- Server protocol: added PROTOCTL EATH=servername, which allows us to
authenticate the server very early in the handshake process. That way,
certain commands and PROTOCTL tokens can 'trust' the server.
See doc/technical/protoctl.txt for details.
- Server protocol: between new Unreal servers we now do the handshake a
little bit different, so it waits with sending the SERVER command until
the first PROTOCTL is received. Needed for next.
- Server protocol: added PROTOCTL SERVERS=1,2,3,4,etc by which a server can
inform the other server which servers (server numeric, actually) it has
linked. See doc/technical/protoctl.txt and next for details.
- When our server was trying to link to some server, and at the same time
another server was also trying to link with us, this would lead to a
server collision: the server would link (twice) ok at first, but then a
second later or so both would quit with 'Server Exists' with quite some
mess as a result. This isn't unique to Unreal, btw.
This happened more often when you had a low connfreq in your link blocks
(aka: quick reconnects), or had multiple hubs on autoconnect (with same
connfreq), or when you (re)started all servers at the same time.
This should now be solved by a new server handshake design, which detects
this race condition and solves it by closing one of the two (or more)
connections to avoid the issue.
This also means that it should now be safe to have multiple hubs with low
connfreq's (eg: 10s) without risking that your network falls apart.
This new server handshake (protocol updates, etc) was actually quite some
work, especially for something that only happened sporadically. I felt it
was needed though, because (re)linking stability is extremely important.
This new feature/design/fix requires extensive testing.
This feature can be disabled by: set { new-linking-protocol 0; };
such as ~q:~c:#test to only silence users on #test, for example. This feature
is enabled by default, but can be disabled during ./Config -advanced. Module
support for this feature must note the following:
- For is_ok function, the extban can either assign extban_is_ok_nuh_extban, which
will deal checking a chained extban (including checking for restricted extbans),
or it can call that function from its own is_ok routine. For the latter case,
remember to pass only the mask part of your ban format (ie, don't just pass para as
otherwise it'll just call your is_ok again).
- For conv_param function, the extban can either assign extban_conv_param_nuh_or_extban,
which will automatically call conv_param for a chained extban, or pretty up a n!u@h mask.
- For is_banned, the extban should call ban_check_mask with the mask part of the parameter.
This will automatically call is_banned for a stacked extban, or match against a n!u@h. n!u@h
is checked against the current user (ie, with the info in the globals ban_ip, etc), so things
can get weird if you call this outside a normal ban check.
Modules must keep in mind that chained extban support is not available (and neither are the three
functions above) if DISABLE_STACKED_EXTBANS is #defined (this is controled by Config). Modules will
not compile/load if they try to use them anyway.
This change should not break extban modules, and should need some more extensive testing.
- Misc fix for disabling extban chains, should've done stuff in our autoconf
stuff instead of hacking configure directly :P .
load (for example when the webserver is down), then the most recent
version of that remote include will be used, and the ircd will still boot
and be able to rehash. Even though this is quite a simple feature, it
can make a key difference when deciding to roll out remote includes on
your network. Previously, servers would be unable to boot or rehash when
the webserver was down, which would be a big problem (often unacceptable).
The latest version of fetched urls are cached in the cache/ directory as
cache/<md5 hash of url>.
Obviously, if there's no 'latest version' and an url fails, the ircd will
still not be able to boot. This would be the case if you added or changed
the path of a remote include and it's trying to fetch it for the first time.
To disable this new behavior, check out REMOTEINC_SPECIALCACHE in
include/config.h.
In the IRCd world correct time is very important. This means that time
should be correct when the IRCd is booted, either by running ntpd/ntpdate
on the system or some other synchronization software, or by using the
built-in timesync feature.
Whenever the clock is adjusted for more than a few seconds AFTER the IRCd
has booted, it can lead to dangerous effects ranging from unfair
timestamps
for nicks and channels (and hence the possibility to takeover channels),
to even completely stalling the IRCd (negative timeshift) or making it so
nobody can connect anymore due to throttling (positive timeshift).
We now try to 'fix' the worst effects such as the IRCd freeze and
throttling. This does not fix the whole problem, so I've added some big
warnings when the clock is adjusted, including an annoying one every 5
minutes if the clock was set backwards, until the time is OK again
(catches up with the original time).
This fixes#0003230 reported by Stealth, and #0002521 reported by durrie.
user target string (nick!user@host:info), insteaf of doing it at like 5 places.
- Spamfilter target 'u' (user): the host field (nick!user@HOST:realname) is now escaped
with brackets if it's an IPv6 address, eg: blah!blah@[1:2:3:4:5:6:7:8]:hello, reported
by aquanight and others (#0003010).
map the command '/BLAH 5' to 'NICK idiot5'. More info in docs on alias block.
- Modulized: badwords system (src/badwords.c is now gone) and StripColors/StripControlCodes
to m_message, multiple netsynch routines to m_server, send_list to m_list, a certain mode
routine to m_svsmode, all /MSG IRC.. webtv stuff to src/modules/webtv.c which is compiled
with m_message.
