- use CONFIGURE_ENV during the configure phase to make sure we pick up
QMAKESPEC
- set CONFIG_SHELL to ${SH} to help /bin/sh challenged platforms like
solaris.
- fix insecure file creation in /tmp, patch from silc cvs
the impact of this issue is very low. it allows an attacker to overwrite
arbitrary files owned by the user running silcd ("silcd", in pkgsrc) IFF
the owner of the process or root send SIGUSR1 signal to the process to dump
stats. the only file owned by the "silcd" user is typically the log file
which resides in a directory inaccessible by anyone except the user itself
and root so the potential attacker would need to guess its name.
http://www.zataz.net/adviso/silc-server-toolkit-06152005.txt
please note that the advisory also incorrectly states that silc-toolkit is
vulnerable too. the code in question is never compiled in the toolkit so
it's not affected.
Bump PKGREVISION.
- Updated translations (cs, el, fr, gl, it, nl, sl, sr, vi, zh_TW).
- Fixed incorrect information displayed in Plugins & scripts window
under unix (xc244-fixpluginns.diff).
- Added "/set irc_whois_front 1" option to show WHOIS in front tab.
- Lots of speed ups under the hood, mainly in handling of URL
highlighting during mouse motion. Also now allows underlining
.name and .info domains [1230265].
- Moved the "Insert color code" menu into the input box's right-
click menu.
- Fixed "Your Message" messing up when starting with a comma
[1230269].
- Added /id command to identify yourself to nickserv.
- Added /gui MSGBOX <text> for scripters.
- Added /menu command which lets plugins/scripts add their own
menu items.
- Added support for passive DCC chat via /DCC PCHAT <nick>.
- Added support for DCC sending and receiving very large files
(above 4 GB).
- Improved layout of "Info" button in the DCC windows.
- Improved layout of the nick-name right-click menu.
- Improved /help command's display of plugins/script commands.
- Fixed two bugs in detaching tabs (or CTRL-I) [1228926].
- Added /uselect command for scripters to select nick names in the
channel userlist (Daniel P. Stasinski).
- Fixed possible crashes while using the SJIS (Japanese) charset.
- Fixed various memory leaks in right-click menus.
Closes PR pkg/31291 from Jason White. Jason reports that 2.34 does
not work with jabber, and 2.36 did.
Change since 2.34:
2.38 Added #defines for older gtk to bypass gstdio
Fixed flags on outgoing key messages to avoid un-idling
Changed default auto-accept of unknown keys to false.
Fixed debug output for version warning (Daniel Atallah)
2.37 Bugfix for possible double-free of buffer
Bugfix for preference directories with unicode chars
Relaxed checks on Gaim version to allow minor version bumps
Same relaxed checks on the Win32 installer (Daniel Atallah)
Change to intl code to compile on Solaris (Jonathan Wang)
2.36 Bugfix for users who use a non-default prefs directory
Bugfix for Windows 2000
2.35 Workaround for Jabber bug in Gaim
* gdomain now has a -K flag
* gsend has the -b flag
* the default gsubrc doesn't escape '+' characters
* the time format of the default gsubrc is adjustable
* sundry WQP bugfixes
* bug fix to stop gsub from dropping messages
* MacOSX compatibility fixes
* gkinfo has a -y flag
* much more verbose diagnostics when things go wrong with keys
* partial fix to the galed spum problem (not a real fix though)
* don't be as verbose about "replaced .. with __" messages
* more mostly unrelated liboop fixes
1) Simplify the way how an emacs version is picked when no emacs
is installed, but a user try to install an Emacs Lisp package.
Just pick up the version set as EMACS_TYPE than searching for
versions already installed etc. If the EMACS_TYPE version is
not supported by the Emacs Lisp Package, just fail. EMACS_TYPE
be default to GNU Emacs 21.
(In other words, users should set EMACS_TYPE as they want.
Otherwise GNU Emacs 21 is used.)
2) All Emacs Lisp Packages *must* prepend EMACS_PKGNAME_PREFIX to
a) the PKGNAME itself, and b) PKGNAME in its dependency lines.
EMACS_PKGNAME_PREFIX is expanded to "xemacs-" when XEmacs is
used. This keeps dependency graph of Emacs-Lisp-packages-
installed-for-XEmacs consistent.
3) Document EMACS_* variables as much as possible.
4) Provide more cookies for PLIST. Maybe utilized later.
Note that the 2) change doesn't affect the default, GNU Emacs 21
behaviour. So no version / revision bumps in this commit.
by Linux) in tests, use value of BUILDLINK_LDADD.dl.
Also use PTHREAD_LIBS instead of hard-coded -lpthread.
Include dlopen.buildlink3.mk and include pthread.buildlink3.mk.
This fixes problem on Linux and DragonFly where it was not
built with silc support.
Okayed by maintainer.
I tested this on Linux, DragonFly and NetBSD 2.0.2.
I also filed a bug report with gaim last week so they can fix the
configure.ac correctly.
* This release features a large refactoring to make further development
easier.
* Part of the works has also been to make the current features more stable
and better implement.
* Most of the work on this release has been carried out by Mikael Hallendal
and Martyn Russel.
* UPDATE NOW!
