The tools required now are supported by Debian GNU/Linux 4.0 "Etch",
RedHat Enterprise Linux 5, and Mac OS X 10.6.x for example.
I read the changelogs of autoconf and automake, and I think ther's nothing
that prevents it from working with these older versions; and we don't want
to force users to upgrade without real benefits.
Please note: the recommended versions are still autoconf 1.11.x and the
most recent autoconf release that works with automake 1.11.x!
Alexander Barton [Sun, 23 Sep 2012 17:58:50 +0000 (19:58 +0200)]
Merge branch 'bug92-xop'
By Alexander Barton (5) and Sebastian Köhler (2)
* bug92-xop:
Fix NAMES/WHO response when client has multi-prefix
Fix prefix of "halfop" when "multi-prefix" is active
Clean up doc/.gitignore
doc/Modes.txt: add version number to new channel modes
Fix some "whitespace glitches"
Tests and documentation for xop
Implemented xop support
Conflicts (because of "multi-prefix fix"):
src/ngircd/irc-info.c
This fixes bug #92 "ngircd does not support XOP usermodes".
Alexander Barton [Fri, 21 Sep 2012 08:36:09 +0000 (10:36 +0200)]
Initialize SSL when needed only, and disable SSL on errors
With this patch, the SSL subsystem will only be initialized if at least
one SSL ports is configured; so you won't get "SSL initialization failed"
messages if you didn't configured it at all.
And if SSL initialization fails, no SSL listen ports will be enabled
later which never could establish a working SSL connection at all ...
Alexander Barton [Sun, 16 Sep 2012 11:05:23 +0000 (13:05 +0200)]
Merge branch 'autoconf-update'
Update GNU autoconf and automake infrastructure.
Tested on modern systems as well as Apple A/UX :-)
* autoconf-update:
AUTOMAKE_OPTIONS: fix ansi2knr option, include path
Don't use AC_FUNC_MALLOC and AC_FUNC_REALLOC
Make our own targets "silent", if enabled
configure.in: use AC_CHECK_{FUNCS|HEADERS}_ONCE
Updated config.{guess|sub} to version 2012-08-14
Make autogen.sh more verbose when VERBOSE=1 is set
configure.in: use AC_SEARCH_LIBS (not AC_CHECK_LIB)
configure.in: use AS_HELP_STRING macro
configure.in: use AC_CANONICAL_HOST (not AC_CANONICAL_TARGET)
configure.in: inttypes.h is an optional header file
Use HAVE_SETSID #define when testing for setsid()
Don't include <stdint.h>, it is included by "portab.h"
Don't check type.h availability, it is required
configure.in: Use AC_CONFIG_FILES macro
configure.in: Don't use AC_C_PROTOTYPES
configure.in: Update checks for required and optional features
configure.in: require autoconf 2.67 and automake 1.11
configure.in: sort some lists (templates, output, ...)
Alexander Barton [Sat, 15 Sep 2012 18:26:59 +0000 (20:26 +0200)]
Don't use AC_FUNC_MALLOC and AC_FUNC_REALLOC
We don't expect the GNU'ish behaviour of of malloc() and realloc() and never
implemented the replacement functions rpl_malloc()/rpl_realloc() -- so these
test result in linking failues on systems that don't have a GNU'ish malloc()
and realloc() even though we don't require it!
Alexander Barton [Tue, 11 Sep 2012 09:40:47 +0000 (11:40 +0200)]
Show a warning on startup if config file is not a full path
ngIRCd is a long-running process and changes its working directory to "/" to
not block mounted filesystems and the like when running as daemon ("not in the
foreground"); therefore the path to the configuration file must be relative to
"/" (or the chroot() directory), which basically is "not relative", to ensure
that "kill -HUP" and the "REHASH" command work as expected later on.
Alexander Barton [Mon, 10 Sep 2012 10:43:44 +0000 (12:43 +0200)]
WHO #channel: don't limit list size
It makes no sense to limit the list size when doing WHO for a channel
and not to return all the users in that channel, so I removed the check.
But if there are more than MAX_RPL_WHO(25) replies, the client requesting
the list will be "penalized" one second more (then 2 in total).
Define EV_SET() for kqueue() on systems that don't have it
Some systems, notably FreeBSD 4.x, do have the kqueue() function but
lack the definition of EV_SET() in their header files -- but don't
worry, we can #define it on our own ;-)
Definition taken from /usr/include/sys/event.h of FreeBSD 8.1.
Patch tested on FreeBSD 4.1 by Götz Hoffart. Thanks!
Alexander Barton [Tue, 28 Aug 2012 21:28:56 +0000 (23:28 +0200)]
Introduce numeric RPL_HOSTHIDDEN_MSG(396)
This numeric is sent to the client each time it changes its displayed
hostname using "MODE +/-x", and if "CloakHost" is set right after the
MOTD has been sent.
Alexander Barton [Tue, 28 Aug 2012 21:26:43 +0000 (23:26 +0200)]
Always cloak client hostname, if needed
Not only cloak the hostname in Client_MaskCloaked(), but also in
Client_HostnameCloaked() -- so move the actual cloaking to this function
and call it in Client_MaskCloaked() to get the (cloaked) hostname.
This fixes USERHOST not displaying the correctly cloaked hostname,
for example.
