ngircd/ngircd

Free, portable and lightweight Internet Relay Chat server

ngIRCd - Internet Relay Chat Server

Introduction

ngIRCd is a free, portable and lightweight Internet Relay Chat (IRC) server for small or private networks, developed under the terms of the GNU General Public License (GPL); please see the file COPYING for licensing information.

The server is quite easy to configure and runs as a single-node server or can be part of a network of ngIRCd servers in a LAN or across the internet. It optionally supports the IPv6 protocol, SSL/TLS-protected client-server and server-server links, the Pluggable Authentication Modules (PAM) system for user authentication, IDENT requests, and character set conversion for legacy clients.

The name ngIRCd stands for next-generation IRC daemon, which is a little bit exaggerated: lightweight Internet Relay Chat server most probably would have been a better name :-)

Status

Development of ngIRCd started back in 2001: The server has been written from scratch in C, tries to follow all relevant standards, and is not based on the forefather, the daemon of the IRCNet.

It is not the goal of ngIRCd to implement all the nasty behaviors of the original ircd or corner-cases in the RFCs, but to implement most of the useful commands and semantics that are used by existing clients.

ngIRCd is used as the daemon in real-world in-house and public IRC networks and included in the package repositories of various operating systems.

Advantages and strengths

• Well arranged (lean) configuration file.
• Simple to build, install, configure, and maintain.
• Supports IPv6 and SSL.
• Can use PAM for user authentication.
• Lots of popular user and channel modes are implemented.
• Supports "cloaking" of users.
• No problems with servers that have dynamic IP addresses.
• Freely available, modern, portable and tidy C source.
• Wide field of supported platforms, including AIX, A/UX, FreeBSD, HP-UX, IRIX, Linux, macOS, NetBSD, OpenBSD, Solaris and Windows with WSL or Cygwin.

Documentation

The homepage of the ngIRCd project is https://ngircd.barton.de.

The INSTALL.md document describes how to install and upgrade ngIRCd. It is included in all distribution archives and available online on GitHub.

Please see the file doc/QuickStart.md in the doc/ directory or on GitHub for information about setting up and running ngIRCd, including some real-world configuration examples.

More information can be found in a couple of files in the doc/ directory (online on GitHub) and in the documentation section on the homepage of ngIRCd.

In addition, ngIRCd comes with two manual pages: ngircd(8) (for the daemon) and ngircd.conf(5) (for its configuration file). They have even more details and list all possible command line parameters and configuration options. You can read them with the man command (when they are installed locally on your system, e.g. man 8 ngircd and man 5 ngircd.conf) or online here:

• Daemon: ngircd(8)
• Configuration file: ngircd.conf(5)

Downloads & Source Code

You can find the latest information about the ngIRCd and the most recent stable release on the news and downloads pages of the homepage.

Visit our source code repository at GitHub if you are interested in the latest development code.

Problems, Bugs, Patches

Please don't hesitate to contact us if you encounter problems:

• On IRC: irc://irc.barton.de/ngircd
• Via the mailing list: [email protected]

See https://ngircd.barton.de/support for details.

If you find any bugs in ngIRCd (which most probably will be there ...), please report them to our issue tracker at GitHub:

• Bug tracker: https://github.com/ngircd/ngircd/issues
• Patches, "pull requests": https://github.com/ngircd/ngircd/pulls

ngircd's Issues

AUTH Service

(Report imported from Bugzilla #46)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform PC.
Assigned to: Alexander Barton.

On 2003-09-14 21:32:28 +0200, Tobias Mathes wrote:

Authoration Service like Q on Quakenet would be very nice.

On 2003-10-20 00:36:59 +0200, Alexander Barton wrote:

I don't know the "Q" command on Quakenet, but it looks to me comparable to
NICKSERV services. Therefore I think hat it shouldn't be implemented in the
server but insted in an -- still to be written -- IRC services package for ngIRCd.

(BTW: sorry for answering that late, but Bugzilla hasn't mailed me this report!?)

On 2003-10-28 01:35:21 +0100, Tobias Mathes wrote:

No problem, I'll ask a friend if he want to write an service for ngircd. I very
like ngircd. ;)

On 2003-11-06 03:08:10 +0100, Alexander Barton wrote:

Great!

If you have any questions regarding the internals of the server, just ask!

english "sample-ngircd.conf" is missing

(Report imported from Bugzilla #28)

Status CLOSED, severity enhancement, in component Documentation (english).
Reported in version unspecified on platform All.
Assigned to: Ilja Osthoff.

On 2002-10-12 12:14:46 +0200, Alexander Barton wrote:

There is no english "sample-ngircd.conf" available at the moment,
but it would be very useful ;-)

Someone interested in writing one?

On 2002-10-12 12:16:39 +0200, Alexander Barton wrote:

Strange ... I thought there has been some english documentation for the
ngircd.conf -- but actually I can't find it (any more)!?

On 2002-11-09 00:10:41 +0100, Ilja Osthoff wrote:

Now there be an english sample.conf.

On 2002-11-09 00:29:48 +0100, Alexander Barton wrote:

I added the english translated file as doc/en/sample-ngircd.conf
to CVS-HEAD. Thanks Ilja!

Squelching duplicate server to server link attempts

(Report imported from Bugzilla #29)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version 0.5.2 on platform All.
Assigned to: Alexander Barton.

On 2002-10-12 18:19:39 +0200, Joshua Coombs wrote:

ngircd, when configured as below, will correct establish a link between two
servers. Periodically, one server will attempt to form a new connection
between the same two servers. ngircd will recognize the duplicate connection
attempt and drop it.

server 1:
[Global]
Name = portland.me.us.x386.net
Info = Server Info Text

[Server]
Host = caern.thesept.com
Name = portland2.me.us.x386.net
Password = test
Port = 6667

server 2:
[Global]
Name = portland2.me.us.x386.net
Info = Server Info Text
Ports = 4000

[Server]
Host = cyrix-dlc.x386.net
Name = portland.me.us.x386.net
Port = 6667
Password = test

The prefered behavior for servers configured as above would be both sides
attempting to connect to each other. Once a connection is established, neither
side attempts more connections untill the existing one drops.

On 2002-11-08 23:29:03 +0100, Alexander Barton wrote:

I added code to solve this problem to CVS-HEAD. It should work,
But actually it is absolutely untested ...

