-*- coding: iso-8859-1 -*- Mailman - The GNU Mailing List Management System Copyright (C) 1998-2016 by the Free Software Foundation, Inc. 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA Here is a history of user visible changes to Mailman. 2.1.25 (xx-xxx-xxxx) Bug fixes and other patches - Defended against certain web attacks that cause exceptions and "we hit a bug" responses when POST data or query fragments contain multiple values for the same parameter. (LP: #1695667) - The fix for LP: #1614841 caused a regression in the options CGI. This has been fixed. (LP: #1602608) - Added a -a option to the (e)grep commands in contrib/mmdsr to account for logs that may have non-ascii and be seen as binary. - Fixed the -V option to bin/list_lists to not show lists whose host is a subdomain of the given domain. (LP: #1695610) 2.1.24 (02-Jun-2017) Security - A most likely unexploitable XSS attach that relies on the Mailman web server passing a crafted Host: header to the CGI environment has been fixed. Apache for one is not vulnerable. Thanks to Alqnas Eslam. New Features - There is a new RCPT_BASE64_HEADER_NAME setting. If this is set to a non-empty string, that string is the name of a header that will be added to personalized and VERPed deliveries with value equal to the base64 encoding of the recipient's email address. This is intended to enable identification of the recipient otherwise redacted from "spam report" feedback loop messages. - cron/senddigests has a new -e/--exceptlist option to send pending digests for all but a named list. (LP: #1619770) - The values for DEFAULT_DIGEST_FOOTER and DEFAULT_MSG_FOOTER have been changed to use a standard signature separator for DEFAULT_MSG_FOOTER and to remove the unneded line of underscores from DEFAULT_DIGEST_FOOTER. (LP: #266269) i18n - The Polish html templates have been recoded to use html entities instead of non-ascii characters. - The Basque (Euskara) translation has been updated by Gari Araolaza. - The German "details for personalize" page has been updated by Christian F Buser. - The Japanese translation has been updated by Yasuhito FUTATSUKI. Bug fixes and other patches - The list-owner@virtual.domain addresses are now added to virtual-mailman as they are exposed in 'list created' emails. (LP: 1694384) - The 'list run by' addresses in web page footers are now just the list-owner address. (LP: #1694384) - Changed member_verbosity_threshold from a >= test to a strictly > test to avoid the issue of moderating every post when the threshold = 1. (LP: #1693366) - Subject prefixing has been improved to always have a space between the prefix and the subject even with non-ascii in the prefix. This will sometimes result in two spaces when the prefix is non-ascii but the subject is ascii, but this is the lesser evil. (LP: #1525954) - Treat message and digest headers and footers as empty if they contain only whitespace. (LP: #1673307) - Ensured that added message and digest headers and footers always have a terminating new-line. (LP: #1670033) - Fixed an uncaught TypeError in the subscribe CGI. (LP: #1667215) - Added recognition for a newly seen mailEnable bounce. - Fixed an uncaught NotAMemberError when a member is removed before a probe bounce for the member is returned. (LP: #1664729) - Fixed a TypeError thrown in the roster CGI when called with a listname containing a % character. (LP: #1661810) - Fixed a NameError issue in bin/add_members with DISABLE_COMMAND_LOCALE_CSET = yes. (LP: #1647450) - The CleanseDKIM handler has been removed from OWNER_PIPELINE. It isn't needed there and has adverse DMARC implications for messages to -owner of an anonymous list. (LP: #1645901) - Fixed an issue with properly RFC 2047 encoding the display name in the From: header for messages with DMARC mitigations. (LP: #1643210) - Fixed an issue causing UnicodeError in sending digests following a change of a list's preferred_language. (LP: #1644356) - Enhanced the fix for race conditions in MailList().Load(). (LP: #266464) - Fixed a typo in Utils.py that could have resulted in a NameError in logging an unlikely occurrence. (LP: #1637745) - Fixed a bug which created incorrect "view more members" links at the bottom of the admin Membership List pages. (LP: #1637061) - The 2.1.23 fix for LP: #1604544 only fixed the letter links at the top of the Membership List. The links at the bottom have now been fixed. - paths.py now adds dist-packages as well as site-packages to sys.path. (LP: #1621172) - INIT INFO has been added to the sample init.d script. (LP: #1620121) 2.1.23 (27-Aug-2016) Security - CSRF protection has been extended to the user options page. This was actually fixed by Tokio Kikuchi as