#!/usr/bin/env python
"""
irkerd - a simple IRC multiplexer daemon

Listens for JSON objects of the form {'to':<irc-url>, 'privmsg':<text>}
and relays messages to IRC channels. Each request must be followed by
a newline.

The <text> must be a string.  The value of the 'to' attribute can be a
string containing an IRC URL (e.g. 'irc://chat.freenet.net/botwar') or
a list of such strings; in the latter case the message is broadcast to
all listed channels.  Note that the channel portion of the URL need
*not* have a leading '#' unless the channel name itself does.

Options: -d sets the debug-message level (probably only of interest to
developers). -l sets a logfile to capture message traffic from
channels.  -n sets the nick and -p the nickserv password. The -V
option prints the program version and exits.

Design and code by Eric S. Raymond <esr@thyrsus.com>. See the project
resource page at <http://www.catb.org/~esr/irker/>.

Requires Python 2.6 or 2.5 with the simplejson library installed, and
the irc client library at version >= 3.4 which requires 2.6: see

http://pypi.python.org/pypi/irc/
"""
# These things might need tuning

HOST = "localhost"
PORT = 6659

XMIT_TTL = (3 * 60 * 60)	# Time to live, seconds from last transmit
PING_TTL = (15 * 60)		# Time to live, seconds from last PING
HANDSHAKE_TTL = 60		# Time to live, seconds from nick transmit
CHANNEL_TTL = (3 * 60 * 60)	# Time to live, seconds from last transmit
DISCONNECT_TTL = (24 * 60 * 60)	# Time to live, seconds from last connect
UNSEEN_TTL = 60			# Time to live, seconds since first request
CHANNEL_MAX = 18		# Max channels open per socket (default)
ANTI_FLOOD_DELAY = 0.5		# Anti-flood delay after transmissions, seconds
ANTI_BUZZ_DELAY = 0.09		# Anti-buzz delay after queue-empty check
CONNECTION_MAX = 200		# To avoid hitting a thread limit

# No user-serviceable parts below this line

version = "1.17"

import sys, getopt, urlparse, time, random, socket, signal, re
import threading, Queue, SocketServer
import irc.client, logging
try:
    import simplejson as json	# Faster, also makes us Python-2.4-compatible
except ImportError:
    import json

# Sketch of implementation:
#
# One Irker object manages multiple IRC sessions.  It holds a map of
# Dispatcher objects, one per (server, port) combination, which are
# responsible for routing messages to one of any number of Connection
# objects that do the actual socket conversations.  The reason for the
# Dispatcher layer is that IRC daemons limit the number of channels a
# client (that is, from the daemon's point of view, a socket) can be
# joined to, so each session to a server needs a flock of Connection
# instances each with its own socket.
#
# Connections are timed out and removed when either they haven't seen a
# PING for a while (indicating that the server may be stalled or down)
# or there has been no message traffic to them for a while, or
# even if the queue is nonempty but efforts to connect have failed for
# a long time.
#
# There are multiple threads. One accepts incoming traffic from all
# servers.  Each Connection also has a consumer thread and a
# thread-safe message queue.  The program main appends messages to
# queues as JSON requests are received; the consumer threads try to
# ship them to servers.  When a socket write stalls, it only blocks an
# individual consumer thread; if it stalls long enough, the session
# will be timed out. This solves the biggest problem with a
# single-threaded implementation, which is that you can't count on a
# single stalled write not hanging all other traffic - you're at the
# mercy of the length of the buffers in the TCP/IP layer.
#
# Message delivery is thus not reliable in the face of network stalls,
# but this was considered acceptable because IRC (notoriously) has the
# same problem - there is little point in reliable delivery to a relay
# that is down or unreliable.
#
# This code uses only NICK, JOIN, PART, MODE, and PRIVMSG. It is strictly
# compliant to RFC1459, except for the interpretation and use of the
# DEAF and CHANLIMIT and (obsolete) MAXCHANNELS features.  CHANLIMIT
# is as described in the Internet RFC draft
# draft-brocklesby-irc-isupport-03 at <http://www.mirc.com/isupport.html>.
# The ",isnick" feature is as described in
# <http://ftp.ics.uci.edu/pub/ietf/uri/draft-mirashi-url-irc-01.txt>.

