diff options
author | Nathan Fritz <nathan@andyet.net> | 2010-10-13 18:15:21 -0700 |
---|---|---|
committer | Nathan Fritz <nathan@andyet.net> | 2010-10-13 18:15:21 -0700 |
commit | 7ad7a29a8f08ff815164731eec50ff1cba46b07a (patch) | |
tree | 444a2d89993072a377155d8f131899a19d18dffa /sleekxmpp/thirdparty | |
parent | b0e036d03c5e850af3a3c1589af7333becf54a86 (diff) | |
download | slixmpp-7ad7a29a8f08ff815164731eec50ff1cba46b07a.tar.gz slixmpp-7ad7a29a8f08ff815164731eec50ff1cba46b07a.tar.bz2 slixmpp-7ad7a29a8f08ff815164731eec50ff1cba46b07a.tar.xz slixmpp-7ad7a29a8f08ff815164731eec50ff1cba46b07a.zip |
new state machine in place
Diffstat (limited to 'sleekxmpp/thirdparty')
-rw-r--r-- | sleekxmpp/thirdparty/__init__.py | 0 | ||||
-rw-r--r-- | sleekxmpp/thirdparty/statemachine.py | 278 |
2 files changed, 278 insertions, 0 deletions
diff --git a/sleekxmpp/thirdparty/__init__.py b/sleekxmpp/thirdparty/__init__.py new file mode 100644 index 00000000..e69de29b --- /dev/null +++ b/sleekxmpp/thirdparty/__init__.py diff --git a/sleekxmpp/thirdparty/statemachine.py b/sleekxmpp/thirdparty/statemachine.py new file mode 100644 index 00000000..140c0819 --- /dev/null +++ b/sleekxmpp/thirdparty/statemachine.py @@ -0,0 +1,278 @@ +""" + SleekXMPP: The Sleek XMPP Library + Copyright (C) 2010 Nathanael C. Fritz + This file is part of SleekXMPP. + + See the file LICENSE for copying permission. +""" +import threading +import time +import logging + +log = logging.getLogger(__name__) + + +class StateMachine(object): + + def __init__(self, states=[]): + self.lock = threading.Lock() + self.notifier = threading.Event() + self.__states= [] + self.addStates(states) + self.__default_state = self.__states[0] + self.__current_state = self.__default_state + + def addStates(self, states): + self.lock.acquire() + try: + for state in states: + if state in self.__states: + raise IndexError("The state '%s' is already in the StateMachine." % state) + self.__states.append( state ) + finally: self.lock.release() + + + def transition(self, from_state, to_state, wait=0.0, func=None, args=[], kwargs={} ): + ''' + Transition from the given `from_state` to the given `to_state`. + This method will return `True` if the state machine is now in `to_state`. It + will return `False` if a timeout occurred the transition did not occur. + If `wait` is 0 (the default,) this method returns immediately if the state machine + is not in `from_state`. + + If you want the thread to block and transition once the state machine to enters + `from_state`, set `wait` to a non-negative value. Note there is no 'block + indefinitely' flag since this leads to deadlock. If you want to wait indefinitely, + choose a reasonable value for `wait` (e.g. 20 seconds) and do so in a while loop like so: + + :: + + while not thread_should_exit and not state_machine.transition('disconnected', 'connecting', wait=20 ): + pass # timeout will occur every 20s unless transition occurs + if thread_should_exit: return + # perform actions here after successful transition + + This allows the thread to be responsive by setting `thread_should_exit=True`. + + The optional `func` argument allows the user to pass a callable operation which occurs + within the context of the state transition (e.g. while the state machine is locked.) + If `func` returns a True value, the transition will occur. If `func` returns a non- + True value or if an exception is thrown, the transition will not occur. Any thrown + exception is not caught by the state machine and is the caller's responsibility to handle. + If `func` completes normally, this method will return the value returned by `func.` If + values for `args` and `kwargs` are provided, they are expanded and passed like so: + `func( *args, **kwargs )`. + ''' + + return self.transition_any( (from_state,), to_state, wait=wait, + func=func, args=args, kwargs=kwargs ) + + + def transition_any(self, from_states, to_state, wait=0.0, func=None, args=[], kwargs={} ): + ''' + Transition from any of the given `from_states` to the given `to_state`. + ''' + + if not (isinstance(from_states,tuple) or isinstance(from_states,list)): + raise ValueError( "from_states should be a list or tuple" ) + + for state in from_states: + if not state in self.__states: + raise ValueError( "StateMachine does not contain from_state %s." % state ) + if not to_state in self.__states: + raise ValueError( "StateMachine does not contain to_state %s." % to_state ) + + + start = time.time() + while not self.__current_state in from_states or not self.lock.acquire(False): + # detect timeout: + remainder = start + wait - time.time() + if remainder > 0: self.notifier.wait(remainder) + else: return False + + try: # lock is acquired; all other threads will return false or wait until notify/timeout + if self.__current_state in from_states: # should always be True due to lock + + # Note that func might throw an exception, but that's OK, it aborts the transition + return_val = func(*args,**kwargs) if func is not None else True + + # some 'false' value returned from func, + # indicating that transition should not occur: + if not return_val: return return_val + + log.debug(' ==== TRANSITION %s -> %s', self.