.ca) for fix to clean sunos-4.1.3 compiles
  and pre-release compilation on many platforms, fixes for bash-isms in
  ezmlm-check, and testing.
- Shuhei Kobayashi (SK; shuhei-k@jaist.ac.jp) for MIME corrections/suggestions,
  X-sequence suggestions, ezmlm-check/make corrections, and pre-release testing.
- Yusuf Goolamabbas (YG; yusufg@krdl.org.sg) and the Mutt developers for
  pointing out outdated MIME, ezmlmrc improvements, outformat bug, suggestion
  about ezmlm-moderate -r switch, and pre-release testing.
- Brian Gentry (BG; gentry@usaccess-inc.com) for reporting the
  ezmlm-moderate-0.12 exit code bug.
- Anand R. Buddhev (ARB; arb@iconnect.co.ke) for fix to clean BSDI 2.1 compiles
  and pre-release testing.
- Masashi Fujita (MF; objectx@polyphony.scei.co.jp) for fix to clean SGI
  compiles, lint advice, MIME suggestions, pre-release testing, a patch for
  the ezmlm-make-0.21 -c bug, ezmlmrc.ja, and finding postmsg bug in 0.301.
- Matthew D. Stock (MDS; stock@perdix.acsu.buffalo.edu) for pre-release testing.
- Ximenes Zalteca for finding an ezmlm-send-0.21 bug.
- Jukka Suomela (JS; jukka@narnia.tky.hut.fi) for inciting the ezmlm-send -cC
  switch.
- John White (johnjohn@triceratops.com) for questions leading to ezmlm-tstdig,
  and for testing the example script.
- Torben Fjerdingstad (TF; unitfj@tfj.uni-c.dk) for testing case-insensitive
  issub.c and subscribe.c, for TARGETS, for suggesting a no-copy-to-SENDER
  option, for ezmlmrc.da, for reporting missing MIME end for base64/QP
  moderated [un]sub confirms, and for pre-release testing.
- Shinya O'Hira (SOH; Shinya_Oohira@justsystem.co.jp) for reporting the
  ezmlm-make-0.22 lock file name bug and for testing the fix. Also for many
  helpful suggestions and hard work testing rfc2047 subject support.
- Sadhu(sadhu@aloha.net) & Tracy Reed (TR; treed@ultraviolet.org) for reporting
  the ezmlm-send-0.22 undefined SENDER bug, and testing the fix.
- Scott Balantyne (SDB; sdb@ssr.com) for inciting the modifications for digest
  out of dir/editor.
- Frank Denis (j@industrie.capgemini.fr) for ezmlmrc.fr.
- Marc Evans (marc@destek.net) for reporting the ezmlm-make-0.221 64-bit bug
  and pre-release testing, and inciting, improving, and testing ezmlm-test.
- Vince Vielhaber (VV; vev@michvhf.com) for reporting ezmlm-check problems with
  non-bash and testing fixes.
- Andrew Pam (AP; xanni@xanadu.net) for suggesting the ezmlm-idx -d switch and
  reporting the ezmlm-gate failure on NULL $SENDER bug.
- Glen Stewart (GS; glen_stewart@associate.com) for reminding me about command
  case insensitivity and ezmlmrc typo fixes.
- Kenji Ikeda (KI; noroi@nt.is.dnp.co.jp) for a patch to ezmlm-idx-0.23 to get
  message numbers in subjects. I've added code to ezmlm-idx-0.30 implementing
  this is a similar manner. Sen Nagata (sen_ml@eccosys.com) for suggesting this
  earlier. Bugfix for missing MIME boundary in -index reply.
- Sebastian Andersson (SA; sa@hogia.net) for ezmlmrc.sv, MIME suggestions,
  ezmlm-weed patch, ezmlmrc version check suggestion and pre-release testing.
- Giorgos Stathakopoulos (GS; stathako@cti.gr) for pre-release testing.
- David Summers (DS; david@summersoft.fay.ar.us) for pre-release testing,