class Connection:
    def __init__(self, irkerd, servername, port):
        self.irker = irkerd
        self.servername = servername
        self.port = port
        self.nick_trial = None
        self.connection = None
        self.status = None
        self.last_xmit = time.time()
        self.last_ping = time.time()
        self.channels_joined = {}
        self.channel_limits = {}
        # The consumer thread
        self.queue = Queue.Queue()
        self.thread = None
    def nickname(self, n=None):
        "Return a name for the nth server connection."
        if n is None:
            n = self.nick_trial
        if fallback:
            return (namestyle % n)
        else:
            return namestyle
    def handle_ping(self):
        "Register the fact that the server has pinged this connection."
        self.last_ping = time.time()
    def handle_welcome(self):
        "The server says we're OK, with a non-conflicting nick."
        self.status = "ready"
        self.irker.debug(1, "nick %s accepted" % self.nickname())
        if password:
            self.connection.privmsg("nickserv", "identify %s" % password)
    def handle_badnick(self):
        "The server says our nick is ill-formed or has a conflict."
        self.irker.debug(1, "nick %s rejected" % self.nickname())
        if fallback:
            # Randomness prevents a malicious user or bot from
            # anticipating the next trial name in order to block us
            # from completing the handshake.
            self.nick_trial += random.randint(1, 3)
            self.last_xmit = time.time()
            self.connection.nick(self.nickname())
        # Otherwise fall through, it might be possible to
        # recover manually.
    def handle_disconnect(self):
        "Server disconnected us for flooding or some other reason."
        self.connection = None
        self.status = "disconnected"
    def handle_kick(self, outof):
        "We've been kicked."
        self.status = "handshaking"
        try:
            del self.channels_joined[outof]
        except KeyError:
            self.irker.logerr("kicked by %s from %s that's not joined"
                              % (self.servername, outof))
        qcopy = []
        while not self.queue.empty():
            (channel, message) = self.queue.get()
            if channel != outof:
                qcopy.append((channel, message))
        for (channel, message) in qcopy:
            self.queue.put((channel, message))
        self.status = "ready"
    def enqueue(self, channel, message):
        "Enque a message for transmission."
        if self.thread is None or not self.thread.is_alive():
            self.status = "unseen"
            self.thread = threading.Thread(target=self.dequeue)
            self.thread.setDaemon(True)
            self.thread.start()
        self.queue.put((channel, message))
    def dequeue(self):
        "Try to ship pending messages from the queue."
        try:
            while True:
                # We want to be kind to the IRC servers and not hold unused
                # sockets open forever, so they have a time-to-live.  The
                # loop is coded this particular way so that we can drop
                # the actual server connection when its time-to-live
                # expires, then reconnect and resume transmission if the
                # queue fills up again.
                if self.queue.empty():
                    # Queue is empty, at some point we want to time out
                    # the connection rather than holding a socket open in
                    # the server forever.
                    now = time.time()
                    xmit_timeout = now > self.last_xmit + XMIT_TTL
                    ping_timeout = now > self.last_ping + PING_TTL
                    if self.status == "disconnected":
                        # If the queue is empty, we can drop this connection.
                        self.status = "expired"
                        break
                    elif xmit_timeout or ping_timeout:
                        self.irker.debug(1, "timing out connection to %s at %s (ping_timeout=%s, xmit_timeout=%s)" % (self.servername, time.asctime(), ping_timeout, xmit_timeout))
                        with self.irker.irc.mutex:
                            self.connection.context = None
                            self.connection.quit("transmission timeout")
                            self.connection = None
                        self.status = "disconnected"
                    else:
                        # Prevent this thread from hogging the CPU by pausing
                        # for just a little bit after the queue-empty check.
                        # As long as this is less that the duration of a human
