Removed PKG_DESTDIR_SUPPORT=destdir, no longer needed since 2011.
2.1.20 (31-Mar-2015)
- A path traversal vulnerability has been discovered and fixed (CVE-2015-2775)
- There is a new Address Change sub-section in the web admin
Membership Management section
- The Russian translation has been updated by Danil Smirnov.
- The Polish translation has been updated by Stefan Plewako.
- A LookupError in SpamDetect on a message with RFC 2047 encoded headers
in an unknown character set is fixed.
- 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.
- 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.
- 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.
- Added "If you are a list member" qualification to some messages from the
options login page.
- 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)
- 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.
- 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
<HEAD> 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.
- 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.
- 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.
- 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.
- 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.
- 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.
- 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.
- 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.
- The Russian templates have been updated by Danil Smirnov.
- The Japanese translation has been updated by SATOH Fumiyasu.
- A minor change in the French translation of a listinfo subscribe form
message has been made.
- 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.
- 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.
- 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.
- 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.
- The Subject: of the list welcome message wasn't always in the user's
preferred language. Fixed.
- Accept email command in Subject: prefixed with Re: or similar with no
intervening space.
- Fixed a UnicodeDecodeError that could occur in the web admin interface
if 'text' valued attributes have unicode values.
- We now catch the NotAMemberError exception thrown if an authenticated
unsubscribe is submitted from the user options page for a nonmember.
- Fixed an archiving bug that would cause messages with 'Subject: Re:'
only to be indexed in the archives without a link to the message.
- The vette log entry for a message discarded by a handler now includes
the list name and the name of the handler.
- The options CGI now rejects all but HTTP GET and POST requests.
- A list's poster password will now be accepted on an Urgent: header.
- Fixed a bug which caused a setting of 2 for REMOVE_DKIM_HEADERS to be
ignored.
- Renamed messages/sr/readme.sr to README.sr.
- 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.
- <label> tags have been added around most check boxes and radio buttons
and their text labels in the admin and admindb web GUI so they can be
(de)selected by clicking the text.
- If checking DNS for dmarc_moderation_action and DNS lookup is not
available, log it.
- Handle missing From: header addresses for DMARC mitigation actions.
0.9.10:
* Do not dereference null pointer on learning.
* Fix some extreme cases in BAYES.
* Add a workaround to avoid bad HTML messages breaking.
* Build with -O2 flags by default.
* Add constraints to limit DNS requests count per task.
* Add workaround for SURBL DNS flood.
* Set error if rspamd cannot learn anything.
0.9.9:
* Don't use RWL_SPAMHAUS_WL (unknown result) for whitelisting (by @fatalbanana)
* Import updated public suffix list (by @fatalbanana)
* Remove debug message
* Fix settings (by @fatalbanana)
* Remove duplicated symbol registration
* Use WAL for fuzzy storage
* RBL fixes (by @fatalbanana):
- silence errors;
- yield unknown results from RBLs;
- fix scoring for DNSWL;
- fix use of RBL name as symbol;
- ignore RBL names that would not be yielded;
* Support captures in regular expressions
* Add captures support to lua_regexp
* Support dist on FreeBSD and Darwin
* Add RCVD_IN_DNSWL_NONE as whitelisting exclusion (by @fatalbanana)
* Multiple fixes to URL detection:
- support port definition;
- fix query and path recognition;
- fix parsing of multiple slashes in URL;
- fix parsing query just after port;
- fix path field in `url:to_table` method;
- improve support of IP based URLs.
* Set ignore_whitelists = true for RECEIVED_SPAMHAUS_XBL (by @fatalbanana)
* Add GTUBE support
* Ignore User header in SA mode
0.9.8:
* Fix critical bug in bayes classifier (#305)
* Fix critical bug in RBL module (by @fatalbanana)
* Fix and rework settings plugin.
* Fix get_all_opts for a case of non-iterable options.
* Use tld for redirector's matching.
0.9.7:
* Add whitelist_exception setting to RBL module (by @fatalbanana)
* Don't use RWL_MAILSPIKE_POSSIBLE or DNSWL_BLOCKED for whitelisting (by
@fatalbanana)
* Fix extreme cases in bayes classifier.
