鋪墊
在大量的實踐中,似乎我們總是通過類似的方式來使用異步編程:
- 監聽事件
- 事件發生執行對應的回調函數
- 回調完成(可能產生新的事件添加進監聽隊列)
- 回到1,監聽事件
因此我們將這樣的異步模式稱為Reactor模式,例如在iOS開發中的Run Loop概念,實際上非常類似于Reactor loop,主線程的Run Loop監聽屏幕UI事件,一旦發生UI事件則執行對應的事件處理代碼,還可以通過GCD等方式產生事件至主線程執行。
上圖是boost對Reactor模式的描繪,Twisted的設計就是基于這樣的Reactor模式,Twisted程序就是在等待事件、處理事件的過程中不斷循環。
from twisted.internet import reactor reactor.run()
reactor是Twisted程序中的單例對象。
reactor
reactor是事件管理器,用于注冊、注銷事件,運行事件循環,當事件發生時調用回調函數處理。關于reactor有下面幾個結論:
- Twisted的reactor只有通過調用reactor.run()來啟動。
- reactor循環是在其開始的進程中運行,也就是運行在主進程中。
- 一旦啟動,就會一直運行下去。reactor就會在程序的控制下(或者具體在一個啟動它的線程的控制下)。
- reactor循環并不會消耗任何CPU的資源。
- 并不需要顯式的創建reactor,只需要引入就OK了。
最后一條需要解釋清楚。在Twisted中,reactor是Singleton(也就是單例模式),即在一個程序中只能有一個reactor,并且只要你引入它就相應地創建一個。上面引入的方式這是twisted默認使用的方法,當然了,twisted還有其它可以引入reactor的方法。例如,可以使用twisted.internet.pollreactor中的系統調用來poll來代替select方法。
若使用其它的reactor,需要在引入twisted.internet.reactor前安裝它。下面是安裝pollreactor的方法:
from twisted.internet import pollreactor pollreactor.install()
如果你沒有安裝其它特殊的reactor而引入了twisted.internet.reactor,那么Twisted會根據操作系統安裝默認的reactor。正因為如此,習慣性做法不要在最頂層的模塊內引入reactor以避免安裝默認reactor,而是在你要使用reactor的區域內安裝。
下面是使用 pollreactor重寫上上面的程序:
from twited.internet import pollreactor pollreactor.install() from twisted.internet import reactor reactor.run()
那么reactor是如何實現單例的?來看一下from twisted.internet import reactor做了哪些事情就并明白了。
下面是twisted/internet/reactor.py的部分代碼:
# twisted/internet/reactor.py import sys del sys.modules['twisted.internet.reactor'] from twisted.internet import default default.install()
注:Python中所有加載到內存的模塊都放在sys.modules,它是一個全局字典。當import一個模塊時首先會在這個列表中查找是否已經加載了此模塊,如果加載了則只是將模塊的名字加入到正在調用import的模塊的命名空間中。如果沒有加載則從sys.path目錄中按照模塊名稱查找模塊文件,找到后將模塊載入內存,并加入到sys.modules中,并將名稱導入到當前的命名空間中。
假如我們是第一次運行from twisted.internet import reactor,因為sys.modules中還沒有twisted.internet.reactor,所以會運行reactory.py中的代碼,安裝默認的reactor。之后,如果導入的話,因為sys.modules中已存在該模塊,所以會直接將sys.modules中的twisted.internet.reactor導入到當前命名空間。
default中的install:
# twisted/internet/default.py def _getInstallFunction(platform): """ Return a function to install the reactor most suited for the given platform. @param platform: The platform for which to select a reactor. @type platform: L{twisted.python.runtime.Platform} @return: A zero-argument callable which will install the selected reactor. """ try: if platform.isLinux(): try: from twisted.internet.epollreactor import install except ImportError: from twisted.internet.pollreactor import install elif platform.getType() == 'posix' and not platform.isMacOSX(): from twisted.internet.pollreactor import install else: from twisted.internet.selectreactor import install except ImportError: from twisted.internet.selectreactor import install return install install = _getInstallFunction(platform)
很明顯,default中會根據平臺獲取相應的install。Linux下會首先使用epollreactor,如果內核還不支持,就只能使用pollreactor。Mac平臺使用pollreactor,windows使用selectreactor。每種install的實現差不多,這里我們抽取selectreactor中的install來看看。
# twisted/internet/selectreactor.py: def install(): """Configure the twisted mainloop to be run using the select() reactor. """ # 單例 reactor = SelectReactor() from twisted.internet.main import installReactor installReactor(reactor) # twisted/internet/main.py: def installReactor(reactor): """ Install reactor C{reactor}. @param reactor: An object that provides one or more IReactor* interfaces. """ # this stuff should be common to all reactors. import twisted.internet import sys if 'twisted.internet.reactor' in sys.modules: raise error.ReactorAlreadyInstalledError("reactor already installed") twisted.internet.reactor = reactor sys.modules['twisted.internet.reactor'] = reactor
在installReactor中,向sys.modules添加twisted.internet.reactor鍵,值就是再install中創建的單例reactor。以后要使用reactor,就會導入這個單例了。