On 2002-11-18 19:42:29 +0100, Alexander Barton wrote:

In the meantime I tested it a little bit. Seems to work :-)

CONNECT ist nur unzureichend implementiert

(Report imported from Bugzilla #39)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.5.3 on platform All.
Assigned to: Alexander Barton.

On 2002-11-23 00:15:57 +0100, Alexander Barton wrote:

Die bisherige Implementation von CONMECT erlaubt es nur, den "passive mode"
(--passive bzw. -p) des ngIRCd zu deaktivieren; ein geziehlter Verbindungsaufbau
zu einem anderen Server ist nicht m

On 2003-01-03 14:59:54 +0100, Alexander Barton wrote:

CONNECT is implemented in CVS-HEAD now:

• "CONNECT "
  Enable and connect an configured server
• "CONNECT "
  Add a new server (ngIRCd tries to connect new servers only once!).

On 2003-01-27 17:02:04 +0100, Alexander Barton wrote:

Seems to work as expected.

WHOWAS is missing

(Report imported from Bugzilla #40)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version 0.7.1 on platform All.
Assigned to: Alexander Barton.

On 2002-11-23 00:25:17 +0100, Alexander Barton wrote:

Der ngIRCd kennt den WHOWAS-Befehl noch nicht, wie er in RFC 2812
beschrieben ist:

3.6.3 Whowas

  Command: WHOWAS

Parameters: *( "," ) [ [ ] ]

On 2005-06-04 13:27:34 +0200, Alexander Barton wrote:

The WHOWAS command is now implemented in CVS HEAD and will be included in
upcoming new releases of ngIRCd.

On 2005-11-07 13:54:59 +0100, Alexander Barton wrote:

Seems to work just fine -- I'll close this bug :-)

Channel Topics und der TOPIC-Befehl werden nicht unterstützt

(Report imported from Bugzilla #9)

Status CLOSED, severity normal, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-02-24 04:48:48 +0100, Alexander Barton wrote:

Der ngIRCd unterst

On 2002-02-24 04:50:01 +0100, Alexander Barton wrote:

Der Server wird Channel Topics unterst

On 2002-02-28 02:24:11 +0100, Alexander Barton wrote:

Ich habe Channel Topics sowie den TOPIC-Befehl implementiert.

getsockopt() liefert Fehler: "No child processes"

(Report imported from Bugzilla #27)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.5.2 on platform Macintosh.
Assigned to: Alexander Barton.

On 2002-10-07 00:50:15 +0200, Alexander Barton wrote:

Wenn der ngIRCd unter A/UX versucht, sich zu anderen Servern zu
verbinden, so liefert getsockopt() folgenden Fehlercode (der
Aufruf von getsockopt() selber ist erfolgreich!):

Can't connect socket: No child processes!

Quelle: src/ngircd/conn.c, Zeile 626 ff.

On 2002-11-11 01:29:21 +0100, Alexander Barton wrote:

Der Fehler kommt gar nicht von getsockopt(), sondern von connect() ...

On 2002-11-11 02:01:54 +0100, Alexander Barton wrote:

Der Fehler-Code von connect() wurde falsch bzw. nicht ausgewertet,
ich habe den Test korrigiert, der Server kann sich nun auch unter
A/UX wieder zu anderen Servern connectieren.

BTW: der Fehler ist auf allen Plattformen vorhanden gewesen, scheint
aber nur unter A/UX (bisher) aufgefallen zu sein ...

Jetzt sollte er jedenfalls behoben sein ;-)

On 2003-03-20 16:46:51 +0100, Alexander Barton wrote:

Seems to work.

Send "lost server" messages compatible to original ircd

(Report imported from Bugzilla #58)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2004-09-03 02:09:25 +0200, Alexander Barton wrote:

The "lost server" messages should be formated in a way compatible to the original ircd, so IRC
clients (like irssi) would be able to print out nicely formated "net split messages" instead of a bunch
of "server leaving" messages.

On 2005-06-04 17:06:09 +0200, Alexander Barton wrote:

As of today Florian Westphal committed a patch to implement this.

On 2005-11-07 13:55:18 +0100, Alexander Barton wrote:

Seems to work just fine -- I'll close this bug :-)

Multiple connections with same Group ID

(Report imported from Bugzilla #44)

Status CLOSED, severity critical, in component Daemon.
Reported in version development on platform Sun.
Assigned to: Alexander Barton.

On 2003-05-08 10:27:05 +0200, Marek Czernohous wrote:

the ngircd seems to connect to multiple other servers even if they have the same
Group-ID.
Even if there is no port configured!

On 2003-05-08 15:17:22 +0200, Alexander Barton wrote:

How can I reproduce this?

On 2003-07-28 15:24:47 +0200, Alexander Barton wrote:

Seems to be solved!?

"Seen"-Support

(Report imported from Bugzilla #34)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-11-21 17:47:44 +0100, Alexander Barton wrote:

Der Server sollte eine Alternative zum "Seen"-Support einiger Bots anbieten:
evtl. l

On 2002-12-02 15:00:29 +0100, Alexander Barton wrote:

Ein Bug f

On 2004-09-03 02:00:59 +0200, Alexander Barton wrote:

A "seen" command is a feature of a bot; it's nothing that has to be implemented by the server.

On 2004-10-14 00:18:18 +0200, Alexander Barton wrote:

Nothing to do at the moment ...

LUSERS zeigt falsche Anzahl Server-Links

(Report imported from Bugzilla #22)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.4.1 on platform All.
Assigned to: Alexander Barton.

On 2002-04-12 12:08:23 +0200, Alexander Barton wrote:

Der LUSERS-Befehl zeigt die Anzahl der Server-Links des jeweiligen
Servers falsch an. Es wird wohl immer "1" geliefert, unabhaengig von
der tatsaechlichen Anzahl.

On 2002-04-14 15:56:18 +0200, Alexander Barton wrote:

Ich habe einen Patch ins CVS eingespielt, die Funktion Count_MyServers() in
client.c hat falsch gezaehlt. Sollte nun behoben sein.

BitchX -- PING

(Report imported from Bugzilla #19)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.3.0 on platform All.
Assigned to: Alexander Barton.

On 2002-03-08 15:55:53 +0100, Georg Buschbeck wrote:

BitchX meldet regelm

On 2002-03-10 23:43:43 +0100, Alexander Barton wrote:

Ich habe IRC_PING() toleranter gemacht, wenn ngIRCd nicht mit "--enable-strict-rfc"
konfiguriert wurde: es akzeptiert dann beliebig viele Parameter. BitchX sendet naemlich
z.B. solche Anfragen: 'PING 1015798870 irc3.my.net ngircd-0.3.0-CVS aio amnopqstv'.
RFC-konform ist das nicht.

