############################################################################## # # Copyright (c) 2002 Nexedi SARL and Contributors. All Rights Reserved. # Jean-Paul Smets-Solanes <jp@nexedi.com> # # WARNING: This program as such is intended to be used by professional # programmers who take the whole responsability of assessing all potential # consequences resulting from its eventual inadequacies and bugs # End users who are looking for a ready-to-use solution with commercial # garantees and support are strongly adviced to contract a Free Software # Service Company # # This program is Free Software; you can redistribute it and/or # modify it under the terms of the GNU General Public License # as published by the Free Software Foundation; either version 2 # of the License, or (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. # ############################################################################## import socket import urllib import threading import sys from types import StringType from collections import defaultdict from cPickle import dumps, loads from Products.CMFCore import permissions as CMFCorePermissions from Products.ERP5Type.Tool.BaseTool import BaseTool from Products.CMFActivity.ActiveResult import ActiveResult from Products.CMFActivity.ActiveObject import DEFAULT_ACTIVITY from Products.CMFActivity.ActivityConnection import ActivityConnection from Products.PythonScripts.Utility import allow_class from AccessControl import ClassSecurityInfo, Permissions from AccessControl.SecurityManagement import newSecurityManager from AccessControl.SecurityManagement import noSecurityManager from AccessControl.SecurityManagement import setSecurityManager from AccessControl.SecurityManagement import getSecurityManager from AccessControl.User import system as system_user from Products.CMFCore.utils import UniqueObject from Products.ERP5Type.Globals import InitializeClass, DTMLFile from Acquisition import aq_base, aq_inner, aq_parent from ActivityBuffer import ActivityBuffer from ActivityRuntimeEnvironment import BaseMessage from zExceptions import ExceptionFormatter from BTrees.OIBTree import OIBTree from Zope2 import app from Products.ERP5Type.UnrestrictedMethod import PrivilegedUser from zope.site.hooks import setSite import transaction from App.config import getConfiguration import Products.Localizer.patches localizer_lock = Products.Localizer.patches._requests_lock localizer_contexts = Products.Localizer.patches._requests LocalizerContext = lambda request: request from Products.MailHost.MailHost import MailHostError from zLOG import LOG, INFO, WARNING, ERROR import warnings from time import time try: from Products.TimerService import getTimerService except ImportError: def getTimerService(self): pass from traceback import format_list, extract_stack # Using a RAM property (not a property of an instance) allows # to prevent from storing a state in the ZODB (and allows to restart...) active_threads = 0 max_active_threads = 1 # 2 will cause more bug to appear (he he) tic_lock = threading.Lock() # A RAM based lock to prevent too many concurrent tic() calls timerservice_lock = threading.Lock() # A RAM based lock to prevent TimerService spamming when busy is_running_lock = threading.Lock() currentNode = None _server_address = None ROLE_IDLE = 0 ROLE_PROCESSING = 1 # Logging channel definitions import logging # Main logging channel activity_logger = logging.getLogger('CMFActivity') # Some logging subchannels activity_tracking_logger = logging.getLogger('Tracking') activity_timing_logger = logging.getLogger('CMFActivity.TimingLog') # Direct logging to "[instancehome]/log/CMFActivity.log", if this directory exists. # Otherwise, it will end up in root logging facility (ie, event.log). from App.config import getConfiguration import os instancehome = getConfiguration().instancehome if instancehome is not None: log_directory = os.path.join(instancehome, 'log') if os.path.isdir(log_directory): from Signals import Signals from ZConfig.components.logger.loghandler import FileHandler log_file_handler = FileHandler(os.path.join(log_directory, 'CMFActivity.log')) # Default zope log format string borrowed from # ZConfig/components/logger/factory.xml, but without the extra "------" # line separating entries. log_file_handler.setFormatter(logging.Formatter("%(asctime)s %(levelname)s %(name)s %(message)s", "%Y-%m-%dT%H:%M:%S")) Signals.registerZopeSignals([log_file_handler]) activity_logger.addHandler(log_file_handler) activity_logger.propagate = 0 def activity_timing_method(method, args, kw): begin = time() try: return method(*args, **kw) finally: end = time() activity_timing_logger.info('%.02fs: %r(*%r, **%r)' % (end - begin, method, args, kw)) def getServerAddress(): """ Return current server address """ global _server_address if _server_address is None: ip = port = '' from asyncore import socket_map for k, v in socket_map.items(): if hasattr(v, 'addr'): # see Zope/lib/python/App/ApplicationManager.py: def getServers(self) type = str(getattr(v, '__class__', 'unknown')) if type == 'ZServer.HTTPServer.zhttp_server': ip, port = v.addr break if ip == '0.0.0.0': ip = socket.gethostbyname(socket.gethostname()) _server_address = '%s:%s' %(ip, port) return _server_address def getCurrentNode(): """ Return current node identifier """ global currentNode if currentNode is None: currentNode = getattr( getConfiguration(), 'product_config', {}, ).get('cmfactivity', {}).get('node-id') if currentNode is None: warnings.warn('Node name auto-generation is deprecated, please add a' '\n' '<product-config CMFActivity>\n' ' node-id = ...\n' '</product-config>\n' 'section in your zope.conf, replacing "..." with a cluster-unique ' 'node identifier.', DeprecationWarning) currentNode = getServerAddress() return currentNode # Here go ActivityBuffer instances # Structure: # global_activity_buffer[activity_tool_path][thread_id] = ActivityBuffer global_activity_buffer = defaultdict(dict) from thread import get_ident MESSAGE_NOT_EXECUTED = 0 MESSAGE_EXECUTED = 1 MESSAGE_NOT_EXECUTABLE = 2 class SkippedMessage(Exception): pass class Message(BaseMessage): """Activity Message Class. Message instances are stored in an activity queue, inside the Activity Tool. """ active_process = None active_process_uid = None call_traceback = None exc_info = None is_executed = MESSAGE_NOT_EXECUTED processing = None traceback = None oid = None is_registered = False def __init__( self, url, oid, active_process, active_process_uid, activity_kw, method_id, args, kw, request=None, portal_activities=None, ): self.object_path = url self.oid = oid self.active_process = active_process self.active_process_uid = active_process_uid self.activity_kw = activity_kw self.method_id = method_id self.args = args self.kw = kw if getattr(portal_activities, 'activity_creation_trace', False): # Save current traceback, to make it possible to tell where a message # was generated. # Strip last stack entry, since it will always be the same. self.call_traceback = ''.join(format_list(extract_stack()[:-1])) self.user_name = getSecurityManager().getUser().getIdOrUserName() # Store REQUEST Info self.request_info = {} if request is not None: if 'SERVER_URL' in request.other: self.request_info['SERVER_URL'] = request.other['SERVER_URL'] if 'VirtualRootPhysicalPath' in request.other: self.request_info['VirtualRootPhysicalPath'] = \ request.other['VirtualRootPhysicalPath'] if 'HTTP_ACCEPT_LANGUAGE' in request.environ: self.request_info['HTTP_ACCEPT_LANGUAGE'] = \ request.environ['HTTP_ACCEPT_LANGUAGE'] self.request_info['_script'] = list(request._script) @staticmethod def load(s, **kw): self = loads(s) self.__dict__.update(kw) return self dump = dumps def getGroupId(self): get = self.activity_kw.get group_method_id = get('group_method_id', '') if group_method_id is None: group_method_id = 'portal_activities/dummyGroupMethod/' + self.method_id return group_method_id + '\0' + get('group_id', '') def _getObject(self, activity_tool): obj = activity_tool.getPhysicalRoot() for id in self.object_path[1:]: obj = obj[id] return obj def getObject(self, activity_tool): """return the object referenced in this message.""" try: obj = self._getObject(activity_tool) except KeyError: LOG('CMFActivity', WARNING, "Message dropped (no object found at path %r)" % (self.object_path,), error=sys.exc_info()) self.setExecutionState(MESSAGE_NOT_EXECUTABLE) else: if (self.oid and self.oid != getattr(aq_base(obj), '_p_oid', None) and # XXX: BusinessTemplate must be fixed to preserve OID 'portal_workflow' not in self.object_path): raise ValueError("OID mismatch for %r" % obj) return obj def getObjectList(self, activity_tool): """return the list of object that can be expanded from this message An expand method is used to expand the list of