part of the fix for LP: #775294 and intended for Mailman 2.1.15, but that fix wasn't completely merged at the time. The full fix also addresses the admindb, and edithtml pages as well as the user options page and the previously fixed admin pages. Thanks to Nishant Agarwala for reporting the issue. CVE-2016-6893 (LP: #1614841) New Features - For header_filter_rules matching, RFC 2047 encoded headers, non-encoded headers and header_filter_rules patterns are now all decoded to unicode. Both XML character references of the form &#nnnn; and unicode escapes of the form \Uxxxx in patterns are converted to unicodes as well. Both headers and patterns are normalized to 'NFKC' normal form before matching, but the normalization form can be set via a new NORMALIZE_FORM mm_cfg setting. Also, the web UI has been updated to encode characters in text fields that are invalid in the character set of the page's language as XML character references instead of '?'. This should help with entering header_filter_rules patterns to match 'odd' characters. This feature is experimental and is problematic for some cases where it is desired to have a header_filter_rules pattern with characters not in the character set of the list's preferred language. For patterns without such characters, the only change in behavior should be because of unicode normalization which should improve matching. For other situations such as trying to match a Subject: with CJK characters (range U+4E00..U+9FFF) on an English language (ascii) list, one can enter a pattern like '^subject:.*[一-鿿]' or '^subject:.*[\u4e00;-\u9fff;]' to match a Subject with any character in the range, and it will work, but depending on the actual characters and the browser, submitting another, even unrelated change can garble the original entry although this usually occurs only with ascii pages and characters in the range \u0080-\u00ff. The \Uxxxx unicode escapes must have exactly 4 hex digits, but they are case insensitive. (LP: #558155) - Thanks to Jim Popovitch REMOVE_DKIM_HEADERS can now be set to 3 to preserve the original headers as X-Mailman-Original-... before removing them. - Several additional templates have been added to those that can be edited via the web admin GUI. (LP: #1583387) - SMTPDirect.py can now do SASL authentication and STARTTLS security when connecting to the outgoiung MTA. Associated with this are new Defaults.py/mm_cfg.py settings SMTP_AUTH, SMTP_USER, SMTP_PASSWD and SMTP_USE_TLS. (LP: #558281) - There is a new Defaults.py/mm_cfg.py setting SMTPLIB_DEBUG_LEVEL which can be set to 1 to enable verbose smtplib debugging to Mailman's error log to help with debugging 'low level smtp failures'. (LP: #1573074) - A list's nonmember_rejection_notice attribute will now be the default rejection reason for a held non-member post in addition to it's prior role as the reson for an automatically rejected non-member post. (LP: #1572330) i18n - The French translation of 'Dutch' is changed from 'Hollandais' to 'Néerlandais' per Francis Jorissen. - Some German language templates that were incorrectly utf-8 encoded have been recoded as iso-8859-1. (LP: #1602779) - Japanese translation and documentation in messages/ja has been updated by Yasuhito FUTATSUKI. Bug fixes and other patches - The admin Membership List letter links could be incorrectly rendered as Unicode strings following a search. (LP: #1604544) - We no longer throw an uncaught TypeError with certain defective crafted POST requests to Mailman's CGIs. (LP: #1602608) - Scrubber links in archives are now in the list's preferred_language rather than the poster's language. (LP: #1586505) - Improved logging of banned subscription and address change attempts. (LP: #1582856) - In rare circumstances a list can be removed while the admin or listinfo CGI or bin/list_lists is running causing an uncaught MMUnknownListError to be thrown. The exception is now caught and handled. (LP: #1582532) - Set the Date: header in the wrapper message when from_is_list or dmarc_moderation_action is Wrap Message. (LP: #1581215) - A site can now set DMARC_ORGANIZATIONAL_DOMAIN_DATA_URL to None or the null string if it wants to avoid using this. (LP: #1578450) - The white space to the left of the admindb Logout link is no longer part of the link. (LP: #1573623) 2.1.22 (17-Apr-2016) i18n - Fixed a typo in the German options.html template. (LP: #1562408) - An error in the Brazilian Portugese translation of Quarterly has been fixed thanks to Kleber A. Benatti. - The Brazilian Portugese translation has been updated by Emerson Ribeiro de Mello. Bug fixes and other patches - All addresses in data/virtual-mailman are now properly appended with VIRTUAL_MAILMAN_LOCAL_DOMAIN