MODE parser is too limited

(Report imported from Bugzilla #35)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.5.3 on platform All.
Assigned to: Alexander Barton.

On 2002-11-22 22:08:17 +0100, Alexander Barton wrote:

Die Funktion IRC_MODE(), die den MODE-Befehl verarbeitet, ist zu limitiert.

Viele Parameter, die der Orginalserver problemlos versteht, werden nicht oder
nicht richtig erkannt und behandelt. Erweiterungen, z.B. f

On 2002-12-14 17:07:19 +0100, Alexander Barton wrote:

I started rewriting it at the moment :-)

On 2002-12-17 00:15:55 +0100, Alexander Barton wrote:

I just commited a new parser to CVS-HEAD. It will be included in 0.6.0.

On 2003-01-27 17:01:28 +0100, Alexander Barton wrote:

I fixed some more bugs in CVS-HEAD which are included in 0.6.1, too.
Seems to work quite well now.

Netzsplits <-> QUIT-Meldung

(Report imported from Bugzilla #20)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.3.0 on platform All.
Assigned to: Alexander Barton.

On 2002-03-10 18:14:40 +0100, Georg Buschbeck wrote:

bei einem Netzsplit kommt als QUIT Meldung der User 'Ping-timeout' anstatt irgendwas, was auf einen Netzsplit hindeutet, auch kriegt man keine Info dar

On 2002-03-10 23:12:26 +0100, Alexander Barton wrote:

Ich habe das nun geaendert: in den Meldungen der User steht " | ",
wobei Server1 der nun nicht mehr erreichbare ist und Server2 der, zu dem Server1
bis gerade eben connectiert war.

Fails to build .deb on Debian stable

(Report imported from Bugzilla #56)

Status CLOSED, severity minor, in component Contrib.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

Original attachment names and IDs:

• ngircd.diff (ID 5)

On 2004-07-15 10:18:02 +0200, Kevin Otte wrote:

The control file as shipped in the last few versions fails to build a .deb package.

On 2004-07-15 10:19:12 +0200, Kevin Otte wrote:

Created attachment 5
Fix for control file

Simple fix to add the commas to the control file to build a proper Debian
package under stable.

On 2004-07-15 11:04:14 +0200, Alexander Barton wrote:

I applied your patch, both to CVS HEAD and branch-0-8-x. Thanks!

On 2004-10-14 00:20:36 +0200, Alexander Barton wrote:

Seems to work, I guess.

Unreg. Verbindungen: QUIT geht nicht

(Report imported from Bugzilla #21)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.4.0 on platform All.
Assigned to: Alexander Barton.

On 2002-04-07 19:48:53 +0200, Alexander Barton wrote:

Wenn eine Verbindung noch nicht registriert ist, d.h. noch kein NICK und USER
an den Server geschickt wurde, so akzeptiert letzterer den QUIT-Befehl nicht.

On 2002-04-14 16:04:46 +0200, Alexander Barton wrote:

Ok, sollte nun behoben sein, ein Patch ist ins CVS eingespielt. QUIT wird nun immer
akzeptiert, IRC_QUIT() unterscheidet nur noch zwischen "Server" und "nicht Server".

IRCop levels

(Report imported from Bugzilla #50)

Status RESOLVED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2004-01-30 23:25:06 +0100, Dobermann wrote:

I suggested to set IRCop levels, to allow some users to kill without possibilities of restarting the
daemon or /die it...

On 2005-11-07 03:01:13 +0100, De Zeurkous wrote:

In kxIRCd, i'm currently implementing a system of user modes for this purpose, only leaving the 'o' (and in channels the 'o' and 'v') modes intact, because of their widespread use in the same way. The modes will of course be very Korax-specific, but maybe such a system will work equally well for IRC in general too...

On 2005-12-09 10:40:29 +0100, Alexander Barton wrote:

I'll have a look at this and probably implement a "restricted operator" mode.

On 2006-03-17 15:19:41 +0100, Alexander Barton wrote:

Changed bug status to "LATER" because of lack of time.
Someone willing to take this over?

"-" wird in nicknames als illegales Zeichen betrachtet

(Report imported from Bugzilla #18)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.3.0 on platform All.
Assigned to: Alexander Barton.

On 2002-03-08 15:53:44 +0100, Georg Buschbeck wrote:

der gedankenstrich wird in nicknames nicht akzeptiert. vermutlich als illegal
betrachtet.

On 2002-03-10 02:11:34 +0100, Alexander Barton wrote:

Ich habe das heute im Zug von Freiburg nach Karlsruhe gefixed, werde es morgen noch
ein wenig testen und dann ins CVS einspielen. Im naechsten Release ist es damit dann
natuerlich auch drinnen.
BTW: fehlen noch andere Zeichen, die das Original zulaesst?

On 2002-03-10 18:22:15 +0100, Alexander Barton wrote:

Okay, Patch ist ins CVS eingespielt.

MODE +I <mask> generates duplicates

(Report imported from Bugzilla #52)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.7.7 on platform All.
Assigned to: Alexander Barton.

On 2004-04-06 23:39:08 +0200, Alexander Barton wrote:

When adding already existent invite masks using the "MODE +I " command
(and most probably ban mask, too!) to a channel, the daemon adds another
identical mask to the list.

Example for such a list:

a!~[email protected]
a!~[email protected]
x!y@z

Ugly :-)

On 2004-04-10 00:03:33 +0200, Alexander Barton wrote:

Hm, I can't reproduce this at the moment!?

Strange, I thought that I've seen such a list ...

Probably this has been an side effect of bug # 53.

On 2004-09-03 01:46:53 +0200, Alexander Barton wrote:

Seems to work.

Channel-Mode "i": Invite-Only

(Report imported from Bugzilla #23)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-05-24 17:26:06 +0200, Alexander Barton wrote:

Der Server sollte den Channel-Mode "i" (Invite Only) unterst

On 2002-06-03 01:32:13 +0200, Alexander Barton wrote:

Den INVITE-Befehl habe ich heute zum Teil implementiert: User
koennen nun in Channels eingeladen werden, Channels unterstuetzen
den Mode "i" auch bereits, ein Einladen in solche Channels ist
jedoch noch nicht implementiert (der Server kennt Invite-Lists
noch nicht ...).