* Fix parsing of urls with '?' at the end of hostname.
* Update interface.
* Fix number of issues with webui interaction.
* Fix saving maps.
* Allow user@ and @domain matches in multimap.
* Fix issues with bounces From processing.
* Fix abs/fabs misuse.
* Fix builds on suse and arch linux distributions.
0.9.6:
* Fix memory leak if mime cannot be parsed.
* Fix dkim cache expiration.
* Fix issues with redirector HTTP response.
* Fix abnormal connection closing with certains messages with a high score
(issue #296)
* Fix redirector installation.
* Use specific POE loop for some systems.
* Fix number of issues in URL redirector.
* Fix selecting URLs for sending to redirector.
Changelog:
Changed Hardware acceleration is now disabled by default to avoid crashing Thunderbird
Fixed A few bugs have been fixed to avoid crashing Thunderbird
Fixed in Thunderbird 38.2
2015-90 Vulnerabilities found through code inspection
2015-88 Heap overflow in gdk-pixbuf when scaling bitmap images
2015-85 Out-of-bounds write with Updater and malicious MAR file
2015-84 Arbitrary file overwriting through Mozilla Maintenance Service with hard links
2015-79 Miscellaneous memory safety hazards (rv:40.0 / rv:38.2)
Notmuch 0.20.2 (2015-06-27)
===========================
Emacs Interface
---------------
Bug fix for marking messages read in `notmuch-tree` mode.
Notmuch 0.20.1 (2015-06-01)
===========================
Test Suite
----------
Work around apparent gdb bug on arm64
Notmuch 0.20 (2015-05-31)
=========================
Command-Line Interface
----------------------
There is a new `mimetype:` search prefix
The new `mimetype:` search prefix allows searching for the
content-type of attachments, which is now indexed. See the
`notmuch-search-terms` manual page for details.
Path to gpg is now configurable
On systems with multiple versions of gpg, you can tell
notmuch which one to use by setting `crypto.gpg_path`
Emacs
-----
Avoid rendering large text attachements.
Improved rendering of CID references in HTML.
Vim
---
Vim client now respects excluded tags.
Notmuch-Mutt
------------
Support messages without Message-IDs.
Library
-------
Undeprecate single message mboxes
It seems more trouble to remove this feature than expected, so
`notmuch new` will no longer nag about mboxes with a single message.
New error logging facility
Clients should call `notmuch_database_status_string` to retrieve
output formerly printed to stderr.
Several bug fixes related to stale iterators
New status code returning API for n_query_search_{messages,thread}
Fix for library `install_name` on Mac OS X
Fix for rounding of seconds
Documentation
-------------
Sphinx is now mandatory to build docs
Support for using rst2man in place of sphinx to build the
docmumentation has been removed.
Improved notmuch-search-terms.7
The man page `notmuch-search-terms(7)` has been extended, merging
some material from the relicensed wiki.
Contrib
-------
`notmuch-deliver` is removed. As far as we know, all functionality
previously provided by `notmuch-deliver` should now be provided by
`notmuch insert`, provided by the main notmuch binary.
nmbug-status
------------
`nmbug-status` now only matches local branches when reading
`status-config.json` from the `config` branch of the `NMBGIT`
repository. To help new users running `nmbug-status`, `nmbug clone`
now creates a local `config` branch tracking `origin/config`. Folks
who use `nmbug-status` with an in-Git config (i.e. you don't use the
`--config` option) who already have `NMBGIT` set up are encouraged to
run:
git checkout config origin/config
in their `NMBGIT` repository (usually `~/.nmbug`).
limitations like NetBSD.
o Explicity disable readline support that - at least - on some Darwin
systems was automatically detected.
o Bump PKGREVISION (the readline change can affect also platforms where readline
was automagically recognised and used)
Thanks to sevan@ for testing the patches.
Upstream changes:
1.012 2015-03-28 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* CPAN RT #100839: Malformed JSON in META.json.
1.011.3 2014-12-10 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* Use "eval 'require ...'" to import optional encoding modules.
Overridden $SIG{__DIE__} can pass "eval 'use ...'".
* Replace META.yml with META.json.