and duplicates are not generated if the site list is in a virtual domain. (LP: #1570630) - DMARC mitigations will now find the From: domain to the right of the rightmost '@' rather than the leftmost '@'. (LP: #1568445) - DMARC mitigations for a sub-domain of an organizational domain will now use the organizational domain's sp= policy if any. (LP: #1568398) - Modified NewsRunner.py to ensure that messages gated to Usenet have a non-blank Subject: header and when munging the Message-ID to add the original to References: to help with threading. (LP: #557955) - Fixed the pipermail archiver to do a better job of figuring the date of a post when its Date: header is missing, unparseable or has an obviously out of range date. This should only affect bin/arch as ArchRunner has code to fix dates at least if ARCHIVER_CLOBBER_DATE_POLICY has not been set to 0 in mm_cfg.py. If posts have been added in the past to a list's archive using bin/arch and an imported mbox, running bin/arch again could result is some of those posts being archived with a different date. (LP: #1555798) - Fixed an issue with CommandRunner shunting a malformed message with a null byte in the body. (LP: #1553888) - Don't collapse multipart with a single sub-part inside multipart/signed parts. (LP: #1551075) 2.1.21 (28-Feb-2016) New Features - There is a new dmarc_none_moderation_action list setting and a DEFAULT_DMARC_NONE_MODERATION_ACTION mm_cfg.py setting to optionally apply Munge From or Wrap Message actions to posts From: domains that publish DMARC p=none. The intent is to eliminate failure reports to the domain owner for messages that would be munged or wrapped if the domain published a stronger DMARC policy. See the descriptions in Defaults.py, the web UI and the bug report for more. (LP: #1539384) - Thanks to Jim Popovitch there is now a feature to automatically turn on moderation for a malicious list member who attempts to flood a list with spam. See the details for the Privacy options ... -> Sender filters -> member_verbosity_threshold and member_verbosity_interval settings in the web admin UI and the documentation in Defaults.py for the DEFAULT_MEMBER_VERBOSITY_* and VERBOSE_CLEAN_LIMIT settings for information. - bin/list_members now has options to display all moderated or all non-moderated members. - There is now a mm_cfg.py setting GLOBAL_BAN_LIST which is like the individual list's ban_list but applies globally to all subscribe requests. See the description in Defaults.py for more details. i18n - The Japanese translation has been updated by Yasuhito FUTATSUKI. - Also thanks to Miloslav Trmac and Yasuhito FUTATSUKI, the l10n for Mailman's bin/ commands has been fixed to display using the character set of the user's work station even when Mailman's character set for the language is different. Because this has not been tested over a wide set of locales, there is an mm_cfg.py switch DISABLE_COMMAND_LOCALE_CSET to disable it if it causes problems. (LP: #558167) - The Polish translation has been updated by Stefan Plewako. - The German translation has been updated by Mirian Margiani and Bernhard Schmidt. - The Russian translation has been updated by Danil Smirnov. - Several Galician templates that were improperly encoded as iso-8859-1 have been fixed. (LP: #1532504) - The Brazilian Portugese translation has been updated by Emerson Ribeiro de Mello. Bug fixes and other patches - If DMARC lookup fails to find a policy, also try the Organizational Domain. Associated with this is a new mm_cfg.py setting DMARC_ORGANIZATIONAL_DOMAIN_DATA_URL which sets the URL used to retrieve the data for the algorithm that computes the Organizational Domain. See https://publicsuffix.org/list/ for info. (LP: #1549420) - Modified contrib/mmdsr to correctly report No such list names that contain ". - User's "Acknowledge" option will now be honored for posts to anonymous lists. (LP: #1546679) - Fixed a typo in the Non-digest options regular_exclude_ignore description thanks to Yasuhito FUTATSUKI. - DEFAULT_PASS_MIME_TYPES has been changed to accept text/plain sub-parts from message/rfc822 parts and multipart parts other than mixed and alternative and also accept pgp signatures. This only applies to newly created lists and other than pgp signatures, still only accepts text/plain. (LP: #1517446) - Modified contrib/mmdsr to report held and banned subscriptions and DMARC lookups in their own categories. - Fixed a bug that could create a garbled From: header with certain DMARC mitigation actions. (LP: #1536816) - Treat a poster's address which matches an equivalent_domains address as a list member for the regular_exclude_ignore check. (LP: #1526550) - Fixed an issue