On 2002-06-10 23:57:20 +0200, Alexander Barton wrote:

Nun kennt der Server auch Invite-Lists; User koennen mit dem INVITE-
Befehl in Channel mit dem Mode "i" eingeladen werden. Nach wie vor
koennen aber noch keine Invite-Lists mit dem MODE-Befehl gesetzt
werden, das kommt aber bald :-))

On 2002-08-27 11:16:44 +0200, Alexander Barton wrote:

Ich habe bei mir bereits eine Version, die INVITE-Lists

On 2002-09-17 00:40:53 +0200, Alexander Barton wrote:

ngIRCd unterst

On 2002-10-07 00:56:05 +0200, Alexander Barton wrote:

Ich habe keine negativen Meldungen bisher bekommen, ich setze den
Status daher mal auf "verified". Kommentare?

ngircd doesn't re-read ngircd.conf if you HUP it

(Report imported from Bugzilla #57)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.7.7 on platform other.
Assigned to: Alexander Barton.

On 2004-08-15 17:29:28 +0200, John Bleichert wrote:

Greetings.

ngircd doesn't seem to re-read its conf file when I HUP it. It re-reads its motd, but not the conf file.
We're seeing this with version 7.7 on FreeBSD 4.10 and version 8.0 on Slackware 10.

Not sure if it's a bug or by design, but wanted to let you know.

On 2004-08-15 17:37:04 +0200, Alexander Barton wrote:

Thanks for your report.

Do you get some messages regarding the "HUP" in the syslog?
And does the operator command "REHASH" work instead?

BTW: are you running ngIRCd inside a chroot environment?

And is the configuration file readable for the user you configured in the
configuration file, variable "ServerUID"?

On 2004-10-14 00:16:23 +0200, Alexander Barton wrote:

Hm, no new information and I can't reproduce this bug ...

On 2005-03-10 18:44:27 +0100, Alexander Barton wrote:

No new information. Closed.

Decompression error

(Report imported from Bugzilla #54)

Status CLOSED, severity critical, in component Daemon.
Reported in version 0.7.7 on platform All.
Assigned to: Alexander Barton.

Original attachment names and IDs:

• Inflate-Debug.patch (ID 4)

On 2004-04-12 13:53:30 +0200, Alexander Barton wrote:

The server disconnects a server-server-link immediately after switching to
compressed mode with the following error message:

Decompression error: code -3 (ni=2136969, ai=101, no=2130660, ao=2047)!?

This is caused by Unzip_Buffer() in src/ngircd/conn-zip.c, line 165.

"Code -3" stands for Z_DATA_ERROR, which is the result of inflate() instead of
Z_OK in this case. According to the zlib documentation
(http://www.gzip.org/zlib/manual.html#inflate), "inflate() returns [...]
Z_DATA_ERROR if the input data was corrupted (input stream not conforming to the
zlib format or incorrect adler32 checksum)".

Anthony, the first reporter of this behaviour, used ngIRCd 0.7.5 on Linux 2.4.17
with zlib 1.1.4 on the one end, and ngIRCd 0.7.0 on FreeBSD 5.1-REL with the
same version on zlib.

You can read the thread on the mailing list here:
http://arthur.ath.cx/pipermail/ngircd-ml/2003-December/thread.html, subject
"Decompression error".

On 2004-04-12 14:27:59 +0200, Alexander Barton wrote:

Created attachment 4
"enhanced" logging of the error message.

The attached patch extends the generated error message to include an error
description string of zlib.

On 2004-04-12 14:30:25 +0200, Alexander Barton wrote:

Heiko reported the second occurrence of this problem: here the peer server is
ircd 2.10.2, therefore it seems to be a bug in the locally running ngIRCd ...

On 2004-09-03 01:53:22 +0200, Alexander Barton wrote:

Hm, any news regarding this bug? If not I''ll close it with "works for me" ...

On 2004-11-10 22:10:10 +0100, Alexander Barton wrote:

No new reports, so I'll close this bug.

Hopefully we don't need to reopen it and it was some sort of black magic ...

On 2005-03-10 18:44:01 +0100, Alexander Barton wrote:

Finally closed.

Quit-Message to all channels

(Report imported from Bugzilla #48)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version development on platform All.
Assigned to: Alexander Barton.

On 2003-12-29 20:24:09 +0100, G wrote:

It would be nice to have a commandline switch posting a configurable message to all channels
and quitting afterwards.

Intention: Instead of throwing out all users when quitting ngircd (for updates) they get a message
announcing the downtime of the server.

Regards
G

On 2003-12-30 15:47:25 +0100, Alexander Barton wrote:

It would be possible (I think quite easy?) to enhance the DIE command (see RFC
2812, section 4.3) to accept an additional parameter as "die reason" which then
will be written to all channels.

On 2006-07-23 17:44:21 +0200, Alexander Barton wrote:

I enhanced DIE to allow a single parameter:

DIE []

This comment, if any, is sent to all locally connected clients
before the server goes down.

an oper /kill on himself crashes the daemon

(Report imported from Bugzilla #43)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.6.0 on platform PC.
Assigned to: Alexander Barton.

On 2003-01-16 15:04:17 +0100, Johann Beretta wrote:

I noticed that if I /oper myself, and then issue a /kill on myself the daemon
will crash.

I've reproduced it a couple of times. I'm running the daemon on Debian 3.0,
kernel 2.4.18 SMP with all current patches and updates.

I've never done a bug report before, so if you need any more information, or a
better explaination, please let me know.

AerynSun:~# ngircd --version
ngircd 0.6.0-SYSLOG+ZLIB+IRCPLUS-i586/pc/linux-gnu

On 2003-01-16 15:22:48 +0100, Alexander Barton wrote:

Thanks for your report!

This bug has been fixed 2002-12-26 in the current development tree (CVS-HEAD),
therefore won't be present in the next official release. Until then, you can
use a "nightly build" from here:

ftp://arthur.ath.cx/pub/Users/alex/ngircd/

Warnungen des gcc unter A/UX: "implicit declaration"

(Report imported from Bugzilla #11)

Status CLOSED, severity minor, in component Daemon.
Reported in version unspecified on platform Macintosh.
Assigned to: Alexander Barton.