                        # reflex arc it is highly unlikely any human will ever
                        # notice.
                        time.sleep(ANTI_BUZZ_DELAY)
                elif self.status == "disconnected" \
                         and time.time() > self.last_xmit + DISCONNECT_TTL:
                    # Queue is nonempty, but the IRC server might be
                    # down. Letting failed connections retain queue
                    # space forever would be a memory leak.
                    self.status = "expired"
                    break
                elif not self.connection:
                    # Queue is nonempty but server isn't connected.
                    with self.irker.irc.mutex:
                        self.connection = self.irker.irc.server()
                        self.connection.context = self
                        # Try to avoid colliding with other instances
                        self.nick_trial = random.randint(1, 990)
                        self.channels_joined = {}
                        try:
                            # This will throw
                            # irc.client.ServerConnectionError on failure
                            self.connection.connect(self.servername,
                                                self.port,
                                                nickname=self.nickname(),
                                                username="irker",
                                                ircname="irker relaying client")
                            if hasattr(self.connection, "buffer"):
                                self.connection.buffer.errors = 'replace'
                            self.status = "handshaking"
                            self.irker.debug(1, "XMIT_TTL bump (%s connection) at %s" % (self.servername, time.asctime()))
                            self.last_xmit = time.time()
                            self.last_ping = time.time()
                        except irc.client.ServerConnectionError:
                            self.status = "disconnected"
                elif self.status == "handshaking":
                    if time.time() > self.last_xmit + HANDSHAKE_TTL:
                        self.status = "expired"
                        break
                    else:
                        # Don't buzz on the empty-queue test while we're
                        # handshaking
                        time.sleep(ANTI_BUZZ_DELAY)
                elif self.status == "unseen" \
                         and time.time() > self.last_xmit + UNSEEN_TTL:
                    # Nasty people could attempt a denial-of-service
                    # attack by flooding us with requests with invalid
                    # servernames. We guard against this by rapidly
                    # expiring connections that have a nonempty queue but
                    # have never had a successful open.
                    self.status = "expired"
                    break
                elif self.status == "ready":
                    (channel, message) = self.queue.get()
                    if channel not in self.channels_joined:
                        self.connection.join(channel)
                        self.irker.debug(1, "joining %s on %s." % (channel, self.servername))
                    # An empty message might be used as a keepalive or
                    # to join a channel for logging, so suppress the
                    # privmsg send unless there is actual traffic.
                    if message:
                        for segment in message.split("\n"):
                            # Truncate the message if it's too long,
                            # but we're working with characters here,
                            # not bytes, so we could be off.
                            # 500 = 512 - CRLF - 'PRIVMSG ' - ' :'
                            maxlength = 500 - len(channel)
                            if len(segment) > maxlength:
                                segment = segment[:maxlength]
                            try:
                                self.connection.privmsg(channel, segment)
                            except ValueError as err:
                                self.irker.debug(1, "irclib rejected a message to %s on %s because: %s" % (channel, self.servername, str(err)))
                            time.sleep(ANTI_FLOOD_DELAY)
                    self.last_xmit = self.channels_joined[channel] = time.time()
                    self.irker.debug(1, "XMIT_TTL bump (%s transmission) at %s" % (self.servername, time.asctime()))
                    self.queue.task_done()
        except:
            (exc_type, _exc_value, exc_traceback) = sys.exc_info()
            self.irker.logerr("exception %s in thread for %s" % \
                              (exc_type, self.servername))