Changes in 1.0:
---------------
* Fix crash when connection goes down using SSL, LM-3 (Mikael Hallendal)
* Fixed some GCC 4.0 warnings (Mikael Hallendal)
* Some API fixups for consistency in LmSLL (Mikael Hallendal)
* Don't try to setup SSL if it's not supported (Martyn Russel)
* Documentation fixes (Mikael Hallendal)
* State fixes when using blocking API, LM-7 (Mikael Hallendal)
* Check that connection is open before using it, LM-6 (Mikael Hallendal)
* Fix infinite loop in blocking open if failed, LM-8 (Mikael Hallendal)
Changes in 0.90:
----------------
* Two crashers for ssl read fixed (Micke, Daniel Lavalliere)
* Fixed a couple of leaks when freeing the connection (Daniel Lavalliere)
* A couple of thread related issues solved (Micke, Daniel Lavalliere)
* Work with latest GnuTLS (foser)
* Improved the example tools some (Micke)
- Updated translations (hi, ko, lt, pa, ru, vi, zh_TW).
- People's away message is now shown in the right-click menu, if
known (Christopher Aillon).
- The "Bind to:" setting can now be set to 0.0.0.0 [1176256].
- Plugin API: Don't crash if a print-event closes the current context
and doesn't eat the event [1175674].
- Disabled parsing of quotation marks for /JOIN, so you can join
channels with a quote in them (Dan Fruehauf).
- Fixed truncation of the URL in the right-click menu. Now handles
UTF-8 properly [1188229].
- Fixed use of CP1255 charset, which would chop the last char when
receiving messages [1122089].
- The DCC windows now allow multiple selection and the columns auto
resize (Dan Fruehauf).
- Added "CTCP Sound to Channel" event [1159445].
- You can now drag and drop files into dialog windows to start file
transfers.
- Fixed: "XChat can't ban users with long idents" (Dan Fruehauf)
[1159447].
- Implemented taskbar flashing on unix. Requires a window manager
or taskbar that supports XUrgency flag (Adil).
backslashes anymore. A single backslash is enough. Changed the
definition in all affected packages. For those that are not caught, an
additional check is placed into bsd.pkginstall.mk.
that these directories will be conditionally removed (based on reference
counts), regardless of the value of PKG_CONFIG. Bump the PKGREVISION
for packages that were modified as a result.
as the INSTALL and DEINSTALL scripts no longer distinguish between
the two types of files. Drop SUPPORT_FILES{,_PERMS} and modify the
packages in pkgsrc accordingly.
"this release of gaim has a few security fixes which mirror
the effects of patch-ae patch-af patch-ag"
ChangeLog says:
version 1.5.0 (8/11/2005):
* Ability to set IRC quit message (Lalo Martins)
* OSCAR file transfers now work for 2 users behind the same NAT
(Jonathan Clark)
* Yahoo! buddy requests to add you to their buddy list now prompt for
authorization
* Added a /clear command for conversations/chats
* Fixed ICQ encoding for messages with offline ICQ users
(Ilya Konstantinov, SF Bug #1179452)
* Default Yahoo! chat roomlist locale to 'us'
file's sole purpose was to provide a dependency on pkg-config and set
some environment variables. Instead, turn pkg-config into a "tool"
in the tools framework, where the pkg-config wrapper automatically
adds PKG_CONFIG_LIBDIR to the environment before invoking the real
pkg-config.
For all package Makefiles that included pkg-config/buildlink3.mk, remove
that inclusion and replace it with USE_TOOLS+=pkg-config.
- An error in the handling of away messages can be exploited to cause
a heap-based buffer overflow by sending a specially crafted away message
to a user logged into AIM or ICQ.
Successful exploitation allows execution of arbitrary code.
- An error in the handling of file transfers can be exploited to crash
the application by attempting to upload a file with a non-UTF8 filename
to a user logged into AIM or ICQ.
Patches from RedHat.
will install Perl modules into the "vendor" directories:
chat/vicq math/udunits
databases/rrdtool mbone/beacon
devel/p5-subversion
Bump their PKGREVISIONs.
"Multiple "memory alignment errors" in libgadu, as used in ekg before 1.6rc2
and other packages, allows remote attackers to cause a denial of service (bus
error) on certain architectures such as SPARC via an incoming message."
Bump PKGREVISION, patch from Gaim CVS.
The jabberd project team is pleased to announce the release of jabberd 2.0s9.
This is a security release. There is a buffer overflow that could be used to
perform a DoS attack and possible code execution. It is *HIGHLY* recommended
that you upgrade!
ChangeLog:
* fixed only one user is loaded correctly for each router acl
* fixed s2s segfault under particular connection timeout conditions
* fixed id is being case sensitive
* fixed Users cannot login after a long period of server inactivity
* fixed handling of stream errors
* fixed version attribute reply in stream
* fixed c2s glibc abort and mysql option flags
* fixed sx io mem leak
* fixed Incorrect SASL error message defined in sx/sasl.c
* fixed 3 buffer overflows in jid.c
* fixed second log-in in with similar resource breaks routing for first login
Main changelog entries are:
Security:
- Fixed a bug in http_encode that might have caused buffer overflows
(although not likely to be exploitable) when trying to encode strings
with non-ASCII characters.
- Newline stripping added to prevent newline-in-friendlyname attacks.
(Which allowed remote people to make BitlBee send raw custom IRC lines.)
Bugs:
- Many crashes
- Yahoo! cleanup code to avoid 100% CPU time usage
- fixes for ICQ and MSN
approved by wiz@
3 May 2005:
- Released 2.0.2
- Fix to co-exist more nicely with other encrypting gaim plugins.