that sometimes left no white space following subject_prefix. (LP: #1525954) - Vette log entries for banned subscriptions now include the source of the request if available. (LP: #1525733) - Submitting the user options form for a user who was asynchronously unsubscribed would throw an uncaught NotAMemberError. (LP: #1523273) - It was possible under some circumstances for a message to be shunted after a handler rejected or discarded it, and the handler would be skipped upon unshunting and the message accepted. (LP: #1519062) - Posts gated to usenet will no longer have other than the target group in the Newsgroups: header. (LP: #1512866) - Invalid regexps in *_these_nonmembers, subscribe_auto_approval and ban_list are now logged. (LP: #1507241) - Refactored the GetPattern list method to simplify extending @listname syntax to new attributes in the future. Changed Moderate.py to use the GetPattern method to process the *_these_nonmembers lists. - Changed CookHeaders to default to using space rather than tab as continuation_ws when folding headers. (LP: #1505878) - Fixed the 'pidfile' path in the sample init.d script. (LP: #1503422) - Subject prefixing could fail to collapse multiple 'Re:' in an incomming message if they all came after the list's subject_prefix. This is now fixed. (LP: #1496620) - Defended against a user submitting URLs with query fragments or POST data containing multiple occurrences of the same variable. (LP: #1496632) - Fixed bin/mailmanctl to check its effective rather than real uid. (LP: #1491187) - Fixed cron/gate_news to catch EOFError on opening the newsgroup. (LP: #1486263) - Fixed a bug where a delayed probe bounce can throw an AttributeError. (LP: #1482940) - If a list is not digestable an the user is not currently set to receive digests, the digest options will not be shown on the user's options page. (LP: #1476298) - Improved identification of remote clients for logging and subscribe form checking in cases where access is via a proxy server. Thanks to Jim Popovitch. Also updated contrib/mmdsr for log change. - Fixed an issue with shunted messages on a list where the charset for the list's preferred_language had been changed from iso-8859-1 to utf-8 without recoding the list's description. (LP: #1462755) - Mailman-Postfix integration will now add mailman@domain entries in data/virtual-mailman for each domain in POSTFIX_STYLE_VIRTUAL_DOMAINS which is a host_name of a list. This is so the addresses which are exposed on admin and listinfo overview pages of virtual domains will be deliverable. (LP: #1459236) - The vette log entry for DMARC policy hits now contains the list name. (LP: #1450826) - If SUBSCRIBE_FORM_SECRET is enabled and a user's network has a load balancer or similar in use the POSTing IP might not exactly match the GETting IP. This is now accounted for by not requiring the last octet (16 bits for ipV6) to match. (LP: #1447445) - DKIM-Signature:, DomainKey-Signature: and Authentication-Results: headers are now removed by default from posts to anonymous lists. (LP: #1444673) - The list admin web UI Mambership List search function often doesn't return correct results for search strings (regexps) that contain non-ascii characters. This is partially fixed. (LP: #1442298) 2.1.20 (31-Mar-2015) Security - A path traversal vulnerability has been discovered and fixed. This vulnerability is only exploitable by a local user on a Mailman server where the suggested Exim transport, the Postfix postfix_to_mailman.py transport or some other programmatic MTA delivery not using aliases is employed. CVE-2015-2775 (LP: #1437145) New Features - There is a new Address Change sub-section in the web admin Membership Management section to allow a list admin to change a list member's address in one step rather than adding the new address, copying settings and deleting the old address. (LP: #266809) i18n - The Russian translation has been updated by Danil Smirnov. - The Polish translation has been updated by Stefan Plewako. Bug fixes and other patches - A LookupError in SpamDetect on a message with RFC 2047 encoded headers in an unknown character set is fixed. (LP: #1427389) - Fixed a bug in CommandRunner that could process the second word of a body line as a command word and a case sensitivity in commands in Subject: with an Re: prefix. (LP: #1426829) - Fixed a bug in CommandRunner that threw an uncaught KeyError if the input to the list-request address contained a command word terminated by a period. (LP: #1426825) 2.2 Branch Backports (released in conjunction with 2.1.19) The following New Features and Bug Fixes have been in an "unofficial, never to be released" Mailman 2.2 branch for several years. Until