On 2002-02-26 02:03:11 +0100, Alexander Barton wrote:

Unter A/UX werden beim Compilieren vom gcc einige Warnungen ausgegeben,
da

On 2002-02-26 02:48:24 +0100, Alexander Barton wrote:

Es handel sich um solche Meldungen:
channel.c:257: warning: implicit declaration of function `strcasecmp'

Und zwar fuer:
strcasecmp, gethostname, bzero, socket, setsockopt, bind, listen, select,
vsnprintf, send, accept, recv, connect, openlog, closelog und syslog.

Ich wuerde das gerne aendern, hoechste Prioritaet hat es aber nicht,
prinzipiell compilieren tut der ngIRCd unter A/UX schliesslich.

On 2002-03-31 14:55:10 +0200, Alexander Barton wrote:

Ein paar dieser Warnungen sind nun weg, seitdem unter A/UX _POSIX_SOURCE
definiert wird. Dazu teste ich im configure-Script explizit auf A/UX und fuege
-D_POSIX_SOURCE zu den CFLAGS hinzu.
Bei den nun noch verbliebenen Warnungen habe ich noch nicht weiter untersucht,
ob und wo die entsprechenden Prototypen stecken ...

On 2002-10-05 02:55:34 +0200, Alexander Barton wrote:

Vermutlich sind das (zumindest u.a.) Prototypen der Funktionen, die "nativ"
unter A/UX fehlen und z.B. aus der libUTIL.a kommen.

On 2004-04-06 23:49:49 +0200, Alexander Barton wrote:

Most (all!?) of this messages should be gone when using the new build scripts
(mainly the new ./configure) as of the actual CVS version; these scripts will be
included starting with version 0.8.0.

On 2010-06-25 00:09:05 +0200, Alexander Barton wrote:

No one tested recent ngIRCd releases on A/UX

Channel-Ops werden nach Server-Joins nicht als solche angezeigt

(Report imported from Bugzilla #24)

Status CLOSED, severity major, in component Daemon.
Reported in version 0.4.3 on platform All.
Assigned to: Alexander Barton.

On 2002-08-07 13:27:29 +0200, Marek Czernohous wrote:

Die Korrektur der User-Modes nach verbinden von Servern untereinander scheint
noch nicht zu klappen. Ein Op eines lokalen Servers wird zwar nach dem join in
den

On 2002-08-27 11:14:41 +0200, Alexander Barton wrote:

Das korrekte Verhaltem ist, da

On 2002-08-27 16:12:40 +0200, Alexander Barton wrote:

Ich habe soeben einen Fix ins CVS eingespiel, in ersten Tests l

On 2002-08-27 16:15:34 +0200, Alexander Barton wrote:

BTW: das Problem bestand nicht bei NJOIN, sondern bei JOIN (irc-channel.c) ...

On 2002-09-17 00:43:53 +0200, Alexander Barton wrote:

Scheint behoben zu sein, der Bug trat in den letzten Tagen nicht mehr auf.

Nick- und Channel-Names werden nicht genuegend validiert

(Report imported from Bugzilla #12)

Status CLOSED, severity critical, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-02-27 15:37:34 +0100, Alexander Barton wrote:

Nicknames (bei NICK) und Channel-Names (bei JOIN) werden nicht genuegend validiert:
so ist es moeglich, ungueltige Namen zu verwenden, mit denen die Server nicht umgehen
konnen und so das IRC-Netz "out of sync" zu bringen.
Die entsprechenden Funktionen Channel_IsValidName() und Client_IsValidNick() werden
zwar bereits aufgerufen, pruefen jedoch zu wenig ...

On 2002-03-02 02:40:22 +0100, Alexander Barton wrote:

Ich habe die beiden Funktionen Channel_IsValidName() und Client_IsValidNick()
ergaenzt: nun sollten sie ordentlich validieren.

MOTD funktioniert nur lokal

(Report imported from Bugzilla #36)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.5.3 on platform All.
Assigned to: Alexander Barton.

On 2002-11-22 23:31:06 +0100, Alexander Barton wrote:

der MOTD-Befehl kann noch nicht an andere Server weitergeleitet werden,
wie es im RFC 2812 vorgesehen ist:

3.4.1 Motd message

  Command: MOTD

Parameters: [ ]

On 2002-11-24 17:36:09 +0100, Alexander Barton wrote:

In CVS-HEAD habe ich soeben einen Patch eingespielt, mit dem MOTD
an andere Server delegiert werden kann.

On 2002-11-30 02:19:27 +0100, Alexander Barton wrote:

Funktioniert in meinen Tests gut.

ngIRCd assert()'s in conn.c, line 620

(Report imported from Bugzilla #51)

Status CLOSED, severity critical, in component Daemon.
Reported in version 0.7.7 on platform All.
Assigned to: Alexander Barton.

Original attachment names and IDs:

• Conn_Close.patch (ID 3)

On 2004-04-06 23:28:55 +0200, Alexander Barton wrote:

I received a bug report that ngIRCd 0.7.7 assert()'s in conn.c, line 620 with
the following message in the /tmp/ngircd-XXX.err file:

assertion "My_Connections[Idx].sock > NONE" failed: file "conn.c", line 620

Thanks, Anthony!

On 2004-04-06 23:31:21 +0200, Alexander Barton wrote:

Created attachment 3
Proposed patch

Proposed patch after having a quick look at it ...

It seems as if ngircd is calling recursively Conn_Close() with an already
closed socket. This patch should detect this situation and return before
checking the socket handle for validity.

On 2004-09-03 01:51:34 +0200, Alexander Barton wrote:

Patch has been applied to both CVS HEAD and branch-0-8-x some weeks (months?) ago. Seems to
work ...

On 2004-10-14 00:19:46 +0200, Alexander Barton wrote:

No new reports, seems to work just fine now.

Evaluate "Modes" in "[Channel]" sections like MODE command

(Report imported from Bugzilla #55)

Status RESOLVED, severity normal, in component Daemon.
Reported in version 0.7.7 on platform All.
Assigned to: Alexander Barton.

On 2004-06-11 15:30:39 +0200, Alexander Barton wrote:

Actually the server doesn't evaluate the "Mode" configuration option in
"[channel]" sections like the MODE command: it only sets known "mode
characters". Therefore it isn't possible to set channel limits (mode "l") and
channel keys ("k").

On 2005-02-09 10:40:00 +0100, Alexander Barton wrote:

... and it doesn't enforce predefined modes if the channel exists already.