  reporting a ezmlm-check bug and SPEC files for rpm:s.
- Steinar Haug (STH; sthaug@nethelp.no) for pointing out missing ';' in
  ezmlmrc (caused problems with /bin/sh on FreeBSD), and the trigger message
  received lines in the digest header.
- Peter Hunter (PH) for suggesting dir/headeradd and list address in digest
  headers and for reporting a "cosmetic" bug in subscription logging.
- Jim Simmons (JS) for finding the n/d switch misnaming in ezmlmrc for text file
  editing.
- Sergiusz Pawlowicz (SP; ser@arch.pwr.wroc.pl) for ezmlmrc.pl, and many entries
  for mimeremove.
- Wanderlei Antonio Cavassin (WAC; cavassin@connectiva.com.br) for
  ezmlmrc.pt_BR.
- Jim Simmons (     ) for finding a missing-newline-mime bug in
  ezmlm-moderate/store/clean.
- Monte Mitzelfelt (MM; monte@gonefishing.org) for suggesting alternative
  From: for ezmlm-manage help messages to break responder loops.
- Louis Larry (LL; luois@kediri.webindonesia.com) for reporting a 0.311 prefix
  handling bug.
- Sigi Remsmurr (SR; service@isk.de), Bill Nugent (whn@topelo.lopi.com), James
  Smallacombe (JS; up@3am) for ezmlmrc corrections/suggestions.
- Matthew Saunders (MS; matts@easynet.net) for ezmlm-glconf.sh corrections and
  reporting crashability of ezmlm-get by abnormal use (fixed).
- Petr Novotny (PN; Petr.Novotny@antek.cz) for ezdomo.tar.gz corrections.
- Evan Champion (EC; evanc@synapse.net) for ezmlm-request bug report.
- Jeff Hill (JH; jhill@hronline.com) for ezmlm-reject bug report.
- Bruno Wolff (BW; bruno@cerberus.umn.edu) for Ultrix fixes.
- Butch Evans (BE; butch@###.com) for reporting trailer problems with
  multipart/alternative messages and testing fix.
- Kragen Sitaker (KS; kragen@pobox.com) for reporting ezmlm-manage -get bug.
- Lars B. Rasmusson (LBR; lbr@mjolner.dk) for prerelease testing with sun cc.
- Mike McLeish (MMcL; Mike.Mcleish@chatsoft.com) for reporting problems
  using the 'x' format and testing the fix.
- Jan Kasprzak (JK; kas@informatics.muni.cz) for ezmlmrc.cs and headerremove
  suggestions.
- Matt McGlynn for reporting 0.32 ezmlmrc -return omission.
- Roman V Isaev (RVI; rm@techno.ru) for ezmlmrc.ru as well as suggestions
  on -allow automation.
- Aria Prima Novianto (APN; aria@isnet.org) for ezmlmrc.id.
- Roberto De Carlo (RDC; rodeca@flashnet.it) for ezmlmrc.it.
- Vicent Mas, Francesc Alted, Sonia Lorente, and Cyndy DePoy for ezmlmrc.es.
- HaiFeng Guo (HFG; haifeng@ms.lawton.com.cn) for ezmlmrc.cn_GB.
- Julian Severn-nek (http://js.hu/) for ezmlmrc.hu.
- Willem <willem@king-pin.nl> for ezmlmrc.nl.
- Benjamin Pflugman (BPF; philemon@spin.de) for reporting multipart/signed
  problems and testing fix.
- Rik Myers (RM; rik@sumthin.nu), Daniel Mattos (DM; daniel@tiii.com) and others
  for helping make ezmlm-test less platform-sensitive.
- Michael Hirohama (MH; michael@sooth.com) for reporting Outlook problems and
  testing a work-around.
- Mate Wierdl (MW; mw@wierdlpc.msci.memphis.edu) for SPEC file kit and
  suggestions on ezmlm-test, "it", etc, and pre-release testing.
- Numerous users for suggestions/corrections for the documentation.
- I'm sure I've forgotten others who have contributed ideas, questions,
  comments. Thanks!

