James Lu
80bebe3827
protocols: also validate TS on incoming SJOIN
2015-08-26 14:59:05 -07:00
James Lu
977b69b8e9
protocols/: Fix args parsing when the sender UID == the last argument
...
args.index() would return 0 instead of the index of the current argument.
2015-08-26 14:57:46 -07:00
James Lu
05be3e6c17
pr/: add prefixmodes in sjoinServer if TS <= theirs
...
The old comparison of our TS < theirs was wrong, because modes are merged (not dropped) when two channels of the same TS join.
2015-08-24 18:37:39 -07:00
James Lu
694e7b87f6
various: make the server description configurable
...
Two options for this: bot::serverdesc, and server::<network>::serverdesc, with the latter taking precedence if present.
2015-08-24 18:14:35 -07:00
James Lu
1d245bf001
protoocols.sjoinServer: only add prefix modes to channel state if our TS < theirs
2015-08-20 01:32:30 -07:00
James Lu
61d95c19cc
protocols: delete empty non-permanent channels
2015-08-15 23:29:40 -07:00
James Lu
c99a6a7b9d
protocols: update topic in channel state when sending outbound TOPIC
2015-08-15 22:34:40 -07:00
James Lu
4352a68357
Check for TS changes on SJOIN (both incoming and outgoing)
...
Restore relay to the devel branch; TS bursts shouldn't be handled there, but in the protocol modules.
TODO: fix prefixmodes being cleared after BURST, and never being reset?
2015-08-15 19:53:09 -07:00
James Lu
0d41e35d9c
Merge branch 'devel' into wip/handle-ts-bursts
...
Conflicts:
plugins/relay.py
2015-08-15 19:28:49 -07:00
James Lu
13e4baba8b
inspircd: also strip listmodes in joinClient ( #58 )
2015-08-14 08:47:23 -07:00
James Lu
7d912bbb28
pr/insp: Strip out listmodes in FJOIN
...
They should always be sent separately according to the protocol documentation: https://wiki.inspircd.org/InspIRCd_Spanning_Tree_1.2/FJOIN
Closes #58 .
2015-08-12 07:05:05 -07:00
James Lu
1a57dfcdc3
prs+IrcUser: add incoming (handle_away) and outgoing (awayClient) AWAY handling
2015-08-12 04:17:01 -07:00
James Lu
bf3116d704
First attempt at handling channel TS @ burst ( #93 )
...
Doesn't quite work yet, more testing needed.
2015-08-05 06:14:34 -07:00
James Lu
8a134aadc5
Remove duplicate assignment of irc.uidgen
2015-08-03 19:44:11 -07:00
James Lu
e5eb58ee73
inspircd: use parse_as instead of forcibly mapping ENCAP to KNOCK
2015-08-03 05:00:32 -07:00
James Lu
e354ada838
relay: fix nick collision loop on SAVE + when both tagged UID and untagged UID exist on one net
...
This fixes a clash when for example: both 42XAAAAAA and _42XAAAAAA exist on a network, and PyLink tries to relay both nicks as _42XAAAAAA/network.
Also, this adds an oldnick argument to the SAVE protocol handler, which is then used by relay.normalizeNick to check whether the original pre-SAVE nick is also in use, in the event of nick collisions.
2015-07-25 18:27:30 -07:00
James Lu
a4da9b5324
protocol/relay: fix handling of KILLs sent to non-relay users
2015-07-24 18:26:31 -07:00
James Lu
007a1d9389
Merge branch 'master' into devel
2015-07-24 11:09:53 -07:00
James Lu
01220b3024
correction: InspIRCd uses RFC1459 case mapping by default
2015-07-24 10:59:04 -07:00
James Lu
868b4503e6
protocols: Fix #81 - that was easy!
...
Split data by only one space at a time, not as many spaces as possible. Thanks to @nathan0.
2015-07-23 19:09:19 -07:00
James Lu
58a8d7134c
Fix handling of inbound CHG*
2015-07-23 11:02:32 -07:00
James Lu
8799e1ccc8
spawnClient: add ts as opt. argument; relay: spawn pseudoclients w/ TS of original user
2015-07-23 00:11:24 -07:00
James Lu
d52fba37b8
add numericServer for sending raw numerics from servers, and WHOIS handling in coreplugin
...
Basic WHOIS handling (user information, server information, IRCop access) are sent. #72
TODO: send channel lists, an extra note for relay clients, user modes, and idle time
2015-07-22 21:15:34 -07:00
James Lu
8a1f965303
Consistently use applyModes in spawnClient, so user modes are stored like ('o', None) instead of ('+o', None)
...
Drop the 'modes' argument in IrcUser, for this is incorrect.
2015-07-22 21:14:22 -07:00
James Lu
686467ffa6
Remove endburst toggling in spawnServer, has_bursted IrcServer flag
...