On 2007-06-13 17:13:47 +0200, Alexander Barton wrote:

Starting with ngIRCd 0.10.2 one can specify channel keys and user limits using the new configuration
variables "key" and "maxusers" in [channel] sections.

This is a workaround, but has been much easier to implement.

Probably "someone" will fix this in the future ...

Bans (+b) do not behave as expected.

(Report imported from Bugzilla #47)

Status CLOSED, severity critical, in component Daemon.
Reported in version 0.7.1 on platform All.
Assigned to: Alexander Barton.

Original attachment names and IDs:

• BanPrivmsg.patch (ID 2)

On 2003-10-10 22:11:05 +0200, Hannes Dorbath wrote:

The only thing +b seems to do on a channel is blocking the user from rejoining
after he/she was kicked-banned out. The person is still allowed to send text to
the channel (from outside) when the channel is not +n and is even still allowed
to talk on a channel when he/she was banned and not kicked out. The last one
should only be possible when he/she got voice or op on the channel.

A behaviour as described in
http://www.irchelp.org/irchelp/misc/ccosmos.html#Heading790 should be desirable.

Thanks for listening, keep up the good work.

On 2003-10-20 00:29:25 +0200, Alexander Barton wrote:

You are right, the actual implementation isn't complete and needs to be fixed ...

Thanks for your report, I'll have a look at this!

On 2003-11-06 02:19:12 +0100, Alexander Barton wrote:

Created attachment 2
Patch for 0.7.1 to solve this bug.

This patch (against CVS-HEAD as of 2003-11-06, 02:00 CET, but should work
against 0.7.1, too) should solve the bug: no one is allowed to PRIVMSG to a
channel he is banned in.

On 2003-11-06 02:22:48 +0100, Alexander Barton wrote:

I checked in the patch to CVS HEAD, it will be included in the next release of
ngIRCd which will be released soon.

It didn't change the QUIT, PART or NICK behavior as suggested by the irchelp.org
article, but I don't think that this is really required.

On 2003-11-13 12:24:40 +0100, Alexander Barton wrote:

Okay, seems to work as expected in version 0.7.5.

On 2003-12-08 15:08:57 +0100, Alexander Barton wrote:

Seems to work, I'll close the bug now.

Bei PRIVMSG werden Channel-Modes nicht geprueft

(Report imported from Bugzilla #13)

Status CLOSED, severity major, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-02-27 15:39:37 +0100, Alexander Barton wrote:

Bei PRIVMSG mit einem Channel als Ziel prueft der Server nicht, ob der Client
ueberhaupt in den entsprechenden Channel schreiben darf (die Channel-Flags
"n" und "m" werden z.B. ignoriert).

On 2002-02-27 18:14:36 +0100, Alexander Barton wrote:

ngIRCd beachtet nun bei PRIVMSG die Channel-Modes "n" und "m".

Komprimierte Server-Server-Links

(Report imported from Bugzilla #41)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-11-25 12:42:01 +0100, Alexander Barton wrote:

Der ngIRCd k

On 2002-11-27 01:02:17 +0100, Alexander Barton wrote:

Ich habe gerade eben eine entsprechende Erweiterung in CVS-HEAD eingespielt:
das implementierte Verfahren ist kompatibel zum Original ircd 2.10.3, verwendet
wird die zlib (www.zlib.org).

Unterst

On 2002-12-18 13:23:35 +0100, Alexander Barton wrote:

Seems to work.

ngircd doesn't work with dircproxy

(Report imported from Bugzilla #26)

Status CLOSED, severity normal, in component Daemon.
Reported in version unspecified on platform PC.
Assigned to: Alexander Barton.

On 2002-09-24 22:17:30 +0200, Kevin Fenzi wrote:

Greetings. Hopefully you can read this report in english. (I don't know german)

ngirc doesn't appear to work with dircproxy. http://www.dircproxy.net/
It looks like it's a problem with something the server sends on connect...
It connects, but won't let any commands work.

It may be related to this message:

[error] The following line is in violation of the IRC protocol.
[error] Please tell the server administrator:
my.ircserver.example.com: I have 6 users, 0 services and 1 servers

On 2002-09-25 00:37:52 +0200, Alexander Barton wrote:

I'll have a look at this problem. Because I used dircproxy a few months
ago by myself, it must be around here :-)

But I can't see a problem with the "I have"-line: the ircd 2.10.3p3
sends this status in exactly the same format.

Perhaps I should have a look at the sources of dircproxy to see what it
expects to receive ...

On 2002-09-27 17:00:07 +0200, Alexander Barton wrote:

Okay, it doesnt work here either. Good for you :-)

But my version of dircproxy (1.0.2) seems not to log anything!?
And I can't find a configuration option to change that ...

I'll try it with a newer version later.

On 2002-10-03 19:10:14 +0200, Alexander Barton wrote:

The server message RPL_MYINFO_MSG (004) had a bad syntax that confused
dircproxy. This is fixed in CVS and ngIRCd 0.5.1 which is available now.

On 2002-10-07 00:54:20 +0200, Alexander Barton wrote:

Kevin, did you have time to check this agin?

Does it work for you?
Can I colose the bug?

On 2002-10-07 03:08:28 +0200, Kevin Fenzi wrote:

Yes, it works great now!
Thanks for tracking down the problem.

Just add ssl support and channel keys and ngircd will be all I want in a irc
server. ;)

On 2002-10-09 11:20:30 +0200, Alexander Barton wrote:

Hint: You can file bugs for feature enhancements, too ;-))

(set the "severity" in Bugzilla to "enhancement)

INVITE propagation

(Report imported from Bugzilla #53)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.7.7 on platform All.
Assigned to: Alexander Barton.

On 2004-04-06 23:42:39 +0200, Alexander Barton wrote:

Some people reported that they don't receive INVITE's of other users. Until now,
I didn't have the time to look closely at this, but my "feeling" is that the
server has problems routing such messages via more than one hop.

But this definitely needs more investigation ...

On 2004-04-09 23:49:29 +0200, Alexander Barton wrote:

Okay, the problem wasn't that the INVITE won't be forwarded to other servers,
but a) that "one-shot" invite list entries generated by INVITE won't be deleted
-- after the user has joined to the channel -- on servers except the one the
user is connected to and b) already existent ban and invite list entries won't
be propagated again to other servers when re-calling INVITE or MODE +b/+I.

On 2004-09-03 01:47:37 +0200, Alexander Barton wrote:

Seems to work now.