            # Maybe this should have its own status?
            self.status = "expired"

            # This is so we can see tracebacks for errors inside the thread
            # when we need to be able to for debugging purposes.
            if debuglvl > 0:
                raise exc_type, _exc_value, exc_traceback
        finally:
            try:
                # Make sure we don't leave any zombies behind
                self.connection.close()
            except:
                # Irclib has a habit of throwing fresh exceptions here. Ignore that
                pass
    def live(self):
        "Should this connection not be scavenged?"
        return self.status != "expired"
    def joined_to(self, channel):
        "Is this connection joined to the specified channel?"
        return channel in self.channels_joined
    def accepting(self, channel):
        "Can this connection accept a join of this channel?"
        if self.channel_limits:
            match_count = 0
            for already in self.channels_joined:
                # This obscure code is because the RFCs allow separate limits
                # by channel type (indicated by the first character of the name)
                # a feature that is almost never actually used.
                if already[0] == channel[0]:
                    match_count += 1
            return match_count < self.channel_limits.get(channel[0], CHANNEL_MAX)
        else:
            return len(self.channels_joined) < CHANNEL_MAX

class Target():
    "Represent a transmission target."
    def __init__(self, url):
        # Pre-2.6 Pythons don't recognize irc: as a valid URL prefix.
        url = url.replace("irc://", "http://")
        parsed = urlparse.urlparse(url)
        irchost, _, ircport = parsed.netloc.partition(':')
        if not ircport:
            ircport = 6667
        self.servername = irchost
        # IRC channel names are case-insensitive.  If we don't smash
        # case here we may run into problems later. There was a bug
        # observed on irc.rizon.net where an irkerd user specified #Channel,
        # got kicked, and irkerd crashed because the server returned
        # "#channel" in the notification that our kick handler saw.
        self.channel = parsed.path.lstrip('/').lower()
        # This deals with a tweak in recent versions of urlparse.
        if parsed.fragment:
            self.channel += "#" + parsed.fragment
        isnick = self.channel.endswith(",isnick")
        if isnick:
            self.channel = self.channel[:-7]
        if self.channel and not isnick and self.channel[0] not in "#&+":
            self.channel = "#" + self.channel
        self.port = int(ircport)
    def valid(self):
        "Both components must be present for a valid target."
        return self.servername and self.channel
    def server(self):
        "Return a hashable tuple representing the destination server."
        return (self.servername, self.port)

class Dispatcher:
    "Manage connections to a particular server-port combination."
    def __init__(self, irkerd, servername, port):
        self.irker = irkerd
        self.servername = servername
        self.port = port
        self.connections = []
    def dispatch(self, channel, message):
        "Dispatch messages for our server-port combination."
        # First, check if there is room for another channel
        # on any of our existing connections.
        connections = [x for x in self.connections if x.live()]
        eligibles = [x for x in connections if x.joined_to(channel)] \
                    or [x for x in connections if x.accepting(channel)]
        if eligibles:
            eligibles[0].enqueue(channel, message)
            return
        # All connections are full up. Look for one old enough to be
        # scavenged.
        ancients = []
        for connection in connections:
            for (chan, age) in connections.channels_joined.items():
                if age < time.time() - CHANNEL_TTL:
                    ancients.append((connection, chan, age))
        if ancients:
            ancients.sort(key=lambda x: x[2]) 
            (found_connection, drop_channel, _drop_age) = ancients[0]
            found_connection.part(drop_channel, "scavenged by irkerd")
            del found_connection.channels_joined[drop_channel]
            #time.sleep(ANTI_FLOOD_DELAY)
            found_connection.enqueue(channel, message)
            return
        # Didn't find any channels with no recent activity
        newconn = Connection(self.irker,
                             self.servername,
                             self.port)
        self.connections.append(newconn)
        newconn.enqueue(channel, message)
    def live(self):
        "Does this server-port combination have any live connections?"
        self.connections = [x for x in self.connections if x.live()]
        return len(self.connections) > 0
    def last_xmit(self):
        "Return the time of the most recent transmission."
        return max(x.last_xmit for x in self.connections)