now, they were never implemented on the official 2.1 branch because of their i18n impacts. Given that there have been a number of i18n impacting changes due to DMARC mitigations in the last few releases, it has been decided to backport these as well. All of these changes have been running in production on several lists for years without problems other than untranslated strings, so they should be reasonably "bug free". New Features - There is a new list attribute 'subscribe_auto_approval' which is a list of email addresses and regular expressions matching email addresses whose subscriptions are exempt from admin approval. (LP: #266609) - Confirmed member change of address is logged in the 'subscribe' log, and if admin_notify_mchanges is true, a notice is sent to the list owner using a new adminaddrchgack.txt template. - Added an 'automate' option to bin/newlist to send the notice to the admin without the prompt. - The processing of Topics regular expressions has changed. Previously the Topics regexp was compiled in verbose mode but not documented as such which caused some confusion. Also, the documentation indicated that topic keywords could be entered one per line, but these entries were not handled properly. Topics regexps are now compiled in non-verbose mode and multi-line entries are 'ored'. Existing Topics regexps will be converted when the list is updated so they will continue to work. - Added real name display to the web roster. (LP: #266754) Bug fixes and other patches - Changed the response to an invalid confirmation to be more generic. Not all confirmations are subscription requests. - Changed the default nonmember_rejection_notice to be more user friendly. (LP: #418728) - Added "If you are a list member" qualification to some messages from the options login page. (LP: #266442) - Changed the 'Approve' wording in the admindbdetails.html template to 'Accept/Approve' for better agreement with the button labels. - Added '(by thread)' to the previous and next message links in the archive to emphasize that even if you got to the message from a subject, date or author index, previous and next are still by thread. 2.1.19 (28-Feb-2015) New Features - The subscribe_auto_approval feature backported from the 2.2 branch and described above has been enhanced to accept entries of the form @listname to auto approve members of another list. (LP: #1417093) - There is a new list attribute dmarc_wrapped_message_text and a DEFAULT_DMARC_WRAPPED_MESSAGE_TEXT setting to set the default for new lists. This text is added to a message which is wrapped because of dmarc_moderation_action in a separate text/plain part that precedes the message/rfc822 part containing the original message. It can be used to provide an explanation of why the message was wrapped or similar info. - There is a new list attribute equivalent_domains and a DEFAULT_EQUIVALENT_DOMAINS setting to set the default for new lists which in turn defaults to the empty string. This provides a way to specify one or more groups of domains, e.g., mac.com, me.com, icloud.com, which are considered equivalent for validating list membership for posting and moderation purposes. - There is a new WEB_HEAD_ADD setting to specify text to be added to the section of Mailman's internally generated web pages. This doesn't apply to pages built from templates, but in those cases, custom templates can be created. (LP: #1409396) - There is a new DEFAULT_SUBSCRIBE_OR_INVITE setting. Set this to Yes to make the default selection on the admin Mass Subscriptions page Invite rather than Subscribe. (LP: #1404511) - There is a new list attribute in the Bounce processing section. bounce_notify_owner_on_bounce_increment if set to Yes will cause Mailman to notify the list owner on every bounce that increments a list member's score but doesn't result in a probe or disable. There is a new configuration setting setting DEFAULT_BOUNCE_NOTIFY_OWNER_ON_BOUNCE_INCREMENT to set the default for new lists. This in turn defaults to No. (LP: #1382150) Changed behavior - Mailman's log files, request.pck files and heldmsg-* files are no longer created world readable to protect against access by untrusted local users. Note that permissions on existing log files won't be changed so if you are concerned about this and don't rotate logs or have a logrotate process that creates new log files instead of letting Mailman create them, you will need to address that. (LP: #1327404) Other changes - The Python Powered logo image has been replaced in the misc/ directory in the source distribution. Depending on how you've installed these images, you may need to copy PythonPowered.png from the misc/ directory in the source or from