Rechtschreibfehler

(Report imported from Bugzilla #25)

Status CLOSED, severity normal, in component Web-Page (german).
Reported in version unspecified on platform Macintosh.
Assigned to: Alexander Barton.

On 2002-08-27 17:05:48 +0200, Ilja Osthoff wrote:

Bei dem Hinweis auf den Bugzilla, steht "finde" statt finden...

On 2002-08-27 17:07:57 +0200, Alexander Barton wrote:

Behoben, Danke! :-)

JOIN only supports one key at a time

(Report imported from Bugzilla #42)

Status CLOSED, severity minor, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-12-17 00:21:48 +0100, Alexander Barton wrote:

The JOIN command supports a list of channels but only one key that must be
equal for all channels. This is not compliant to RFC 2812, section 3.2.1.

On 2005-08-15 13:41:47 +0200, Hannes Dorbath wrote:

Any chance to get that fixed sometime soon? It may not be a problem for many
people, but it's a blocker for us, as we use your IRCD as a support system for
customers and all channels are +k.

Thanks

On 2005-08-25 10:32:27 +0200, Alexander Barton wrote:

Now that I know that someone would use this feature I'll have a look at it :-)

On 2005-09-02 17:49:44 +0200, Alexander Barton wrote:

The CVS-HEAD version as of now supports more than one channel key at a time when joining multiple
channels.

Will be included in the next "major" release of ngIRCd.

The patch is available via Web-CVS:
<http://arthur.ath.cx/cgi-bin/viewcvs.cgi/ngircd/ngircd/src/ngircd/irc-channel.c?
sortby=date&r2=1.32&r1=1.31>

On 2005-11-07 13:55:31 +0100, Alexander Barton wrote:

Seems to work just fine -- I'll close this bug :-)

NJOIN mit verstuemmelten Nick-Names

(Report imported from Bugzilla #17)

Status CLOSED, severity critical, in component Daemon.
Reported in version 0.3.0 on platform PC.
Assigned to: Alexander Barton.

Original attachment names and IDs:

• syslog.txt (ID 1)

On 2002-03-05 00:33:23 +0100, Alexander Barton wrote:

Unter nicht ganz klaren Umstaenden korrumpiert der ngIRCd entweder beim Verschicken
oder Empfangen von NJOIN-Befehlen den Nick des jeweiligen Clients. Danach beendet sich
der ganze Daemon stillschweigend ohne weitere Meldungen im Syslog (die korrupten Nicks
sind im Syslog zu sehen, siehe Attachment!).

On 2002-03-05 00:34:23 +0100, Alexander Barton wrote:

Created attachment 1
Syslog-Meldungen mit korrumpierten Nicks

On 2002-03-05 00:54:18 +0100, Alexander Barton wrote:

Das ist schwerwiegend. Keine Frage. Leider kann ich es im Moment nicht nachvollziehen
oder gar das Problem im Source erkennen. Ich werde die vermutlich betroffenen Funktionen
aber nochmal durchsehen ...

On 2002-03-10 02:20:11 +0100, Alexander Barton wrote:

Hm, kann ich nach wie vor nicht nachvollziehen, im Source kann ich so auch
keinen Bug sehen ... bitte auf jeden Fall wieder melden, sollte soetwas erneut
auftreten!

STATS wird nicht unterst

(Report imported from Bugzilla #37)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version 0.5.3 on platform All.
Assigned to: Alexander Barton.

On 2002-11-22 23:46:26 +0100, Alexander Barton wrote:

Der ngIRCd unterst

On 2002-11-23 00:25:58 +0100, Alexander Barton wrote:

Ein Bug f

On 2002-12-02 15:02:13 +0100, Alexander Barton wrote:

Der ngIRCd kennt nun den STATS-Befehl, im Moment allerdings nur die
Unterfunktion "l": damit werden Daten zu aktiven Verbindungen angezeigt.

BTW: der vorherige Kommentar h

On 2002-12-26 19:07:50 +0100, Alexander Barton wrote:

"STATS m" is supported in 0.6.0, too.

Server blockiert bei Connect zu unerreichbarem Host

(Report imported from Bugzilla #16)

Status CLOSED, severity blocker, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-02-28 15:03:00 +0100, Alexander Barton wrote:

Wenn der Server eine Verbindung zu einem anderen Server aufbauen will,
dieser jedoch nicht erreichtbar ist (also auch kein "refused" liefert),
so blockiert der ganze ngIRCd fuer eine laengere Zeit, bis ein Timeout
eintritt (ggf. mehrere Minuten!).

Diese Situation ist bei Netzen, zu denen Server mit dynamischen Adressen
gehoeren, jedoch nicht untypisch.

Da so der gesammte Server unbenutzbar wird, mu

On 2002-03-02 01:50:47 +0100, Alexander Barton wrote:

Das connect() wird nun asyncron ausgefuehrt, das Problem sollte damit behoben sein.

Ein anderes habe ich ebenfalls gleich behoben: bei allen ausgehenden Verbindungen wurden
keine Socket-Options gesetzt ...

Befehl "opper" noch nicht implementiert

(Report imported from Bugzilla #10)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.2.1 on platform All.
Assigned to: Alexander Barton.

On 2002-02-24 13:34:03 +0100, Marek Czernohous wrote:

"opper unknown command"

On 2002-02-24 15:15:10 +0100, Alexander Barton wrote:

Der Befehl heißt OPER :-)

In ServerUID und ServerGID k

(Report imported from Bugzilla #30)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-10-13 05:02:40 +0200, Alexander Barton wrote:

In den beiden Konfigurationsvariablen ServerUID und ServerGID k

On 2002-11-09 00:30:43 +0100, Alexander Barton wrote:

Soeben habe ich einen entsprechenden Fix in CVS-HEAD eingespielt.

On 2002-11-30 02:17:16 +0100, Alexander Barton wrote:

Funktioniert bisher, wie es soll.