class Irker:
    "Persistent IRC multiplexer."
    def __init__(self, debuglevel=0):
        self.debuglevel = debuglevel
        self.irc = irc.client.IRC()
        self.irc.add_global_handler("ping", self._handle_ping)
        self.irc.add_global_handler("welcome", self._handle_welcome)
        self.irc.add_global_handler("erroneusnickname", self._handle_badnick)
        self.irc.add_global_handler("nicknameinuse", self._handle_badnick)
        self.irc.add_global_handler("nickcollision", self._handle_badnick)
        self.irc.add_global_handler("unavailresource", self._handle_badnick)
        self.irc.add_global_handler("featurelist", self._handle_features)
        self.irc.add_global_handler("disconnect", self._handle_disconnect)
        self.irc.add_global_handler("kick", self._handle_kick)
        self.irc.add_global_handler("all_raw_messages", self._handle_all_raw_messages)
        thread = threading.Thread(target=self.irc.process_forever)
        thread.setDaemon(True)
        self.irc._thread = thread
        thread.start()
        self.servers = {}
    def logerr(self, errmsg):
        "Log a processing error."
        sys.stderr.write("irkerd: " + errmsg + "\n")
    def debug(self, level, errmsg):
        "Debugging information."
        if self.debuglevel >= level:
            sys.stderr.write("irkerd: %s\n" % errmsg)
    def _handle_ping(self, connection, _event):
        "PING arrived, bump the last-received time for the connection."
        if connection.context:
            connection.context.handle_ping()
    def _handle_welcome(self, connection, _event):
        "Welcome arrived, nick accepted for this connection."
        if connection.context:
            connection.context.handle_welcome()
    def _handle_badnick(self, connection, _event):
        "Nick not accepted for this connection."
        if connection.context:
            connection.context.handle_badnick()
    def _handle_features(self, connection, event):
        "Determine if and how we can set deaf mode."
        if connection.context:
            cxt = connection.context
            arguments = event.arguments
            # irclib 5.0 compatibility, because the maintainer is a fool
            if callable(arguments):
                arguments = arguments()
            for lump in arguments:
                if lump.startswith("DEAF="):
                    if not logfile:
                        connection.mode(cxt.nickname(), "+"+lump[5:])
                elif lump.startswith("MAXCHANNELS="):
                    m = int(lump[12:])
                    for pref in "#&+":
                        cxt.channel_limits[pref] = m
                    self.debug(1, "%s maxchannels is %d"
                               % (connection.server, m))
                elif lump.startswith("CHANLIMIT=#:"):
                    limits = lump[10:].split(",")
                    try:
                        for token in limits:
                            (prefixes, limit) = token.split(":")
                            limit = int(limit)
                            for c in prefixes:
                                cxt.channel_limits[c] = limit
                        self.debug(1, "%s channel limit map is %s"
                                   % (connection.server, cxt.channel_limits))
                    except ValueError:
                        self.logerr("ill-formed CHANLIMIT property")
    def _handle_disconnect(self, connection, _event):
        "Server hung up the connection."
        self.debug(1, "server %s disconnected" % connection.server)
        connection.close()
        if connection.context:
            connection.context.handle_disconnect()
    def _handle_kick(self, connection, event):
        "Server hung up the connection."
        target = event.target
        # irclib 5.0 compatibility, because the maintainer continues
        # to be a fool.
        if callable(target):
            target = target()
        self.debug(1, "irker has been kicked from %s on %s" % (target, connection.server))
        if connection.context:
            connection.context.handle_kick(target)
    def _handle_all_raw_messages(self, _connection, event):
        "Log all messages when in watcher mode."
        with open(logfile, "a") as logfp:
            logfp.write("%03f|%s|%s\n" % \
                             (time.time(), event.source, event.arguments[0]))
    def handle(self, line):
        "Perform a JSON relay request."
        try:
            request = json.loads(line.strip())
            if not isinstance(request, dict):
                self.logerr("request is not a JSON dictionary: %r" % request)
            elif "to" not in request or "privmsg" not in request:
                self.logerr("malformed request - 'to' or 'privmsg' missing: %r" % request)
            else:
                channels = request['to']
                message = request['privmsg']
                if not isinstance(channels, (list, basestring)):
                    self.logerr("malformed request - unexpected channel type: %r" % channels)
                if not isinstance(message, basestring):
                    self.logerr("malformed request - unexpected message type: %r" % message)
                else:
                    if not isinstance(channels, list):
                        channels = [channels]
                    for url in channels:
                        if not isinstance(url, basestring):
                            self.logerr("malformed request - URL has unexpected type: %r" % url)
                        else:
                            target = Target(url)
                            if not target.valid():
                                return
                            if target.server() not in self.servers:
                                self.servers[target.server()] = Dispatcher(self, target.servername, target.port)
                            self.servers[target.server()].dispatch(target.channel, message)
                            # GC dispatchers with no active connections
                            servernames = self.servers.keys()
                            for servername in servernames:
                                if not self.servers[servername].live():
                                    del self.servers[servername]
                            # If we might be pushing a resource limit
                            # even after garbage collection, remove a
                            # session.  The goal here is to head off
                            # DoS attacks that aim at exhausting
                            # thread space or file descriptors.  The
                            # cost is that attempts to DoS this
                            # service will cause lots of join/leave
                            # spam as we scavenge old channels after
                            # connecting to new ones. The particular
                            # method used for selecting a session to
                            # be terminated doesn't matter much; we
                            # choose the one longest idle on the
                            # assumption that message activity is likely
                            # to be clumpy.
                            if len(self.servers) >= CONNECTION_MAX:
                                oldest = min(self.servers.keys(), key=lambda name: self.servers[name].last_xmit())
                                del self.servers[oldest]
        except ValueError:
            self.logerr("can't recognize JSON on input: %r" % line)
        except RuntimeError:
            self.logerr("wildly malformed JSON blew the parser stack.")