INFORMATION ON BUILDING/USING EZMLM WITH MYSQL SUPPORT

Copyright 1999 Frederik Lindberg, lindberg@id.wustl.edu
Copyright 2007 Bruce Guenter, bruce@untroubled.org
		You may use under GPL.

For information on MySQL, see http://www.tcx.se.

Most of this information is available in FAQ.

If you are interested in contributing/testing a subscriber db interface
for another SQL server, please see README.std and the routines here, and
contact bruce@untroubled.org (it may already be in process). See end of
this file for other ways to contribute.

conf-mysql must be edited to reflect your system. On many systems, you
also need to include ``-lsocket'', as well as change the paths to the
/usr/local equivalents. For the i386.rpm-based systems, you need at
least MySQL-devel to build the files. Look at your mysql docs for more info.

TABLES USED FOR (My)SQL SUPPORT

The basic philosophy is that the database can be on any host (if you use
SENDER restrictions, connectivity to the main host is more important than
to the sublists), and you choose the database and "table root" names. The
default database is ``ezmlm'' and the default table root is ``list''. Each
list has a separate table root. Any number of lists can share a database.

The main list address table is named with the table root only, others have
that name with various suffixes. In the following ``list'' is used as the
table root.

ADDRESS TABLES
list		subscriber addresses
list_slog	subscriber address log
list_allow	subscriber aliases for posts on SENDER checked lists.
list_allow_slog	subscriber log for list_allow
list_deny	blacklisted addresses for posts on SENDER checked lists.
list_deny_slog	log for list_deny.
list_mod	moderator addresses
list_mod_slog	log for list_mod
list_digest	subscriber log for digest list.
list_digest_slog	log for list_digest

MESSAGE LOGGING TABLES
list_cookie	message cookie table for main list
list_mlog	message logging table for main list
list_digest_cookie	message cookie table for digest list
list_digest_mlog	message logging table for digest list

SUBLIST SPLIT TABLES
list_name	sublist split table for main list
list_digest_name	sublist split table for digest list.

The address tables contain (address,domain,hash,h,num). For normal
lists only the address field is used. For main->sublist clusters, the other
fields are used for load splitting. The domain is the first up to 3 characters
of the last part of the domain name. The hash is a address hash [0-52] differnt
from the one used by ezmlm for splitting within DIR/subscribers. When using
the address field as a primary key, the size of the index was unreasonable.
Therefore, ``num'' is used as a dummy primary key, and ``h'' (a 32 bit hash
of the address) is used as an index. This markedly speeds up (un)sub with
large (>30,000 rows) subscriber tables.

The *_slog tables contain the same info as DIR/Log, i.e. address, timestamp,
entry-type, entry-direction, and fromline. The entry-type is the first letter
of the type of entry (probe, manual, `` '' for normal), entry-direction is
``+'' for addition, ``-'' for removal. Fromline is the From: header contents
taken from the subscribe confirm message or from ezmlm-sub (if used with -n).
It is blank for all address removals, and may be blank also for additions. It
is used by the list-log.xx command. It is trivial to JOIN this table with the
address table to get e.g. subsciber names, subscription dates, etc. These
tables also have the 32-bit hash ``h'' as an index. Joins should be done on
``h'' as well as ``address'' for better performance.

The *_cookie tables contain message number, timestamp, and cookie. For each
message a pseudo-random cookie is generated that is ``impossible'' to guess
beforehand. For lists with sublists, this is used as basic authentication,
i.e. the sublist will refuse to process a message that doesn't contain the
correct cookie or that the sublist has already successfully processed.

The *_mlog tables contain log entries from main and sublists. These are
timestamp, listno, done. Listno is the lowest listnumber for an active list
entry with the name of this sublist as looked up in the *_name table. Done
is -1 for bounce, 0 for arrived, 1 for finished processing, and 2 for receipt
received. The routines are set up so that only the first attempt for each
combination (listno,code) is logged.

The *_name tables contain listno,name,domain,hash_lo,hash_hi,msgnum_lo,
msgnum_hi,notuse. Listno is auto_increment and unique. Name is the name of the
sublist. domain is the last up to 3 characters of the top domain name for
addresses served by this list (default = ''). It is is '', the list servers
all_domains_that_are_not_served_by_another list (in addition to domain '').
Of the addresses that match the domain criterion, the list serves the subset
with hash between hash_lo and hash_hi (defaults 0, 52). Any entry is ingnored if
notuse != 0 OR the current message number is not between msgnum_lo and
msgnum_hi.

For normal lists that are not distributed (i.e. they are a single list),
entries in the *_name tables are not needed and logging is not very
relevant.

For most lists, the only addresses that are stored in the SQL database are
the subscribers of list and digest, and the ``allow'' aliases. It is NOT
normally advisable to store moderator addresses there, since they are
needed only at the main list and secrecy is more important. ``Deny'' addresses
are few and again only needed at the main list. ``Allow'' are put in the
SQL database when using the default ezmlmrc file only to make all relevant
addresses manipulatable via the SQL server. The other tables are created, in
case they are  wanted (the cost for having them as empty table is zero). The
basedir/sql file is the decision point. If it exists, an SQL table is used;
if not a local ezmlm db is used.

CONTRIBUTIONS REQESTED

I would be very grateful if there are users out there willing to do any of
the following and contribute it to this package. Please check with me first
(bruce@untroubled.org), as the project may already be in progress/done.

1. Interfaces for other SQL servers. Oracle, SyBase, ...

2. A GUI admin utility to add/remove/manipulate the sublist split, essentially
   by modifying list_[digest_]name in a safe way. Ideally WWW if it can be
   done securely. If you use some standard interface (JDBC/DBD) it would be
   useful also with other SQL severs. This could even be an Access program
   using ODBC, although writing it for a platform running qmail/ezmlm makes
   most sense.