1 Mar 2005:
- Initial autoconfiscation, thanks to Greg Troxel <gdt@ir.bbn.com>.
around at either build-time or at run-time is:
USE_TOOLS+= perl # build-time
USE_TOOLS+= perl:run # run-time
Also remove some places where perl5/buildlink3.mk was being included
by a package Makefile, but all that the package wanted was the Perl
executable.
easily controllable IRC client for your other POE components and sessions.
You create an IRC component and tell it what events your session cares about
and where to connect to, and it sends back interesting IRC events when they
happen. You make the client do things by sending it events.
changes since 1.3.1:
* Fix system log start times for some protocols
* SILC compiles with newer SILC toolkit versions (Pekka Riikonen)
* Fixed a bug where buddy icon cache files were left in the icon
cache directory after they were no longer in use.
* Attempt to detect the file type of a buddy icon when saving.
* Additional Yahoo! boot protection (Peter Lawler)
* A few Yahoo! memory leaks plugged (Peter Lawler)
* Fixed handling of the new Yahoo! profile page. (Joshua Honeycutt,
Peter Lawler)
* Fixed localized Yahoo! room lists. Please refer to the Yahoo!
section of the Gaim FAQ for details. (Peter Lawler)
* Enabled sending files to ICQ users using ICQ 5.02 and newer
(Jonathan Clark)
passing -Xc to sunpro defines __STDC__=1 and the build fails in
ircsig.c because <signal.h> only declares sigaction if __STD__=0.
fixes build on Solaris w/ sunpro.
Changes:
2.9.4
(1) SSL support can now be used incoming.
(2) Added three new configuration entries to handle new features:
listenex [--ssl][--limit <maxusers>][--localhost <hostname>] <port>
privatekey <filename>
publickey <filename>
2.9.3
(1) SECURITY FIX: FD_SETSIZE overflow DOS
(2) SSL support can now be used if configured with --with-ssl
SSL connection is done by passing -s
ex. /quote conn -s ircs.server
Note: this is only partially secure since SSL is only supported
outgoing.
2.9.2
(1) Added flush to logging
(2) General code changes to fix compiling on some compilers
(3) Fixed problems binding to listening address
2.9.1
(1) SECURITY FIX: password check, was only letting incorrect passwords in
(2) Fixed IP binding on listen
2.9.0
(1) Added trailing newlines to log records
(2) Added extra error handling on accepting connections
(3) Fixed buffer overflow in getnickuserhost (reported by Leon Juranic)
(4) Added extra check for gethostbyname2
(5) Made password check more thorough
(6) Fixed ipv6 dns resolving to random ipv4
2.8.9
(1) Fixed backspace security flaw (reported by Yak)
(2) Fixed compile errors related to compound statements
(3) Rewrote logic of /vip command
(4) Rewrote docked session listing code
2.8.8
(1) Added support for setting a specific IP to listen on
(listen <[host:]port> [maxusers])
(2) Enhanced bncsetup to use new conf format and new question to handle
specific host entering.
(3) Changed the Makefile to list libraries last, some crypt libraries
predefined their own main function which prevented compiling.
(4) Reorganized the connection code to fix a bug where a user gets
disconnected while connecting to an irc server.
(5) Socket length was not being set before accept. (Thanks chris)
(6) Removed old hack code for systems that do not support snprintf.
(7) Increased error checking in recv code.
(8) Better parsing of messages
(9) Server buffers always cleared when connecting to a new server.
(10) Initial IPv6 support. Added -6 option to the CONN command
(i.e. CONN -6 irc.ipv6.org) which is only necessary on ambigious
addresses or when connecting via dns.
(11) Listen host can be specified in conf as an ipv6 address by putting
the address in []'s (i.e. LISTEN [2000:610:0:23::]:6669)
Requested by Peter Avalos <pavalos@theshell.com> in private e-mail.
the modules are statically-linked into the ircservices executable.
This fixes the installation of chat/ircservices on platforms where
dlopen() doesn't obey its "mode" argument, e.g. RTLD_NOW. Unfortunately,
NetBSD/amd64 currently falls into this category (port-amd64/30570),
but this will also fix installation on any a.out NetBSD or OpenBSD
platform.
Approved for commit during the deep freeze by <agc>.
Changes:
1.0:
====
Only minor bugfixes were made to the previous version.
- Fixed channel public key list saving on backup router on JOIN
command reply.
- New optimized logging.
0.9.21:
=======
A small bugfix release.
- Added default limit how many channels one client can join (50).
- Added missing getopt.[ch].
- Fixed compilation with pkg-config files
0.9.20:
=======
A bugfix release to the SILC Server. In addition of various bugfixes,
this version now also includes new math library that from now on will be
included in all SILC distributions.
- Added more liberal channel names from the previous more stricter
identifier string change.
- Added SERVICE command to server, though services aren't supported yet.
- Fixed MOTD command to send empty reply if motd does not exist.
- Fixed LIST command.
- Fixed query to stop if client goes away.
- Added pkg-config check to the configure.
- Several other bugfixes were made.
The project was declared dead on February 27th 2003, although the last version
dates from August 14th 2001. Not to mention that the program does not work
any more due to all the changes in Yahoo's authentication mechanisms in the
meantime.
Changes:
- security fixes for DoS issues:
http://gaim.sourceforge.net/security/index.php?id=18http://gaim.sourceforge.net/security/index.php?id=19
- Fix Yahoo! privacy bug
- Fix Jabber Get Info crash on busted servers
- The file transfer details section now also displays the full
path to the local file sent/received.