__current_state, to_state) + self._set_state( to_state ) + return return_val # some 'true' value returned by func or True if func was None + else: + log.error( "StateMachine bug!! The lock should ensure this doesn't happen!" ) + return False + finally: + self.notifier.set() # notify any waiting threads that the state has changed. + self.notifier.clear() + self.lock.release() + + + def transition_ctx(self, from_state, to_state, wait=0.0): + ''' + Use the state machine as a context manager. The transition occurs on /exit/ from + the `with` context, so long as no exception is thrown. For example: + + :: + + with state_machine.transition_ctx('one','two', wait=5) as locked: + if locked: + # the state machine is currently locked in state 'one', and will + # transition to 'two' when the 'with' statement ends, so long as + # no exception is thrown. + print 'Currently locked in state one: %s' % state_machine['one'] + + else: + # The 'wait' timed out, and no lock has been acquired + print 'Timed out before entering state "one"' + + print 'Since no exception was thrown, we are now in state "two": %s' % state_machine['two'] + + + The other main difference between this method and `transition()` is that the + state machine is locked for the duration of the `with` statement. Normally, + after a `transition()` occurs, the state machine is immediately unlocked and + available to another thread to call `transition()` again. + ''' + + if not from_state in self.__states: + raise ValueError( "StateMachine does not contain from_state %s." % from_state ) + if not to_state in self.__states: + raise ValueError( "StateMachine does not contain to_state %s." % to_state ) + + return _StateCtx(self, from_state, to_state, wait) + + + def ensure(self, state, wait=0.0, block_on_transition=False ): + ''' + Ensure the state machine is currently in `state`, or wait until it enters `state`. + ''' + return self.ensure_any( (state,), wait=wait, block_on_transition=block_on_transition ) + + + def ensure_any(self, states, wait=0.0, block_on_transition=False): + ''' + Ensure we are currently in one of the given `states` or wait until + we enter one of those states. + + Note that due to the nature of the function, you cannot guarantee that + the entirety of some operation completes while you remain in a given + state. That would require acquiring and holding a lock, which + would mean no other threads could do the same. (You'd essentially + be serializing all of the threads that are 'ensuring' their tasks + occurred in some state. + ''' + if not (isinstance(states,tuple) or isinstance(states,list)): + raise ValueError('states arg should be a tuple or list') + + for state in states: + if not state in self.__states: + raise ValueError( "StateMachine does not contain state '%s'" % state ) + + # if we're in the middle of a transition, determine whether we should + # 'fall back' to the 'current' state, or wait for the new state, in order to + # avoid an operation occurring in the wrong state. + # TODO another option would be an ensure_ctx that uses a semaphore to allow + # threads to indicate they want to remain in a particular state. + + # will return immediately if no transition is in process. + if block_on_transition: + # we're not in the middle of a transition; don't hold the lock + if self.lock.acquire(False): self.lock.release() + # wait for the transition to complete + else: self.notifier.wait() + + start = time.time() + while not self.__current_state in states: + # detect timeout: + remainder = start + wait - time.time() + if remainder > 0: self.notifier.wait(remainder) + else: return False + return True + + + def reset(self): + # TODO need to lock before calling this? + self.transition(self.__current_state, self.__default_state) + + + def _set_state(self, state): #unsynchronized, only call internally after lock is acquired + self.__current_state = state + return state + + + def current_state(self): + ''' + Return the current state name. + ''' + return self.__current_state + + + def __getitem__(self, state): + ''' + Non-blocking, non-synchronized test to determine if we are in the given state. + Use `StateMachine.ensure(state)` to wait until the machine enters a certain state. + ''' + return self.__current_state == state + + def __str__(self): + return "".join(( "StateMachine(", ','.join(self.__states), "): ", self.__current_state )) + + + +class _StateCtx: + + def __init__( self, state_machine, from_state, to_state, wait ): + self.state_machine = state_machine + self.from_state = from_state + self.to_state = to_state + self.wait = wait + self._locked = False + + def __enter__(self): + start = time.time() + while not self.state_machine[ self.from_state ] or not self.state_machine.lock.acquire(False): + # detect timeout: + remainder = start + self.wait - time.time() + if remainder > 0: self.state_machine.notifier.wait(remainder) + else: + log.debug('StateMachine timeout while waiting for state: %s', self.from_state ) + return False + + self._locked = True # lock has been acquired at this point + self.state_machine.notifier.clear() + log.debug('StateMachine entered context in state: %s', + self.state_machine.current_state() ) + return True + + def __exit__(self, exc_type, exc_val, exc_tb): + if exc_val is not None: + log.exception( "StateMachine exception in context, remaining in state: %s\n%s:%s", + self.state_machine.current_state(), exc_type.__name__, exc_val ) + + if self._locked: + if exc_val is None: + log.debug(' ==== TRANSITION %s -> %s', + self.state_machine.current_state(), self.to_state) + self.state_machine._set_state( self.to_state ) + + self.state_machine.notifier.set() + self.state_machine.lock.release() + + return False # re-raise any exception + +if __name__ == '__main__': + + def callback(s, s2): + print 1, s.transition('on', 'off', wait=0.0, func=callback, args=[s,s2]) + print 2, s2.transition('off', 'on', func=callback, args=[s,s2]) + return True + + s = StateMachine(('off', 'on')) + s2 = StateMachine(('off', 'on')) + print 3, s.transition('off', 'on', wait=0.0, func=callback, args=[s,s2]), + print s.current_state(), s2.current_state() |