1.011.2 2014-11-27 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* Added module CPAN to configure_require for inc::Test::AutoInstall.
* Added POD2::Base to recommended feature.
* Fix: wrong prototype for _resolve_S().
* Move repository to github.
1.011.1 2013-10-07 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* Fix: "shift_jis-2004" did not depend on Encode::JISX0213 but
Encode::ShiftJIS2004.
* Chg: Added gb18030 to extended mapping of GB2312.
* Updated inc::Module::Install to version 1.01.
1.011.0 2013-09-26 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* Use Encode::JISX0213 instead of Encode::JIS2K which was not
fully implemented.
* Added support for iso-2022-jp-2 and JIS X 0213 encodings.
* Reformat Changes file to suit to CPAN::Changes::Spec.
1.010.1 2013-08-25 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* No new features.
* CPAN RT #86917: Invalid META.
* Move Japanese documentation under POD2::JA.
1.010 2013-04-09 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* Fix: UTF-16 & UTF-32 encoders are died by the texts without BOM.
* Added a test #6 and some cases to test #4.
* Changed layout of distribution.
1.009.3 2012-12-30 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* No new features.
* CPAN RT #77715: Spelling mistake in documentation.
* Bug on CPAN site: cannot render PODs including "=item" with
non-ASCII.
1.009.2 2012-06-02 Hatuka*nezumi - IKEDA Soji <hatuka(at)nezumi.nu>
* Fix: detect_7bit_charset(): if charset was unknown, returns undef
instead of "US-ASCII".
* COPYING: Updated the address of FSF.
Upstream changes:
1.907 2015-02-03 17:48:10-05:00 America/New_York
- restore 5.8.x compatibility
1.906 2015-02-03 16:48:11-05:00 America/New_York
- the test suite tests parses more directly, which eliminates a number
of misleading (and possibly wrong) test cases
- escaped characters in a quoted string are now properly unescaped
for the ->phrase method, which means things should round trip better
from their decomposed form
Add LICENSE
Upstream changes:
Version 0.8.21
- fixed the case in the header names (Content-type -> Content-Type)
- fixed the encoding of non-ascii stuff in headers (RFC 2047)
- fixed using the ctype even for body with MailFile
Version 0.8.23
- fixed problems with long subjects containing accentuated characters
- fixed problems with IO::Socket::SSL
- trims addresses passed in an array (per request)
- changed licence per request from Fedora
- fix to allow very long passwords
- fix error in TestServer()
Update DEPENDS
Comment out CONFLICTS and SUPERSEDES lines since p5-Email-MIME-Creator and p5-Email-MIME-Modifier had been removed from pkgsrc 5 years ago.
Upstream changes:
1.934 2015-08-01 20:33:57-04:00 America/New_York
- fix debug_structure for multipart mail with only one subpart
1.933 2015-07-25 09:30:56-04:00 America/New_York
- correct Email::Simple prereq
- skip omitting content-type attributes with undef value
1.932 2015-07-24 22:24:06-04:00 America/New_York
- stable release of previous trial releases
- see v1.930 and v1.931, below
1.931 2015-07-12 17:31:01-04:00 America/New_York (TRIAL RELEASE)
- When doing parts_set to become single-part, do not lose the charset
and set the mail to us-ascii.
This is almost certainly a half measure, and this whole hunk of code
should be improved.
1.930 2015-03-25 22:58:42-04:00 America/New_York (TRIAL RELEASE)
- rely on the new header_raw methods in new Email::Simple
1.929 2015-02-17 09:31:35-05:00 America/New_York
- try not to lose unrecognized content-type parameters
- avoid copying the body content when not needed during construction
1.928 2014-12-15 21:47:49-05:00 America/New_York
- when converting from single part to multipart, drop
Content-Transfer-Encoding and Content-Type's charset property
- update subparts in walk_parts if stringification changes
1.927 2014-12-04 10:19:58-05:00 America/New_York
- add header_str_pairs; like header_pairs, but decoded
- remove Email:MIME::Creator from synopsis
1.926 2014-01-28 23:28:41-05:00 America/New_York
- improve EOL processing with quoted-printable (thanks, Tomohiro
Hosaka!)