- Yahoo! has the following new "/" commands: /join, /buzz
- Updated our gaim.desktop file, thanks to all our terrific
- translators for sending in translations of the changes
- Improvements to how Gaim handles new message notification
- Updated translations (de, sq, zh_CN).
- Fixed crash of server list connect button when no network is
selected while using GTK's auto-find feature [1166669].
- Fixed handling of WhoIs Special event on some networks where it
could chop off the first character [1164315].
- Plugin API changes: Added "nickserv" field to xchat_get_info.
- Python: Fixed get_list() incorrectly failing when the list
contained a time field [1171525].
- Perl: Make scripts using calls with fully qualified subs work again
[1170139] (Lian Wan Situ).
- Fixed input-box input-method (GTK I.M.) problem [1168239].
- Fixed: Ignore and Notify windows incorrectly used the stock CLOSE
button instead of DELETE [1170655].
- Placed Close/Connect buttons in correct position in server list
[1165474].
- Updated translations (ca, de, lt, nl, ru, sk, sr, vi).
- Added command line args -u and -p.
- Fixed handling of "MODE -o+o nick nick" (#1094026).
- Plugin API changes:
* Added "Key Press" print event.
* Added "state_cursor" for xchat_get_prefs.
* Added xchat_strip and xchat_free functions.
* Added "lasttalk" field to "users" list.
* Added "charset" field to xchat_get_info.
- Perl plugin changes (Lian Wan Situ):
* Move each script into their own unique package/namespace. Scripts
containing multiple packages will not be loaded.
* When warning messages are emitted you will now be told which
script it came from.
* Xchat::set_context will now accept Xchat::set_context( $channel )
and Xchat::set_context( $channel, $server ) in addition to
Xchat::set_context( $context ).
* Fix display of loaded scripts in the Plugins and Scripts window.
- TCL: Fixed crash with invalidated TCL timer (#1110306) (Daniel P.
Stasinski).
- /TIMER now supports timeouts to one decimal place.
- Fixed possible crash of open-file dialog on 64-bit machines.
- Pressing CTRL-O in the DCC Receive window will now open your
downloads folder.
- Win32: Default download folder changed to "My Documents\Downloads".
- Added -quiet arg to the /charset command.
- The /country command now supports a wildcard search.
- The user is now warned when real/user name is left blank in the
server list window.
- Added the /URL command.
- Added a text event for all unknown WHOIS reply lines.
- Added /ALLCHANL which sends to the current server only.
- Actions (/ME) are now treated like PRIV/CHAN for purposes of the
ignore list.
package) into "Makefile".
- Don't include "pkgsrc/devel/glib2/buidlink3.mk" because the package
inherits that dependence automatically from the "gtk2+" package.
While here, reintegrate Makefile.common into Makefile since xchat-gnome
has been removed from pkgsrc and there is no further need for Makefile.common.
tron@ says do it.
Several changes are involved since they are all interrelated. These
changes affect about 1000 files.
The first major change is rewriting bsd.builtin.mk as well as all of
the builtin.mk files to follow the new example in bsd.builtin.mk.
The loop to include all of the builtin.mk files needed by the package
is moved from bsd.builtin.mk and into bsd.buildlink3.mk. bsd.builtin.mk
is now included by each of the individual builtin.mk files and provides
some common logic for all of the builtin.mk files. Currently, this
includes the computation for whether the native or pkgsrc version of
the package is preferred. This causes USE_BUILTIN.* to be correctly
set when one builtin.mk file includes another.
The second major change is teach the builtin.mk files to consider
files under ${LOCALBASE} to be from pkgsrc-controlled packages. Most
of the builtin.mk files test for the presence of built-in software by
checking for the existence of certain files, e.g. <pthread.h>, and we
now assume that if that file is under ${LOCALBASE}, then it must be
from pkgsrc. This modification is a nod toward LOCALBASE=/usr. The
exceptions to this new check are the X11 distribution packages, which
are handled specially as noted below.
The third major change is providing builtin.mk and version.mk files
for each of the X11 distribution packages in pkgsrc. The builtin.mk
file can detect whether the native X11 distribution is the same as
the one provided by pkgsrc, and the version.mk file computes the
version of the X11 distribution package, whether it's built-in or not.
The fourth major change is that the buildlink3.mk files for X11 packages
that install parts which are part of X11 distribution packages, e.g.
Xpm, Xcursor, etc., now use imake to query the X11 distribution for
whether the software is already provided by the X11 distribution.
This is more accurate than grepping for a symbol name in the imake
config files. Using imake required sprinkling various builtin-imake.mk
helper files into pkgsrc directories. These files are used as input
to imake since imake can't use stdin for that purpose.
The fifth major change is in how packages note that they use X11.
Instead of setting USE_X11, package Makefiles should now include
x11.buildlink3.mk instead. This causes the X11 package buildlink3
and builtin logic to be executed at the correct place for buildlink3.mk
and builtin.mk files that previously set USE_X11, and fixes packages
that relied on buildlink3.mk files to implicitly note that X11 is
needed. Package buildlink3.mk should also include x11.buildlink3.mk
when linking against the package libraries requires also linking
against the X11 libraries. Where it was obvious, redundant inclusions
of x11.buildlink3.mk have been removed.
- internal libgadu now work with gnutls instead openssl library.