3. a WWW GUI that allows users to subscribe/unsubscribe in a safe way. A random
   password would be created the first time and stored in a new address->pw
   table and mailed to the subscriber address. With that password, the user
   would be able to [un]subscribe to lists, edit the name (for compatibility
   implemented by adding a subscribe line to list_[digest]slog). Add/remove
   aliases. Ideally, it should also allow searching by subscriber name. This
   would search *_slog.fromline. If less that 'x' alternatives are found, the
   user would be presented with names (not addresses), allowing the user to
   cause the subscription name and password to be sent to the respective
   subscription address. With that info, the subscriber can then unsubscribe,
   even if s/he has forgotten the subscription address. It is complicated
   slightly by the fact that ``fromline'' is the crude line and needs to be
   rfc822 parsed. Again, use of a standard interface is encouraged to make it
   compatible also with other SQL servers.

The aim of all this is to make it easy to use ezmlm to run very large lists,
easy to set up sites that handle subscriber interaction, archive access, etc,
and hopefully easier to integrate many ezmlm as done by some WWW sites today.

INFORMATION ON BUILDING/USING EZMLM WITH POSTGRESQL SUPPORT

Original source:
(c) 1999, Frederik Lindberg, lindberg@id.wustl.edu
		You may use under GPL.

and for PostgreSQL modifications:
(c) 1999, Magnus St�l�ker, stalaker@umc.se

Rewritten as a plugin by:
(c) 2007 Bruce Guenter, bruce@untroubled.org

For information on PostgreSQL, see http://www.postgresql.org/

This version of the PostgresSQL supports the basic subscriber address
database and subscription logging. The log that is searched by the -log.
command is the local listdir/Log file, NOT the SQL database. Message
logging is not supported, and central admin of list clusters is not
supported. Still, the current support covers the functionality used in
99% of lists.

Most of this information is available in FAQ.

If you are interested in contributing/testing a subscriber db interface
for another SQL server, please see README.std and the routines here,
and contact bruce@untroubled.org (it may already be in process). See end
of this file for other ways to contribute.

conf-mysql must be edited to reflect your system. On many systems, you
also need to include ``-lsocket'', as well as change the paths to the
/usr/local equivalents. For the i386.rpm-based systems, you need at
least MySQL-devel to build the files. Look at your mysql docs for more info.

TABLES USED FOR (My)SQL SUPPORT

The basic philosophy is that the database can be on any host (if you use
SENDER restrictions, connectivity to the main host is more important than
to the sublists), and you choose the database and "table root" names. The
default database is ``ezmlm'' and the default table root is ``list''. Each
list has a separate table root. Any number of lists can share a database.

The main list address table is named with the table root only, others have
that name with various suffixes. In the following ``list'' is used as the
table root.

ADDRESS TABLES
list		subscriber addresses
list_slog	subscriber address log
list_allow	subscriber aliases for posts on SENDER checked lists.
list_allow_slog	subscriber log for list_allow
list_deny	blacklisted addresses for posts on SENDER checked lists.
list_deny_slog	log for list_deny.
list_mod	moderator addresses
list_mod_slog	log for list_mod
list_digest	subscriber log for digest list.
list_digest_slog	log for list_digest

MESSAGE LOGGING TABLES (not supported by this version of the interface)
list_cookie	message cookie table for main list
list_mlog	message logging table for main list
list_digest_cookie	message cookie table for digest list
list_digest_mlog	message logging table for digest list

SUBLIST SPLIT TABLES (not supported by this version of the interface)
list_name	sublist split table for main list
list_digest_name	sublist split table for digest list.

The address tables contain (address,domain,hash,h,num). For normal
lists only the address field is used. For main->sublist clusters, the other
fields are used for load splitting. The domain is the first up to 3 characters
of the last part of the domain name. The hash is a address hash [0-52] differnt
from the one used by ezmlm for splitting within DIR/subscribers. When using
the address field as a primary key, the size of the index was unreasonable.
Therefore, ``num'' is used as a dummy primary key, and ``h'' (a 32 bit hash
of the address) is used as an index. This markedly speeds up (un)sub with
large (>30,000 rows) subscriber tables.

The *_slog tables contain the same info as DIR/Log, i.e. address, timestamp,
entry-type, entry-direction, and fromline. The entry-type is the first letter
of the type of entry (probe, manual, `` '' for normal), entry-direction is
``+'' for addition, ``-'' for removal. Fromline is the From: header contents
taken from the subscribe confirm message or from ezmlm-sub (if used with -n).
It is blank for all address removals, and may be blank also for additions. It
is used by the list-log.xx command. It is trivial to JOIN this table with the
address table to get e.g. subsciber names, subscription dates, etc. These
tables also have the 32-bit hash ``h'' as an index. Joins should be done on
``h'' as well as ``address'' for better performance.