- avoid uninitialized warnings [rt.cpan.org #91707]
1.925 2013-11-08 07:01:03 America/New_York
- improve reliability of mail reconstruction during walk_parts (thanks,
Tomohiro Hosaka!)
Upstream changes:
2.208 2015-07-20 20:34:04-04:00 America/New_York
- same as v2.207
- ...but bump perl required to v5.8
2.207 2015-07-04 14:23:36-04:00 America/New_York (TRIAL RELEASE)
- fold headers passed to header arg in create method
https://github.com/rjbs/Email-Simple/issues/7
2.206 2015-03-26 23:30:51-04:00 America/New_York
- the changes from 2.204 are back (and re-listed here), but now with
the added benefit of not breaking Email::MIME; see the git repository
for the sordid details
- provide header_raw and header_raw_set for consistent behavior between
Email::Simple and Email::MIME; also, header_raw_pairs
2.205 2015-03-26 14:54:45-04:00 America/New_York
- undo changes from 2.204 until safe Email::MIME upgrade path is
available
2.204 2015-03-25 22:50:49-04:00 America/New_York
- provide header_raw and header_raw_set for consistent behavior between
Email::Simple and Email::MIME; also, header_raw_pairs
IMAPFilter 2.6.2 - 30 Jun 2015
- Bug fix; drop connection on local certificate mismatch.
IMAPFilter 2.6.1 - 29 Jun 2015
- Support for loading the system's CA certificates from a file.
- Bug fix; flusing of standard output/error in a some cases.
IMAPFilter 2.6 - 16 Jun 2015
- Optimizations that vastly improve performance of the meta-searching
functionality, by making subsequent search requests limit their scope based
on the results already returned by previous requests (previously this was
the case only for the match_*() methods).
- Lua 5.3 compatibility (the codebase can still be compiled with versions 5.2
and 5.1).
- A new "limit" option can be used as a work-around for problems that some
servers have with long requests.
- The documentation now clarifies how to make the client block indefinitely.
- Bug fix; the check_status() method now returns 4 numbers on error as
described in the documentation (just negative values in this case).
- Bug fix; flushing of default output on password prompt.
- Bug fix; program fault when very long requests were to be sent.
IMAPFilter 2.5.7 - 16 Nov 2014
- Support for automatic validation of SSL connections, using the operating
system's installed trusted CA certificates (usually from authorities that
are shipped with web browsers).
- A new "wakeonany" option makes it possible to end IDLE mode on any event,
not just on new messages.
- A new "reenter" option controls whether IDLE mode should be re-entered,
after the recovery of the session, which makes it possible to continue with
the execution of the next line in the configuration file.
- The documentation now mentions that the contain_*() methods do
case-insensitive searching (an IMAP limitation).
- The documentation now has a better explanation of the arrived_*() and
sent_*() methods.
- Bug fix; the value returned by the copy method was incorrect in some cases.
With all supported Postfix releases, the default settings have been
updated so that they no longer enable export-grade ciphers, and no
longer enable the SSLv2 and SSLv3 protocols. These ciphers and
protocols have little if any legitimate use today, and have instead
become a vehicle for downgrade attacks. There are no other code
changes.
Postfix documentation has been updated to reflect the new default
settings and their rationale; the RELEASE_NOTES give suggestions
for how to enable the old ciphers and protocols if your infrastructure
requires them.
Finally, abandoning deprecated ciphers and protocols does not really
improve TLS security without measures to better authenticate remote
servers. Secure DNS and TLSA are steps in that direction.
* Fix mozilla-common.mk inclusion.
Changelog:
What's New
Fixed Copy/Paste into plain text editor deletes newlines from quoted text (bug 1143570)
Fixed Cross-posts won't send because Newsgroups: groups are separated with comma+space, not just comma (bug 1151448)
Fixed Cannot send email through exchange server (NTLM) (bug 1174159)
Fixed Doesn't display GB2312 encoded texts correctly for Chinese Characters (bug 1174580)
Fixed OAuth2 authentication for GMail does not work when specified server is imap.gmail.com or smtp.gmail.com. (bug 1176773)
Known Issues
unresolved Import from Outlook and Eudora disabled, code currently not working (bug 1175055)