SelectReactor # twisted/internet/selectreactor.py @implementer(IReactorFDSet) class SelectReactor(posixbase.PosixReactorBase, _extraBase)
IReactorFDSet接口主要對描述符的獲取、添加、刪除等操作的方法。這些方法看名字就能知道意思,所以我就沒有加注釋。
# twisted/internet/interfaces.py class IReactorFDSet(Interface): def addReader(reader): def addWriter(writer): def removeReader(reader): def removeWriter(writer): def removeAll(): def getReaders(): def getWriters(): reactor.listenTCP()
示例中的reactor.listenTCP()注冊了一個監聽事件,它是父類PosixReactorBase中方法。
# twisted/internet/posixbase.py @implementer(IReactorTCP, IReactorUDP, IReactorMulticast) class PosixReactorBase(_SignalReactorMixin, _DisconnectSelectableMixin, ReactorBase): def listenTCP(self, port, factory, backlog=50, interface=''): p = tcp.Port(port, factory, backlog, interface, self) p.startListening() return p # twisted/internet/tcp.py @implementer(interfaces.IListeningPort) class Port(base.BasePort, _SocketCloser): def __init__(self, port, factory, backlog=50, interface='', reactor=None): """Initialize with a numeric port to listen on. """ base.BasePort.__init__(self, reactor=reactor) self.port = port self.factory = factory self.backlog = backlog if abstract.isIPv6Address(interface): self.addressFamily = socket.AF_INET6 self._addressType = address.IPv6Address self.interface = interface ... def startListening(self): """Create and bind my socket, and begin listening on it. 創建并綁定套接字,開始監聽。 This is called on unserialization, and must be called after creating a server to begin listening on the specified port. """ if self._preexistingSocket is None: # Create a new socket and make it listen try: # 創建套接字 skt = self.createInternetSocket() if self.addressFamily == socket.AF_INET6: addr = _resolveIPv6(self.interface, self.port) else: addr = (self.interface, self.port) # 綁定 skt.bind(addr) except socket.error as le: raise CannotListenError(self.interface, self.port, le) # 監聽 skt.listen(self.backlog) else: # Re-use the externally specified socket skt = self._preexistingSocket self._preexistingSocket = None # Avoid shutting it down at the end. self._shouldShutdown = False # Make sure that if we listened on port 0, we update that to # reflect what the OS actually assigned us. self._realPortNumber = skt.getsockname()[1] log.msg("%s starting on %s" % ( self._getLogPrefix(self.factory), self._realPortNumber)) # The order of the next 5 lines is kind of bizarre. If no one # can explain it, perhaps we should re-arrange them. self.factory.doStart() self.connected = True self.socket = skt self.fileno = self.socket.fileno self.numberAccepts = 100 # startReading調用reactor的addReader方法將Port加入讀集合 self.startReading()
整個邏輯很簡單,和正常的server端一樣,創建套接字、綁定、監聽。不同的是將套接字的描述符添加到了reactor的讀集合。那么假如有了client連接過來的話,reactor會監控到,然后觸發事件處理程序。
reacotr.run()事件主循環
# twisted/internet/posixbase.py @implementer(IReactorTCP, IReactorUDP, IReactorMulticast) class PosixReactorBase(_SignalReactorMixin, _DisconnectSelectableMixin, ReactorBase) # twisted/internet/base.py class _SignalReactorMixin(object): def startRunning(self, installSignalHandlers=True): """ PosixReactorBase的父類_SignalReactorMixin和ReactorBase都有該函數,但是 _SignalReactorMixin在前,安裝mro順序的話,會先調用_SignalReactorMixin中的。 """ self._installSignalHandlers = installSignalHandlers ReactorBase.startRunning(self) def run(self, installSignalHandlers=True): self.startRunning(installSignalHandlers=installSignalHandlers) self.mainLoop() def mainLoop(self): while self._started: try: while self._started: # Advance simulation time in delayed event # processors. self.runUntilCurrent() t2 = self.timeout() t = self.running and t2 # doIteration是關鍵,select,poll,epool實現各有不同 self.doIteration(t) except: log.msg("Unexpected error in main loop.") log.err() else: log.msg('Main loop terminated.')