- fixed crash while start.
- Tlen.pl configuration moved to .gg2/tlen file
- Then to Everaldo www.everaldo.com for his Jabber icons.
- close bugs #177,#146,#170,#141,#148
- fixes of parsing links in chat window
- and others.
Changes:
- Fixes for two remotely exploitable crash bugs. See
http://gaim.sourceforge.net/security/ for more information.
- Removed parts of the font selection dialog that were not respected
- Fix being invited to a multi user chat on MSN
- Multiple SILC accounts should work now (Pekka Riikonen)
- Fix times on jabber chat backlogs
- Fix gevolution plugin to compile with e-d-s 1.0 or 1.2
- Fix gevolution plugin to remember buddy name when someone added you
and you then add them
- Formatting in jabber chats works
- Fix to prevent MSN disconnecting if you change status while
- connecting
- Change to correctly handle adding jabber buddies on ejabberd servers
Mostly from MAINTAINER via PR pkg/30204
And always is defined as share/examples/rc.d
which was the default before.
This rc.d scripts are not automatically added to PLISTs now also.
So add to each corresponding PLIST as required.
This was discussed on tech-pkg in late January and late April.
Todo: remove the RCD_SCRIPTS_EXAMPLEDIR uses in MESSAGES and elsewhere
and remove the RCD_SCRIPTS_EXAMPLEDIR itself.
0.11.7.3:
A serious bug in naim's HTML rendering engine has been identified and
corrected. Additionally, changes to the packaging have reduced the source
tarball size from 529 kB in 0.11.7.2 to 452 kB for 0.11.7.3. naim now
completely passes "make distcheck", and distcheck is used to create source
tarballs. FireTalk now allows you to send messages to the :RAW target before
signing on, to allow passwords to be sent to the server during signon. The
IRC driver now handles /names replies from servers that use non-standard
status identifiers (such as those who implement "halfop" and other
statuses). The UI includes code to suppress duplicate messages and redundant
message targets. The embedded Libtool was upgraded to match the stock 1.5.10
release. A new "proto_user_onlineval" chain was added to allow module
authors to track buddy online status.
0.11.7.3.1:
On some systems, naim's build system was unable to generate a list of
special characters to read, which caused naim to interpret the Home key as
Ŋ, etc. naim's build system has been changed to work around the issue.
No other changes have been made since 0.11.7.3.
patch by Evan Hall in PR 29964 (slightly modified)
0.80 - 04/13/2005
* Changed keepalive messages to FLAP channel 0x05 messages, as recommended.
* Re-implemented idle time reporting. Any keystroke in bsflite now resets
your idle time, rather than only on sending a message.
* New option "respond_idle_only" to send away message auto-response only
when away and idle.
* Fixed the terminal line erasing problem previously detailed in the README
both in UNIX (using vt100 sequences), Windows (using console functions)
and DOS (using conio functions). Plan 9 is unaffected since the rio
terminal properly supports reverse line wrap.
* Screen names longer than 16 characters are trimmed to the first
13 characters plus "..." in the buddy list display. Use 'wf' to
see the entire name. This is necessary now with @mac.com and
@thematrixonline.com accounts.
* New command 'G' : write away messages in ~/.bsflite/awaymessages,
one per line, and use 'G' to set them.
* Finally added server-side buddy lists. Buddy add/delete now stores the
information on the server. If you have a buddies file, run this
new version with it once, then delete it.
* Optional "netspeak cleaner" - see cleaner.c and config file
* Cleaned up and sped up login sequence.
* Fixed a mysterious bug where mobile users would disappear from the
buddy list even though they were online.
* imcomm now supports sending more than one packet at a time, which
seems to be how other clients do some things. The fetch awaymsg/profile
command now takes advantage of this, which seems to improve the rate
limit situation a bit. Also, the 'b' command works most of the time
without triggering a rate limit warning.
* Buddy list display now adjusts with screen size.
* Added a configuration file. See config-example for details.
* Changed default paths to ~/.bsflite/[config,log,profile,buddies,icon]
* Fixed the endlessly repeating error messages on certain platforms.
* l command now shows the last 20 lines instead of 10.
* Implemented MD5-based login in IMComm instead of the old, "roasted"
password algorithm. This improves things on three fronts:
- your password is no longer sent in plaintext
- you can login with a @mac.com account
- you can login with an ICQ UIN
This is the way the official client has done things for a while. I used
MD5 routines developed by L. Peter Deutsch (md5.c, md5.h in imcomm/)
for this, at least until I become inspired enough to write my own.
Changes:
- convert to options.mk
1.0.2:
======
Due to the new SILC Server release with support to the UTF-8 nicknames and
channel names new version of the SILC Client is released as it seems the
older ones do not function very well with the new server. This version
includes support to the UTF-8 encoded nicknames and channel names, however,
certain problems may still exists, as this is a slight premature release.
- Added support for UTF-8 encoded nicknames and channel names.
- Command queue implemented for better UTF-8 support.
- --with-silc-includes and --with-silc-libs added gto configure to use
pre-installed SILC Toolkit.
Changes:
- convert to options.mk
0.9.19p1:
=========
A little update with this 0.9.19p1. After such a major release problems
were expected and the p1 fixes some crashes. Upgrade strongly
recommended.