class IrkerTCPHandler(SocketServer.StreamRequestHandler):
    def handle(self):
        while True:
            line = self.rfile.readline()
            if not line:
                break
            irker.handle(line.strip())

class IrkerUDPHandler(SocketServer.BaseRequestHandler):
    def handle(self):
        data = self.request[0].strip()
        #socket = self.request[1]
        irker.handle(data)

if __name__ == '__main__':
    debuglvl = 0
    namestyle = "irker%03d"
    password = None
    logfile = None
    (options, arguments) = getopt.getopt(sys.argv[1:], "d:l:n:p:V:")
    for (opt, val) in options:
        if opt == '-d':		# Enable debug/progress messages
            debuglvl = int(val)
            if debuglvl > 1:
                logging.basicConfig(level=logging.DEBUG)
        elif opt == '-l':	# Logfile mode - report traffic read in
            logfile = val
        elif opt == '-n':	# Force the nick
            namestyle = val
        elif opt == '-p':	# Set a nickserv password
            password = val
        elif opt == '-V':	# Emit version and exit
            sys.stdout.write("irkerd version %s\n" % version)
            sys.exit(0)
    fallback = re.search("%.*d", namestyle)
    irker = Irker(debuglevel=debuglvl)
    irker.debug(1, "irkerd version %s" % version)
    try:
        tcpserver = SocketServer.TCPServer((HOST, PORT), IrkerTCPHandler)
        udpserver = SocketServer.UDPServer((HOST, PORT), IrkerUDPHandler)
        for server in [tcpserver, udpserver]:
            server = threading.Thread(target=server.serve_forever)
            server.setDaemon(True)
            server.start()
        try:
            signal.pause()
        except KeyboardInterrupt:
            raise SystemExit(1)
    except socket.error, e:
        sys.stderr.write("irkerd: server launch failed: %r\n" % e)

# end