This means another ~1500 lines of code are now in modules (and thus can be upgraded on
the fly), which brings the total of modulized lines at 32K.
CALLBACKTYPE_CLOAK). This passes 'aClient *sptr, char *host' instead of only 'char *host'
to the cloaking module, which can be useful if you need to cloak on something other than
IP/host. Suggested by fez (#0002275).
Module may still provide only CALLBACKTYPE_CLOAK though, in fact this is what the official
cloaking module does. So no updating of cloaking modules needed.
A side-effect of this "extra cloaking" callback is that we needed to change make_virthost()
which now has an extra parameter in front, and another side-effect is that calling the
CALLBACKTYPE_CLOAK may not work since only *_EX might be available. To my knowledge there
are very few modules (only 1 I know) that will have a problem due to this, so sounds like
an affordable tradeoff.
a lot of crashes. Both are now fixed. Reported by Zell, Yamake, and others (#2875, #2704).
Fix provided by Xuefer. This also gets rid of some annoying and useless compile warnings
as well.
- When checking if a user is banned, we always check the cloakhost too. Previously we could
not do this if the user had a /VHOST (=a minority of the cases, but still...). In short,
this is some extra protection to combat ban evasion.
- Performance of is_banned() *slightly* improved (just 1-2 usec, but 7 usec if no bans).
- [Module coders] For extban routines, we now offer a routine extban_is_banned_helper(buf)
which can be used instead of the ban_realhost/etc static chars stuff, see
extban_modeq_is_banned for a (real-life) example of how this is used.
- [Services coders!] Added PROTOCTL CLK (requires NICKv2) which adds an extra field in the
NICK command (when a user connects) right before the infofield (gecos).
The added field contains the cloaked host, that is: the masked host if +x would have been
set. This field is ALWAYS sent, regardless of whether the user is actually +x or not.
Services can then store this field in memory, to know the host of the user if the user
is set +x (+x-t). This is a (better) alternative to PROTOCTL VHP, with no race conditions,
and avoids some other VHP problems.
VHP will stay supported though... so it's not mandatory to switch over.
- '/quote dns i' now shows the nameserver settings (which is taken from /etc/resolv.conf
on *NIX, and from the registry on Windows)
- We no longer depend on a C++ compiler (was useless c-ares dependency caused by libtool)
- '/REHASH -dns' now rereads the resolver data from resolv.conf/registry, no IRCd restart
needed anymore. It's currently kinda experimental however, but I *think* it will work ok.
Unfortunately the above features required some ugly hacks if curl was enabled, so if you
use curl (Remote includes), feel free to test on your OS (Linux, but especially FreeBSD
and the other *NIXes) to see if things still compile (make clean; ./Config && make).
clients connecting trough a CGI:IRC gateway that is in cgiirc { }. This might also fix a bug
where (g)zlines were not applied to CGI:IRC clients, reported by devil (#0002850).
- Modulized NAMES command (can now be upgraded on the fly, if ever needed).
- Added NAMESX support, seeing both mIRC (5.17) and XChat support this. What this does is
send all rights of all users on the channel in the NAMES reply (eg: @+Syzop if the user is +ov)
instead of only the highest one (@Syzop in previous example). We only do so if the client
explicitly requested this via a NAMESX in a PROTOCTL message (eg: 'PROTOCTL NAMESX').
Note that there is a glitch: since most clients only send the PROTOCTL NAMESX after they
see NAMESX listed in the 005 announce message this has the effect that if there are
set::auto-join channels present (where users are automatically joined to by the server) the
extended NAMES reply will not be sent for those channels, because from the IRC server' point
of view the join happened before the PROTOCTL and hence it does not know the client wanted
NAMESX at that point (the result is not catastrophic: the old-style NAMES is sent for those
channels). Anyway, for all non-autojoin channels this works great. So still worth adding IMO.
Originally suggested in #0000606.
Side note: this does not mean we dropped the idea of (also) having a challenge-response
system for good ;).
We now support the webirc ('webirc_password' in CGI:IRC) method, which is kinda superior
to the older method ('realhost_as_password').
See the Unreal documentation (section '4.36 - Cgiirc Block') for details on how to configure.
- Changed quoting color in unreal32docs.. looks better now IMO (only English docs updated).
"trusted" and the IRCd will show the users' _real_ host/ip everywhere on IRC, instead of the
_CGI:IRC-gateway_ host/ip.
To do so you must set 'realhost_as_password' to 1 in your cgiirc.conf. And add the
CGI:IRC gateway(s) you fully trust to set::cgiirc::hosts.
means no longer weird issues with +b *\* etc not banning nicks with \ in it.
ExtBan ~c/~r get special treatment and will use our match_esc [match with escaping]
routine, that way you can ban channels such as "#f*ck" via "+b ~c:#f\*ck".