the $prefix/icons/ installed directory to another location for your web server. (LP: #1408575) i18n - The Polish translation has been updated by Stefan Plewako. - The Interlingua translation has been updated by Martijn Dekker. - The Japanese message catalog has been updated by SATOH Fumiyasu. - Mailman's character set for Romanian has been changed from iso-8859-2 to utf-8 and the templates and messages recoded. This change will require running 'bin/arch --wipe' on any existing Romanian language lists in order to recode the list's archives, and will require recoding any edited templates in lists/LISTNAME/ro/*, templates/DOMAIN/ro/* and templates/site/ro/*. It may also require recoding any existing iso-8859-2 text in list attributes. (LP: #1418735) - Mailman's character set for Russian has been changed from koi8-r to utf-8 and the templates and messages recoded. This change will require running 'bin/arch --wipe' on any existing Russian language lists in order to recode the list's archives, and will require recoding any edited templates in lists/LISTNAME/ru/*, templates/DOMAIN/ru/* and templates/site/ru/*. It may also require recoding any existing koi8-r text in list attributes. (LP: #1418448) - Mailman's versions.py has been augmented to help with the above two character set changes. The first time a list with preferred_language of Romanian or Russian is accessed or upon upgrade to this release, any list attributes which have string values such as description, info, welcome_msg, etc. that appear to be in the old character set will be converted to utf-8. This is done recursively for the values (but not the keys) of dictionary attributes and the elements of list and tuple attributes. - The Russian message catalog and templates have been further updated by Danil Smirnov. - The Romanian message catalog has been updated. (LP: #1415489) - The Russian templates have been updated by Danil Smirnov. (LP: #1403462) - The Japanese translation has been updated by SATOH Fumiyasu. (LP: #1402989) - A minor change in the French translation of a listinfo subscribe form message has been made. (LP: #1331194) Bug fixes and other patches - Because of privacy concerns with the 2.2 backport adding real name to list rosters, this is controlled by a new ROSTER_DISPLAY_REALNAME setting that defaults to No. You may wish to set this to Yes in mm_cfg.py. - Organization: headers are now unconditionally removed from posts to anonymous lists. Regexps in ANONYMOUS_LIST_KEEP_HEADERS weren't kept if the regexp included the trailing ':'. This is fixed too. (LP: #1419132) - The admindb interface has been fixed so the the detail message body display doesn't lose part of a multi-byte character, and characters which are invalid in the message's charset are replaced rather than the whole body not being converted to the display charset. (LP: #1415406) - Fixed a bug in bin/rmlist that would throw an exception or just fail to remove held message files for a list with regexp special characters in its name. (LP:#1414864) - When applying DMARC mitigations, CookHeaders now adds the original From: to Cc: rather than Reply-To: in some cases to make MUA 'reply' and 'reply all' more consistent with the non-DMARC cases. (LP: #1407098) - The Subject: of the list welcome message wasn't always in the user's preferred language. Fixed. (LP: #1400988) - Accept email command in Subject: prefixed with Re: or similar with no intervening space. (LP: #1400200) - Fixed a UnicodeDecodeError that could occur in the web admin interface if 'text' valued attributes have unicode values. (LP: #1397170) - We now catch the NotAMemberError exception thrown if an authenticated unsubscribe is submitted from the user options page for a nonmember. (LP: #1390653) - Fixed an archiving bug that would cause messages with 'Subject: Re:' only to be indexed in the archives without a link to the message. (LP: #1388614) - The vette log entry for a message discarded by a handler now includes the list name and the name of the handler. (LP: #558096) - The options CGI now rejects all but HTTP GET and POST requests. (LP: #1372199) - A list's poster password will now be accepted on an Urgent: header. (LP: #1371678) - Fixed a bug which caused a setting of 2 for REMOVE_DKIM_HEADERS to be ignored. (LP: #1363278) - Renamed messages/sr/readme.sr to README.sr. (LP: #1360616) - Moved the dmarc_moderation_action checks from the Moderate handler to the SpamDetect handler so that the Reject and Discard actions will be done before the message might be held by header_filter_rules, and the Wrap Message and Munge From actions will be done on messages held by header_filter_rules if the message is approved. (LP: #1334450) -