0.9.19:
=======
And after a long break new SILC Server is out. This version finalizes the
SILC protocol version 1.2 development and introduces UTF-8 nicknames, channel
names, usernames and host names. It is now possible to create practically any
kind of nicknames and channel names. Practically all letters, numbers and
punctuation marks are supported. Special characters, control characters and
various odd symbol characters however are not allowed. Several minor and
major bugs has been fixed as well. Upgrading is strongly recommended. Old
clients that does not yet support UTF-8 encoded nicknames and channel names
are still able to connect and function normally as long as they do not need
to handle odd UTF-8 encoded names.
- Added support for UTF-8 encoded identifier strings, such as nicknames
and channel names.
- Fixed founder mode handling on JOIN on normal/backup on empty
channels.
- Fixed WATCH command handling on backup router.
- Fixed WATCH command announcing. The WATCH and SILC Gaim buddy list
should work better now.
- Simplified INVITE and BAN string handling in server. Announcing
INVITE and BAN strings should work more reliably now.
- Fixed several bugs from the backup and resuming protocol.
- Fixed, hopefully, all the wrong server statistics numbers.
- Fixed CLOSE command to handle backup routers correctly.
- Fixed various detaching and resuming bugs.
- Fixed announcing to not announce unregistered (ghost) clients.
- Fixed reconnect_keep_trying and QoS settings in server config files.
- Several other bugfixes were made.
4.14.0
------
Because of the mistake I made in the previous version, the 'u' letter was
eaten by the RTF parser when processing messages from windows icq clients.
It looked for unicode chars, but ate all 'u' characters it used to found.
Fixed that.
A memory leak problem when checking RSS was fixed.
Gadu-Gadu module used to produce a segfault when it was unable to obtain
a registration token from the server. Now it doesn't crash, though the
impossibility to register a new UIN is still there. The serve simply
doesn't find an appropriate document by the registration URL. Too bad.
Applied a small patch by Paul Chitescu so that it's now possible to start
centericq with all protocols set to offline.
Another type of external event was added. Now you can execute your actions
when someone goes offline. Hope someone finds that useful.
4.20.0
------
PGP encryption support was implemented for the Jabber module. In a nutshell
this means that with a proper setup your communications get a way more secure.
GPGME library is needed to build the PGP support.
Fixed the RTF parser that takes charge when a message from a Windows ICQ
client is received. One problem was that it used to treat backspace characters
wrong sometimes. Another problem with utf-16 was fixed with a patch submitted
by Vadim Nosovsky.
After ICQ dual login detection centericq used to try re-connecting, though it
wasn't supposed to. Fixed.
During a long time noone reported me the bug related to libicq2000's header
files getting installed each time "make install" was run for centericq. Finally
Wolfram Schlich noted this one, so I fixed it.
Logging on Yahoo! didn't work on Sun Sparc machines. Fixed that. That might
also revive Gadu-Gadu on bigendian architectures.
A couple of interface related bugs were fixed, such as a mistake in the group
organization dialog as well as not clearing pending messages for contacts in
non-chat communication mode.
Security fixes for three DoS vulnerabilities:
Remote DoS on receiving malformed HTML
Remote DoS on receiving certain messages over IRC
Jabber remote crash
Changes:
- URL escaping now works with UTF-8 text. This may break some old
log files.
- Revert to XOR auth for ICQ as the md5 is not fully functional
- Fix bug with going away while in a jabber chat
- MSN bug fixes (Felipe Contreras)
- Escape things properly in IRC
From MAINTAINER via PR pkg/29888
EPIC4-2.2
*** News -- 11/09/2004 -- Changes to way spaces are handled *** IMPORTANT ***
Up until this point, the technical definition of what is a "space"
(the term used for an Internal Field Separator (IFS), that is, the
thing that separates one word from another word)
EPIC has always had three sets of IFSs
1) Character 32 only ("space")
2) Characters 9, 10, 11, 12, 13, and 32 ("my_isspace()")
3) Whatever your locale says is a space ("isspace")
In the C locale (the default for unix users), sets #2 and #3 are
exactly the same. I don't know about other locales.
Now we are going to only use set #3. Changing Set #2 to Set #3 is
easy, and nobody should notice any difference there. The one that
is going to cause pain is Set #1. Consider this situation:
One<tab>Two
Is this one word, or two words? In some places in epic, it is one
word, and in other places it is two words. As a result of this
change, it will now always be two words every place.
Here is a list of the places that were using Set #1 that will change
to use Set #3, and you should be on the lookout for changes!
*) /AWAY separated the end of its flags with spaces.
This means
/AWAY -ALL<tab>Hi there!
will now work properly.
*) $pop(....) removes the last word from the argument list.
Whereas $pop(one two<tab>three) used to return "two<tab>three"
is will now only return "three"
*) Commands are separated from the argument list by a "space".
Whereas
/one<tab>two blah blah blah
parsed "one<tab>two" as the command, now "one" is the command
and "two" is part of the argument list.
*) In expressions, a variable name may be separated from an
operator with a space. Whereas things like
@var<tab> =foo
might have failed because <tab> is not a valid character in
a variable name, it will now treat <tab> the same as a space.
*) Places that expect a number did not accept a string that contained
a <tab> so that things like
<tab><tab>9
was not accepted as a number. Now it will be.
*) The /IGNORE command used to consider a string containing a <tab>
to not be empty. This means that
/IGNORE <tab>
would not list the ignorance list, but rather would try to show
the ban value for an ignore matching <tab> which isn't reasonable.
*) The % wildcard pattern stopped matching only when it saw char 32.