Neither are being used at all; they're essentially dead code.
2015-07-22 20:39:38 -07:00
James Lu
35cdfbf7e6
Declare IRCd casemapping in protocol modules, and respect these in utils.nickToUid
...
This adds a new utils.toLower(irc, text) function which returns the lowercased version of <text> based on <irc>'s declared case mapping.
Closes #75 .
2015-07-22 20:31:45 -07:00
James Lu
ea7681502f
fix some bugs, introduce others
2015-07-21 17:33:36 -07:00
James Lu
9b8ef0a45d
make relay and ts6 a bit less, well, buggy
2015-07-21 12:44:01 -07:00
James Lu
72be5ca79c
inspircd: make handle_part return a list of channels, not just one
...
Some IRCds, like TS6, allow sending multiple channels (as a comma-separated list) in PART.
Update relay accordingly.
2015-07-20 19:52:52 -07:00
James Lu
3494d4f794
inspircd: remove RSQUIT handler
...
Events like this shouldn't be handled at the protocol level, ever. There's little point in adding this globally, since plugins should be able to choose what happens when they receive an SQUIT.
2015-07-20 18:05:42 -07:00
James Lu
a10a18d9cc
Irc: make prefixmodes a dict mapping chars to their prefixes
...
This is needed by TS6 support (#40 ), which send the characters (@+) instead of modes
2015-07-20 16:48:59 -07:00
James Lu
b1e409ff3e
relay: unbreak topic handling ( closes #68 )
...
Introducing a new .topicset attributing in IrcChannel denoting whether we've received a TOPIC for this channel from the uplink yet.
2015-07-20 13:18:04 -07:00
James Lu
fe9c317f2c
inspircd: make channel user tracking more fault tolerant (in sjoinServer / handle_part)
2015-07-19 23:46:19 -07:00
James Lu
bf444f3f6a
pr/insp: Servers should use FTOPIC, not TOPIC to send topics
2015-07-19 17:27:53 -07:00
James Lu
0cb9065d53
pr/insp: coerse 'reginvite' named mode to 'regonly'
2015-07-19 16:53:51 -07:00
James Lu
0540e10d50
relay: don't add bans that don't match nick!user@host
...
Closes #55 .
2015-07-19 16:53:51 -07:00
James Lu
a51cfcb7b2
Fix fakeirc and tests for relay ( #56 )
2015-07-17 22:52:55 -07:00
James Lu
fdea348cbe
move hook calling out of pr/insp and into Irc
2015-07-17 16:09:50 -07:00
James Lu
95853924ad
Revert "protocols/: thread calls to each event hook to prevent blocking globally"
...
This reverts commit 99fd6060a7
, which caused a bunch of race conditions in relay :/
Ref #59 .
2015-07-17 15:13:57 -07:00
James Lu
f41e1a2eb5
Fixes for connection handling?
...
XXX: using SQUIT on the PyLink server doesn't respect autoconnect delays; why?
2015-07-17 15:08:24 -07:00
James Lu
99fd6060a7
protocols/: thread calls to each event hook to prevent blocking globally
...
Closes #59 .
2015-07-17 14:35:02 -07:00
James Lu
9bef93c341
More stuff for PINGing uplink and autoreconnect ( #42 #59 #57 )
...
Doesn't quite work yet; still blocks everything on disconnect for some reason...
2015-07-17 13:41:14 -07:00
James Lu
0aa2f98705
Move client spawning, command / KILL / KICK handling outside the protocol module
...
New plugin 'coreplugin' takes care of command/KILL/KICK handling, while client spawning is moved into Irc.spawnMain() and is used by the former.
Closes #33 .
2015-07-16 21:52:44 -07:00
James Lu
73c625a429
pr/insp: don't raise KeyError when removing channel from user fails
2015-07-16 16:48:54 -07:00
James Lu
d97fce8205
IrcUser & pr/insp: Implement user channel tracking
2015-07-16 12:20:40 -07:00
James Lu
f9d82153bb
pr/insp: map FIDENT/FHOST/FNAME hooks to their CHG* counterparts
2015-07-15 23:23:46 -07:00
James Lu
e413a21e68
pr/insp: send the list of affected users in handle_squit
2015-07-15 19:08:49 -07:00
James Lu
798476c850
pr/insp: don't crash when a server sends TOPIC
2015-07-15 19:08:36 -07:00
James Lu
6e37e1c05d
make Irc.connected a threaded event object, setting it to True ONLY when we receive server capabilities from our uplink
...
The previous behavior set this to True as soon as we ran connect(), but this caused problems because the default capabilities (i.e. nicklen)
that Irc() initializes won't match the real value of the network.
2015-07-15 13:49:12 -07:00
James Lu
67190d8329
pr/insp: send the user data of the kileld user in handle_kill
2015-07-14 23:23:35 -07:00