mianLoop就是最終的主循環了,在循環中,調用doIteration方法監控讀寫描述符的集合,一旦發現有描述符準備好讀寫,就會調用相應的事件處理程序。
# twisted/internet/selectreactor.py @implementer(IReactorFDSet) class SelectReactor(posixbase.PosixReactorBase, _extraBase): def __init__(self): """ Initialize file descriptor tracking dictionaries and the base class. """ self._reads = set() self._writes = set() posixbase.PosixReactorBase.__init__(self) def doSelect(self, timeout): """ Run one iteration of the I/O monitor loop. This will run all selectables who had input or output readiness waiting for them. """ try: # 調用select方法監控讀寫集合,返回準備好讀寫的描述符 r, w, ignored = _select(self._reads, self._writes, [], timeout) except ValueError: # Possibly a file descriptor has gone negative? self._preenDescriptors() return except TypeError: # Something *totally* invalid (object w/o fileno, non-integral # result) was passed log.err() self._preenDescriptors() return except (select.error, socket.error, IOError) as se: # select(2) encountered an error, perhaps while calling the fileno() # method of a socket. (Python 2.6 socket.error is an IOError # subclass, but on Python 2.5 and earlier it is not.) if se.args[0] in (0, 2): # windows does this if it got an empty list if (not self._reads) and (not self._writes): return else: raise elif se.args[0] == EINTR: return elif se.args[0] == EBADF: self._preenDescriptors() return else: # OK, I really don't know what's going on. Blow up. raise _drdw = self._doReadOrWrite _logrun = log.callWithLogger for selectables, method, fdset in ((r, "doRead", self._reads), (w,"doWrite", self._writes)): for selectable in selectables: # if this was disconnected in another thread, kill it. # ^^^^ --- what the !@#*? serious! -exarkun if selectable not in fdset: continue # This for pausing input when we're not ready for more. # 調用_doReadOrWrite方法 _logrun(selectable, _drdw, selectable, method) doIteration = doSelect def _doReadOrWrite(self, selectable, method): try: # 調用method,doRead或者是doWrite, # 這里的selectable可能是我們監聽的tcp.Port why = getattr(selectable, method)() except: why = sys.exc_info()[1] log.err() if why: self._disconnectSelectable(selectable, why, method=="doRead")
那么假如客戶端有連接請求了,就會調用讀集合中tcp.Port的doRead方法。
# twisted/internet/tcp.py @implementer(interfaces.IListeningPort) class Port(base.BasePort, _SocketCloser): def doRead(self): """Called when my socket is ready for reading. 當套接字準備好讀的時候調用 This accepts a connection and calls self.protocol() to handle the wire-level protocol. """ try: if platformType == "posix": numAccepts = self.numberAccepts else: numAccepts = 1 for i in range(numAccepts): if self.disconnecting: return try: # 調用accept skt, addr = self.socket.accept() except socket.error as e: if e.args[0] in (EWOULDBLOCK, EAGAIN): self.numberAccepts = i break elif e.args[0] == EPERM: continue elif e.args[0] in (EMFILE, ENOBUFS, ENFILE, ENOMEM, ECONNABORTED): log.msg("Could not accept new connection (%s)" % ( errorcode[e.args[0]],)) break raise fdesc._setCloseOnExec(skt.fileno()) protocol = self.factory.buildProtocol(self._buildAddr(addr)) if protocol is None: skt.close() continue s = self.sessionno self.sessionno = s+1 # transport初始化的過程中,會將自身假如到reactor的讀集合中,那么當它準備 # 好讀的時候,就可以調用它的doRead方法讀取客戶端發過來的數據了 transport = self.transport(skt, protocol, addr, self, s, self.reactor) protocol.makeConnection(transport) else: self.numberAccepts = self.numberAccepts+20 except: log.deferr()
doRead方法中,調用accept產生了用于接收客戶端數據的套接字,將套接字與transport綁定,然后把transport加入到reactor的讀集合。當客戶端有數據到來時,就會調用transport的doRead方法進行數據讀取了。
Connection是Server(transport實例的類)的父類,它實現了doRead方法。
# twisted/internet/tcp.py @implementer(interfaces.ITCPTransport, interfaces.ISystemHandle) class Connection(_TLSConnectionMixin, abstract.FileDescriptor, _SocketCloser, _AbortingMixin): def doRead(self): try: # 接收數據 data = self.socket.recv(self.bufferSize) except socket.error as se: if se.args[0] == EWOULDBLOCK: return else: return main.CONNECTION_LOST return self._dataReceived(data) def _dataReceived(self, data): if not data: return main.CONNECTION_DONE # 調用我們自定義protocol的dataReceived方法處理數據 rval = self.protocol.dataReceived(data) if rval is not None: offender = self.protocol.dataReceived warningFormat = ( 'Returning a value other than None from %(fqpn)s is ' 'deprecated since %(version)s.') warningString = deprecate.getDeprecationWarningString( offender, versions.Version('Twisted', 11, 0, 0), format=warningFormat) deprecate.warnAboutFunction(offender, warningString) return rval
_dataReceived中調用了示例中我們自定義的EchoProtocol的dataReceived方法處理數據。
至此,一個簡單的流程,從創建監聽事件,到接收客戶端數據就此結束了。