The *_cookie tables contain message number, timestamp, and cookie. For each
message a pseudo-random cookie is generated that is ``impossible'' to guess
beforehand. For lists with sublists, this is used as basic authentication,
i.e. the sublist will refuse to process a message that doesn't contain the
correct cookie or that the sublist has already successfully processed.

The *_mlog tables contain log entries from main and sublists. These are
timestamp, listno, done. Listno is the lowest listnumber for an active list
entry with the name of this sublist as looked up in the *_name table. Done
is -1 for bounce, 0 for arrived, 1 for finished processing, and 2 for receipt
received. The routines are set up so that only the first attempt for each
combination (listno,code) is logged.

The *_name tables contain listno,name,domain,hash_lo,hash_hi,msgnum_lo,
msgnum_hi,notuse. Listno is auto_increment and unique. Name is the name of the
sublist. domain is the last up to 3 characters of the top domain name for
addresses served by this list (default = ''). It is is '', the list servers
all_domains_that_are_not_served_by_another list (in addition to domain '').
Of the addresses that match the domain criterion, the list serves the subset
with hash between hash_lo and hash_hi (defaults 0, 52). Any entry is ingnored if
notuse != 0 OR the current message number is not between msgnum_lo and
msgnum_hi.

For normal lists that are not distributed (i.e. they are a single list),
entries in the *_name tables are not needed and logging is not very
relevant.

For most lists, the only addresses that are stored in the SQL database are
the subscribers of list and digest, and the ``allow'' aliases. It is NOT
normally advisable to store moderator addresses there, since they are
needed only at the main list and secrecy is more important. ``Deny'' addresses
are few and again only needed at the main list. ``Allow'' are put in the
SQL database when using the default ezmlmrc file only to make all relevant
addresses manipulatable via the SQL server. The other tables are created, in
case they are  wanted (the cost for having them as empty table is zero). The
basedir/sql file is the decision point. If it exists, an SQL table is used;
if not a local ezmlm db is used.

CONTRIBUTIONS REQESTED

I would be very grateful if there are users out there willing to do any of
the following and contribute it to this package. Please check with me first
(bruce@untroubled.org), as the project may already be in progress/done.

1. Interfaces for other SQL servers. Oracle, SyBase, ...

2. A GUI admin utility to add/remove/manipulate the sublist split, essentially
   by modifying list_[digest_]name in a safe way. Ideally WWW if it can be
   done securely. If you use some standard interface (JDBC/DBD) it would be
   useful also with other SQL severs. This could even be an Access program
   using ODBC, although writing it for a platform running qmail/ezmlm makes
   most sense.

3. a WWW GUI that allows users to subscribe/unsubscribe in a safe way. A random
   password would be created the first time and stored in a new address->pw
   table and mailed to the subscriber address. With that password, the user
   would be able to [un]subscribe to lists, edit the name (for compatibility
   implemented by adding a subscribe line to list_[digest]slog). Add/remove
   aliases. Ideally, it should also allow searching by subscriber name. This
   would search *_slog.fromline. If less that 'x' alternatives are found, the
   user would be presented with names (not addresses), allowing the user to
   cause the subscription name and password to be sent to the respective
   subscription address. With that info, the subscriber can then unsubscribe,
   even if s/he has forgotten the subscription address. It is complicated
   slightly by the fact that ``fromline'' is the crude line and needs to be
   rfc822 parsed. Again, use of a standard interface is encouraged to make it
   compatible also with other SQL servers.

The aim of all this is to make it easy to use ezmlm to run very large lists,
easy to set up sites that handle subscriber interaction, archive access, etc,
and hopefully easier to integrate many ezmlm as done by some WWW sites today.

STANDARD EZMLM DATABASE INTERFACE

The sub-std.c file comprises the ezmlm subscriber database interface. issub and
subscribe are backwards compatible with ezmlm-0.53 in terms of function,
but use a case-insensitive hash for address storage, and take a few
extra args to be plug-in compatible with alternative interfaces. Logging to
the subscriber log (DIR/Log) is done from subscribe rather than externally,
and the subscriber from line is logged as well.

searchlog allows access to DIR/Log.

putsubs does all output of subscriber addresses, both to qmail for subscibers,
moderator sendouts, etc, and to list subscribers in reply to the -list command.

Alternative subscriber db interface routines should perform equivalent services,
and fall back to these services if the alternative interface is not configured,
e.g. if DIR/subdb doe not exist.