Alexander Barton [Mon, 27 Aug 2012 21:27:30 +0000 (23:27 +0200)]
Merge branch 'umode-B'
This patch series allows ngIRCd to support the user mode "B" ("Bot flasg"):
it is settable and unsettable by every (non-restricted) client.
According to DNS777, this is how Unreal and InspIRCd do behave, so do we :-)
By Alexander Barton (1) and DNS777 (1)
* umode-B:
Add new user mode "B" to doc/Modes.txt
Implement an Unreal-like user mode "B" ("Bot mode")
Brian Collins [Mon, 27 Aug 2012 15:14:47 +0000 (16:14 +0100)]
Fix NAMES response when client has multi-prefix
Two fixes here: IRC_Send_NAMES was checking the capability of the
wrong client when responding, and it didn't return any prefix for
clients that had either +v or +o but not both.
Alexander Barton [Sun, 26 Aug 2012 17:14:29 +0000 (19:14 +0200)]
Merge branch 'move-connection-password' of git://arthur.barton.de/ngircd-alex
This patch series converts the statically allocated password buffer in the
CLIENT structure into a dynamically (and only when needed) allocated buffer
which is referenced by the CONNECTION structure.
This a) saves memory for clients not using passwords at all and b) allows
for "arbitrarily" long passwords.
By Brett Smith (5) and Alexander Barton (2).
* 'move-connection-password' of git://arthur.barton.de/ngircd-alex:
Login_User(): use "conn" insted of calling Client_Conn(Client)
Free already saved password when storing a new one
Indentation and style fixes.
Connection password is not constant.
Implementation clean-ups.
Dynamically allocate memory for connection password.
Move client password from the Client to the Connection struct.
Alexander Barton [Sun, 26 Aug 2012 11:24:54 +0000 (13:24 +0200)]
Enhance "ServiceMask" to handle a list of masks
The "ServiceMask" variable in "Server" blocks now can handle more than
one mask using the new MatchCaseInsensitiveList() function.
This makes marking "service clients" much more specific, which is a
good thing per se, but which is the prerequisite for reasonably
blocking these nick names, too (see commit a6dd2e3 for details).
Alexander Barton [Sun, 26 Aug 2012 10:33:21 +0000 (12:33 +0200)]
Block nicknames that are reserved for services
This patch introduces the new function Conf_NickIsBlocked() which checks
if a given nick name matches with the "service mask" of a configured server.
And Client_CheckNick() uses this information to deny such names for regular
IRC users.
So nick names intended for IRC services are more protected and can't be used
by regular users even when the "services pseudo-server" isn't connected to
the network.
But please note:
Up to now, there can be only one "ServiceMask" pattern per server, which
most probably blocks much more nick names than really required ...
So "ServiceMask" should allow more than one pattern which can be more
specific, and most probably it should be possible to block nick names in
the global server configuration as well.
Nick names introduced by other servers/services are never restricted.
Brett Smith [Thu, 23 Aug 2012 15:07:08 +0000 (11:07 -0400)]
Move client password from the Client to the Connection struct.
This is a relatively naive implementation, basically doing the bare minimum
necessary to make the switchover go. Subsequent commits can focus on
improving the implementation.
Alexander Barton [Sat, 18 Aug 2012 21:50:20 +0000 (23:50 +0200)]
"make uninstall": remove ngircd.conf if not modified
Now "make uninstall" removes the installed "ngircd.conf" file, if it
is still equal to our "sample-ngircd.conf" file and therefore hasn't
been modified by the user. If it has been modified, it isn't removed
and a notice is displayed to the user.
In addition, "make install" now displays a message when no ngircd.conf
file exists and the "sample-ngircd.conf" file will be installed as a
starting point.
Implemented support for hashed hostnames for CloakHost. The admin can
use '%x' in both the CloakHost and CloakHostModeX setting. The config
option CloakHostModeX was renamed to CloakHostSalt. This salt is used
for both cloaking options.
CloakHostModeX can now contain '%x'. It will be replace by the hash of
the original client hostname. The new config option CloakHostModeXSalt
defines the salt for the hash function. When CloakHostModeXSalt is not
set a random salt will be generated after each server restart.
Alexander Barton [Mon, 11 Jun 2012 08:44:02 +0000 (10:44 +0200)]
Merge branch 'bug124-CloakHostModeX'
* bug124-CloakHostModeX:
Describe "CloakHostModeX" in sample-ngircd.conf an ngircd.conf(5)
Rename "CloakModeHost" option to "CloakHostModeX"
Introduce new configuration option "CloakModeHost"
Correctly handle asynchronously re-established server links
Don't try to establish an outgoing server link after DNS lookup when this
server re-connected on its own in the meantime.
In addition, log a warning message if we try to update the connection
index of an already connected server structure -- and ignore it.
Up to now, both behaviour could lead to a race when the remote server
connects to this daemon while it still prepares the outgoing connection:
- The local server prepares the new outgoing connection ...
- in the meantime the remote server becomes connected and registered.
- Now the new outgoing connection overwrites the (correct) socket handle,
- then the 2nd connection becomes disconnected: "already registered",
- and the 1st connection becomes unhandled ("gets lost") because the
configuration structure is reset because of the wrong socket handle.