Now it will stop whenever it seems a Set 3 type space.
So whereas before "%" matched "one<tab>two" now it won't,
because "one<tab>two" is two words, not one word.
*) In /xdebug extractw mode, tabs before or after double quotes were
not considered to start or end a double quoted word. So previously
this string:
one <tab>"two three" four
contained four words, because the <tab> before "two nullified
the double quoting. Now that <tab> is treated like any other
space, the above word has three words, not four.
Places that use Set 1 and will NOT be changing to Set 3, because
they are parsing IRC protocol data, which stipulates that the IFS
MUST be a space solely:
*) CTCP requests and replies are formatted only with spaces
*) Words in protocol messages are separated only with spaces
*** News -- 11/09/2004 -- You may now mangle ALT_CHAR characters
And ALT_CHAR mangling is included in ALL. This was an oversight
that was fixed in epic5.
*** News -- 11/02/2004 -- Mangling "ALL,-BOLD" no longer mangles "ALL_OFF"
In general, if you use ANSI (ALL includes ANSI), the mangled string
has its six attributes (COLOR, REVERSE, UNDERLINE, BOLD, BLINK,
ALT_CHAR) rewritten into canonical form. This will add some ALL_OFFs
to your string that weren't there originally. So if you strip all of
the attributes (as ALL does), then epic will strip ALL_OFF off as well.
This retains backwards compatability with ALL.
But if you use ANSI and don't want to strip all 6 of the attributes,
then it's important that ALL_OFFs are not removed, otherwise your
string will not appear as it should (the ANSInator uses ALL_OFFs to
turn off attributes). So EPIC automatically will not strip ALL_OFFs
if you use ANSI and do not mangle one of the 6 attributes.
Examples:
$stripcrap(ALL this is ^B^_bold underline^_ not bold)
will strip everything as it has always done.
$stripcrap(ALL,-BOLD this is ^B^_bold underline^_^B not bold)
will strip everything but not bold and all_offs, because if all_offs
are stripped, then "not bold" will be in bold!
The entire point of this is to allow /set mangle_logfile ALL to
work the way it has always worked, and $stripcrap(ALL,-BOLD ...)
to work the way it *should* work.
*** News -- 10/06/2004 -- Support for +e and +I numerics from efnet
EFNet has +e and +I channel modes, which act like +b does.
These numerics are now handled by epic like +b is.
*** News -- 10/06/2004 -- New status bar expandos, %{2}W and %{3}W
These two new expandos expand to the same value as %W.
%W Expands on each input window on each screen that has
two or more visible windows.
%{2}W Expands on all visible windows on all screens.
%{3}W Expands on each input window on each screen, even on
screens that have only one visible window.
*** News -- 10/06/2004 -- Support for ircnet's "numeric nick" feature
On ircnet, each user is given a unique numeric identifier, which
is their one true nickname. In addition to this numeric id, they
can have a rfc1459 nickname, but they are not required to have one.
Further, the special numeric id 0 is reserved and refers to the
user's own numeric id. EPIC now fully supports all of this,
particularly in the following ways:
/NICK <unique id>
/NICK 0
To turn off your rfc1459 nickname.
$serverctl(GET <refnum> UNIQUE_ID)
$serverctl(SET <refnum> UNIQUE_ID)
To retrieve and change your unique nickname.
Changing your unique id is probably a bad idea.
/USERHOST <unique id>
You may USERHOST unique id's now.
'epic 0'
You may now use the nickname 0 on the epic command
line if you don't want to use an rfc1459 nickname
on an ircnet server.
Using unique numeric id's on non-ircnet servers is probably
fraught with peril. Try to avoid that.
*** News -- 10/06/2004 -- /SET INDENT maxes out at 1/3 screen width
Previously, if you did /set indent on, and the width of the
first word on the first line was > 1/3 of the screen's width,
then the second (and subsequent) lines were not indented at all.
This was the historical ircII behavior. It seems more sensible
to cap the indent level at 1/3 of the screen's width and indent
it to there.
EPIC4-2.0
*** News -- 02/03/2004 -- ERASE_LINE behavior slightly modified
The ERASE_LINE input function has been modified to not overwrite the
input buffer when it is called on an empty input line.
EPIC4-1.2.9
*** News -- 01/25/2004 -- Slight changes to /ON SSL_SERVER_CERT
|Rain| pointed out that /ON SSL_SERVER_CERT was not terribly useful
because $1 (cert subject) and $2 (cert issuer) could contain spaces,
and so you couldn't really use $1 and $2 to fetch them. He proposed
that we hook a url-encoded version of the subject and issuer (so they
won't contain any spaces) and you can use $urldecode($1) to get the
subject and $urldecode($2) to get the issuer.
EPIC4-1.2.7
*** News -- 01/18/2004 -- New functions $cofilter() and $corfilter().
These are compliments of $copattern().
*** News -- 12/27/2003 -- New $dccctl() modes.
Added $dccctl(get [ref] writable) and $dccctl(writables). This is
useful for sending lots of data out on a raw or chat dcc. Epic will
(iirc) normally drop the data in such cases rather than blocking or
buffering. This isn't a complete solution. Writable means that a
certain amount of data can be written, so data can still be lost.
*** News -- 12/27/2003 -- Added /SET DCC_DEQUOTE_FILENAMES
This turns the double quoted space decoding feature added recently for
received DCC SEND requests on and off.