Fix triggered by bugreport of vonitsanet (#0002782).
- Redid some net synching code to make it more efficient (#2716).
- Fixed spamfilter crash problem: the action 'viruschan' is now no longer incompatible
with target 'user'. Reported by Monk (#0002570).
This should get rid of some annoying untracable (and usually rare) crashbugs in the
old resolver. Besides that, it makes things look more clean and understandable.
This should be the fix for the following bugids (all the same issue): #2499, #2551, #2558,
#2559, #2603, #2642, #2502, #2501, #2618, #2616.
Feedback and testing is very much welcomed (syzop@unrealircd.com).
It also sends a numeric to the user saying the command has been processed, but a copy
has been sent to ircops. I feel this is a good idea for privacy reasons (anti-spy),
though I don't know how users will react to this. If you are using this on your network
and get users bothering you about it (or before that ;p), it's probably a good idea
to explain it somewhere on your site or FAQ :).
Example usage:
/spamfilter add p warn - Testing_mirc_decode_filter \$decode\(.*\)
[WARNING] The numeric text is likely to change in the next few weeks (early-cvs-commit).
- If a class block was removed and any other blocks would be referencing the class block
(such as: allow::class, oper::class, link::class), then this would cause a crash.
Reported by Mike_ (#0002646).
and 2000 lines total that can be hotfixed if needed ;). The effort involved in moving all
this sucks a lot though :/. This might need some more testing to make sure it doesn't break
anything.
- Updated support OS list in documentation.
- Build in some additional checks (especially for Chinese).
- Fixed a bug in chinese character range (affecting 3.2*)
- Relaxed nick character checking from remote servers (rely on NICKCHARS= PROTOCTL
to deal with problems). This is useful to prevent any kills in case we slightly
change the characters that are allowed in a language.
If 2 servers try to link and the allowed nick characters do not fully match, then
the link will be rejected. Note that this will not prevent you from 3.2.2<->3.2.3/CVS
charsets mistakes, but only with linking CVS/3.2.3+ servers. Suggested by Troco (#0002360)
This might need some additional testing, but initial results are positive :).
still cutoff if the nick is too long. Basically this is the same way as Hybrid does it
so it should work ok :).
- Added nick character system. This allows you to choose which (additional) characters
to allow in nicks via set::allowed-nickchars. See unreal32docs.html -> section 3.16
for a list of available languages and more info on how to use it.
Current list: dutch, french, german, italian, spanish, euro-west, chinese-trad,
chinese-simp, chinese-ja, chinese.
If you wonder why your language is not yet included or why a certain mistake is present,
then please understand that we are most likely not experienced (at all) in your language.
If you are a native of your language (or know the language well), and your language
is not included yet or you have some corrections, then contact syzop@vulnscan.org or
report it as a bug on http://bugs.unrealircd.org/
throttle the number of joins per-user to X in Y seconds. Idea from Angrywolf (who
wrote a module that did this before). This might need some more testing :).
It's enabled by default but can be #undef'ed in include/config.h (line 449).
(perhaps this should be a different function?). Anyway, this means less diskspace
is needed (~1.5mb or more), and it also makes it a bit easier for RBAC (#2300).
- Made a new function DoMD5() which is ssl/non-ssl independent. Also made the cloaking
module and the auth functions use it. Hopefully I didn't break anything ;). Suggested
by Bugz (#2298).
- Modulized a lot of commands and related subfunctions: NICK (750 lines), USER (200),
MODE (2300), WATCH (250), JOIN (600), PART (250), MOTD (100), OPERMOTD (100),
BOTMOTD (100), LUSERS (100). More will follow soon (probably including more subfunctions
related to existing commands).
- [Module coders] Added new function: do_cmd(cptr, sptr, cmd, parc, parv) which is an
uniform method to call any other commands. For more info, see description in src/parcket.c.
This will be used for any further modulization of commands that need to call other
commands, like NICK (will be done soon).
to send to normal users w/the snomask set.
- Fixed dcc filtering a bit more.
- Made usermode 'g' operonly since it didn't do much, reported by DukePyrolator (#0002024).
work just the same as the HOOKTYPE_LOCAL_* variants).
- Module coders: HOOKTYPE_REMOTE_CONNECT is now also called during net-merge. You can use
IsSynched(sptr->srvptr) to find out if it's called due to a net merge (0) or a connect (1).
- Added spamfiler 'user' (u) target. This regex is checked against nick!user@host:realname
when a user connects. This makes it easy to ban drones with simple patterns.
For example: '/spamfilter add u gzline 86400 Drone[0-9]+!.+@.+:Drone[0-9]'
would kill any drones that have both a nick and realname with 'Drone' followed by digits.
/connect or autoconnect) and was not present in the cache. Reported and traced by sh0
(#0001976).
- Fixed compile bug at *NIX caused by ModuleGetErrorStr fix.
Bram Matthys