Channel-Logging

(Report imported from Bugzilla #32)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-11-21 15:01:52 +0100, Alexander Barton wrote:

Der Server sollte das Loggen von Channeln unterst

On 2002-11-22 20:50:26 +0100, Alexander Barton wrote:

Auch wenn das Logging f

On 2002-11-22 22:56:11 +0100, Alexander Barton wrote:

Hmpf. Die Idee hat mir zwar gut gefallen, funktioniert aber nicht:

Server bekommen nur Meldungen geschickt, wenn sie einen User haben, f

Reload der Konfigurationsdatei zur Laufzeit

(Report imported from Bugzilla #33)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-11-21 15:04:06 +0100, Alexander Barton wrote:

Der Daemon sollte zur Laufzeit die Konfigurationsdatei neu einlesen k

On 2002-11-22 19:07:22 +0100, Alexander Barton wrote:

Ich habe soeben einen entsprechenden Patch in CVS-HEAD eingespielt: der
Server kennt nun den Befehl RELOAD und liest die Konfigurationsdatei
danach neu ein. Ebenso, wenn ein HUP-Signal empfangen wird.

Es k

On 2002-12-18 13:22:19 +0100, Alexander Barton wrote:

Seems to work quite well ;-)

Server doesn't validate server name

(Report imported from Bugzilla #45)

Status CLOSED, severity major, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2003-05-08 14:48:33 +0200, Alexander Barton wrote:

The daemon doesn't validate the IRC network name of the server (enough);
So you can configure "Name = Invalid !"%&/

On 2004-11-10 22:13:36 +0100, Alexander Barton wrote:

Probably we could use the Match() function (file match.c/match.h) and a "simple" pattern to validate the server name? We'll see ...

On 2006-07-23 18:43:44 +0200, Alexander Barton wrote:

I checked in code to conf.c that sould resolve this issue.

redundante Server-Verbindungen

(Report imported from Bugzilla #14)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-02-28 03:01:15 +0100, Holger wrote:

M

On 2002-02-28 03:16:14 +0100, Alexander Barton wrote:

Klingt gut. Wenn der erste Server nicht connectiert werden kann, so wird der naechste
versucht. Da jetzt schon beliebig viele Server definiert werden koennen, braechte es in
der Konfigurationsdatei "nur" noch einer Moeglichkeit, sie in Gruppen zusammen fassen
zu koennen (man kann schliesslich auch n unabhaengige Server haben, zu denen sich der
ngIRCd connectieren soll).

Im Server selber sollte das kein all zu grosser Aufwand sein: Check_Servers() in conn.c
muesste einfach den naechsten in der Gruppe probieren, wenn der letzte als "versucht"
markiert ist.

Werde ich bei Gelegenheit umsetzen.

On 2002-03-10 02:14:30 +0100, Alexander Barton wrote:

Habe ich heute -- wie Bug 18 -- im Zug nach Karlsruhe mal angegangen. Muss ich
allerdings morgen noch testen. Wenn es morgen(?) dann laeuft, spiele ich es ins CVS
ein. Denke mal, dass es damit auf jeden fall im naechsten Release enthalten sein wird.

On 2002-03-10 19:48:31 +0100, Alexander Barton wrote:

Ist nun implementiert, jedoch noch nicht all zu ausfuehrlich getestet.
Die Beispiel-Konfiguration habe ich entsprechend erweitert.

On 2002-03-31 15:25:05 +0200, Alexander Barton wrote:

Nach meinen -- wenn auch kleinen -- Tests, funktioniert das tatsaechlich :-)

On 2002-05-08 14:07:58 +0200, Alexander Barton wrote:

Wird wohl noch nirgends wirklich eingesetzt,
scheint aber zu funktionieren ...

Channel Keys ("Channel Pa

(Report imported from Bugzilla #31)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2002-11-08 23:51:56 +0100, Alexander Barton wrote:

Der Server sollte Pa

On 2002-12-17 00:16:48 +0100, Alexander Barton wrote:

I just added the channel modes k ("key") and l ("member limit") to CVS-HEAD.
Both modes will be available in the upcoming 0.6.0 release.

On 2003-01-27 17:02:45 +0100, Alexander Barton wrote:

Seems to work.

Administrative interface

(Report imported from Bugzilla #49)

Status RESOLVED, severity enhancement, in component Daemon.
Reported in version unspecified on platform All.
Assigned to: Alexander Barton.

On 2003-12-30 15:40:09 +0100, Alexander Barton wrote:

It would be handy to have an "administrative" interface to control the daemon:
e. g. restart and shutdown it, probably list (and kill?) users, and get statistics.

Most of this actions can be performed using regular IRC commands an operator
privileges, but I think it would be a "good thing" to have a simpler interface:
statistics should be updated by the server and not be polled by the "admin client".

On 2006-03-17 15:16:48 +0100, Alexander Barton wrote:

Changed bug to "LATER", because of lack of time :-/
Someone willing to take this over?

NICK wird schon vor USER an andere Server geforwarded

(Report imported from Bugzilla #15)

Status CLOSED, severity normal, in component Daemon.
Reported in version 0.2.1 on platform All.
Assigned to: Alexander Barton.

On 2002-02-28 14:36:54 +0100, Alexander Barton wrote:

Bekommt ein Server von einem User ein NICK, so wird dieses bereits vor dem
Empfang des USER-Befehls vom User an andere Server weitergeleitet. Dieses
lehnen es dann korrekt mit ERR_NOSUCHNICK (401) ab.

Nachdem der Server das passende USER dann empfangen hat, wird das korrekte
(fuer Server-Server-Links erweiterte) NICK an alle anderen Server im Netz
verschickt, so dass die eigentliche User-Registrierung dennoch klappt.

On 2002-02-28 14:37:25 +0100, Alexander Barton wrote:

Der Fehler liegt in irc-login.c, Funktion IRC_NICK() und kann wohl relativ
einfach behoben werden. Ich werde das noch vor dem naechsten Release fixen.

On 2002-03-02 01:52:42 +0100, Alexander Barton wrote:

Habe ich behoben. Das falsche erste NICK wird nicht mehr an andere
Server geforwarded.

TIME wird nicht unterst

(Report imported from Bugzilla #38)

Status CLOSED, severity enhancement, in component Daemon.
Reported in version 0.5.3 on platform All.
Assigned to: Alexander Barton.

On 2002-11-23 00:10:57 +0100, Alexander Barton wrote:

Der Server kennt den TIME-Befehl noch nicht. RFC 2812:

3.4.6 Time message

  Command: TIME

Parameters: [ ]

On 2002-11-24 19:48:01 +0100, Alexander Barton wrote:

TIME ist nun implementiert (Patch ist in CVS-HEAD). Auch das forwarden an
andere Server sollte funktionieren.

On 2002-11-30 02:18:41 +0100, Alexander Barton wrote:

Scheint zu funktionieren.