*** News -- 12/27/2003 -- $rest() also removes chars from the right.
Using a negative first arg causes $rest() to strip chars from the right
hand side of the string. This makes it a compliment to $right() as
well as $left(), but if a script relies on it returning the original
string the way it did before, it may now cause compatibility problems.
If it does, talk to CrazyEddy on #epic efnet.
*** News -- 12/27/2003 -- Added $splicec() in functions script.
This works much the same way as $splice(), but on characters, not
words. It won't work seemlessly on local vars though.
EPIC4-1.2.5
*** News -- 12/13/2003 -- Default of AUTO_REJOIN is now OFF
Per the poll taken on the epic mailing list, the preponderance of
opinion being in favor of changing the default value of AUTO_REJOIN
to OFF, as of commit 672, the default value is changed. You will
want to put /SET AUTO_REJOIN ON in your ~/.epicrc to get the old
value.
*** News -- 12/11/2003 -- Command parser ignores spaces after semicolons
Since the dawn of the modern ircII syntax, the following syntax
has caused consternation with scripters:
/eval {two; three}
because it runs the /two alias, and then outputs the text "three"
to the current target (channel-or-query). This is because the
command parser does not ignore spaces after semicolons; a leading
space results in ircII assuming you are wanting to run the empty
command (yes, there is actually an empty command), which just
happens to be the /say command. Well, enough is enough. I'm
putting a stop to this. If you depend on this horrifically broken
behavior, then the time has come to switch to using /SAY. The
rest of us will enjoy not having to freak out about spaces after
semicolons.
*** News -- 12/11/2003 -- New /WINDOW verb, /WINDOW SCROLLADJ [ON|OFF]
When the default value of ON is used, when you "GROW" a window,
or otherwise cause the size of a window to be increased, the top
of the scrollable view is moved back so you can see more of the
lines that have recently scrolled off the screen. If you change
this to OFF, the behavior is more like ircII, where when something
scrolls off the window, nothing will bring it back unless you go
into scrollback mode.
*** News -- 12/11/2003 -- New /ON, /ON WINDOW_BEFOREKILL
The value of $* is the window refnum (number) of the window that
is about to be killed. You cannot stop the kill from happening,
but you do get this opportunity to save any information about the
window that you might want. *DANGER*DANGER*DANGER* EPIC does not
try to stop you from doing something really dumb in this /ON, and
if you try to change the window being killed, or you affect the
visibility or invisibility of any window, you run the definite chance
of confusing EPIC and causing a panic. This is not a bug, this is
just your peril if you ignore this warning. Use /DEFER to run
commands safely from within /ONs.
EPIC4-1.2.3
*** News -- 12/05/2003 -- New built in function $numlines() (fudd)
The first argument to $numlines() is the width of a line (which you
can get from $geom() if you want to use that), and the rest of the
text is some text. The return value is the number of screen lines
that will be taken up if you were to /echo the text to a window of
the given number of columns.
*** News -- 12/03/2003 -- New command line arguments, -o and -O
POSIX termios allowed systems to define "extended characters" (IEXTEN)
which are interpreted by the kernel and are not passed onto the
application. In 4.4BSD in particular, these "extended characters"
are ^V and ^O. This means that by default, you cannot use these
two characters in a binding. By using the -o option, you ask
EPIC to specifically turn these extended characters on, and by
using the -O option, you ask epic to specifically turn these
extended characters OFF. If you do not specify either option,
epic does not change the current default value.
EPIC4-1.2.2
*** News -- 12/03/2003 -- New /WINDOW verb, /WINDOW SWAPPABLE [ON|OFF]
You may now control whether your windows are "swappable". The
default (backwards-compatable) value is ON. If you set this to
OFF on a visible window, then that visible window cannot be made
hidden, via /WINDOW HIDE or /WINDOW SWAP. If you set this to OFF
on a hidden window, then that hidden window cannot be made visible,
via /WINDOW SHOW, /WINDOW SWAP, /WINDOW BACK or /WINDOW NEXT. You
must reset this value back to ON before you try to swap the window
back in or out.
*** News -- 12/03/2003 -- New /SET, /SET STATUS_NOSWAP
Whenever a window has /WINDOW SWAPPABLE OFF, the %{1}W status expando
will expand to the value of this /SET
*** News -- 12/03/2003 -- SWAPPABLE accessable via $windowctl()
You can now $windowctl(GET <refnum> SWAPPABLE) to get the window
<refnum>'s current swappable status.
* Support for Perl scripts is turned off
* Support for skins compatible with Gadu-Gadu
* Fixes and stabilisation of the code for status changes
* Fixes in tlen.pl plug-in
* Now builds basic modules on MacOSX
* Improvements of "dockapp" plug-in
dircproxy is an IRC proxy server ("bouncer") designed for people
who use IRC from lots of different workstations or clients, but wish
to remain connected and see what they missed while they were away.
You connect to IRC through dircproxy, and it keeps you connected to
the server, even after you detach your client from it. While you're
detached, it logs channel and private messages as well as important
events, and when you re-attach it'll let you know what you missed.
This can be used to give you roughly the same functionality as
using ircII and screen together, except you can use whatever IRC
client you like, including X ones!
* THIS IS AN UNSTABLE DEVELOPMENT RELEASE OF DIRCPROXY. IT HAS HIGHER
* THEN USUAL RISK OF DESTROYING YOUR SYSTEM.
*
* FOR A STABLE RELEASE USE THE LATEST 1.0.x