objects and to turn a big recursive transaction affecting many objects into multiple transactions affecting only one object at a time (this can prevent duplicated method calls).""" obj = self.getObject(activity_tool) if obj is None: return () if 'expand_method_id' in self.activity_kw: return getattr(obj, self.activity_kw['expand_method_id'])() return obj, def getObjectCount(self, activity_tool): if 'expand_method_id' in self.activity_kw: try: obj = self._getObject(activity_tool) return len(getattr(obj, self.activity_kw['expand_method_id'])()) except StandardError: pass return 1 def changeUser(self, user_name, activity_tool): """restore the security context for the calling user.""" portal = activity_tool.getPortalObject() portal_uf = portal.acl_users uf = portal_uf user = uf.getUserById(user_name) # if the user is not found, try to get it from a parent acl_users # XXX this is still far from perfect, because we need to store all # information about the user (like original user folder, roles) to # replay the activity with exactly the same security context as if # it had been executed without activity. if user is None: uf = portal.aq_parent.acl_users user = uf.getUserById(user_name) if user is None and user_name == system_user.getUserName(): # The following logic partly comes from unrestricted_apply() # implementation in ERP5Type.UnrestrictedMethod but we get roles # from the portal to have more roles. uf = portal_uf role_list = uf.valid_roles() user = PrivilegedUser(user_name, None, role_list, ()).__of__(uf) if user is not None: user = user.__of__(uf) newSecurityManager(None, user) transaction.get().setUser(user_name, '/'.join(uf.getPhysicalPath())) else : LOG("CMFActivity", WARNING, "Unable to find user %r in the portal" % user_name) noSecurityManager() return user def activateResult(self, active_process, result, object): if not isinstance(result, ActiveResult): result = ActiveResult(result=result) signature = self.activity_kw.get('signature') if signature: result.edit(id=signature) # XXX Allow other method_id in future result.edit(object_path=object, method_id=self.method_id) active_process.postResult(result) def __call__(self, activity_tool): try: obj = self.getObject(activity_tool) if obj is not None: old_security_manager = getSecurityManager() try: # Change user if required (TO BE DONE) # We will change the user only in order to execute this method self.changeUser(self.user_name, activity_tool) # XXX: There is no check to see if user is allowed to access # that method ! method = getattr(obj, self.method_id) transaction.get().note( 'CMFActivity ' + '/'.join(self.object_path) + '/' + self.method_id ) # Store site info setSite(activity_tool.getParentValue()) if activity_tool.activity_timing_log: result = activity_timing_method(method, self.args, self.kw) else: result = method(*self.args, **self.kw) finally: setSecurityManager(old_security_manager) if method is not None: if self.active_process and result is not None: self.activateResult( activity_tool.unrestrictedTraverse(self.active_process), result, obj) self.setExecutionState(MESSAGE_EXECUTED) except: self.setExecutionState(MESSAGE_NOT_EXECUTED, context=activity_tool) def notifyUser(self, activity_tool, retry=False): """Notify the user that the activity failed.""" portal = activity_tool.getPortalObject() user_email = portal.getProperty('email_to_address', portal.getProperty('email_from_address')) email_from_name = portal.getProperty('email_from_name', portal.getProperty('email_from_address')) fail_count = self.line.retry + 1 if retry: message = "Pending activity already failed %s times" % fail_count else: message = "Activity failed" path = '/'.join(self.object_path) mail_text = """From: %s <%s> To: %s Subject: %s: %s/%s Node: %s Failures: %s User name: %r Uid: %u Document: %s Method: %s Arguments: %r Named Parameters: %r """ % (email_from_name, activity_tool.email_from_address, user_email, message, path, self.method_id, getCurrentNode(), fail_count, self.user_name, self.line.uid, path, self.method_id, self.args, self.kw) if self.traceback: mail_text += '\nException:\n' + self.traceback if self.call_traceback: mail_text += '\nCreated at:\n' + self.call_traceback try: portal.MailHost.send(mail_text) except (socket.error, MailHostError), message: LOG('ActivityTool.notifyUser', WARNING, 'Mail containing failure information failed to be sent: %s' % message) def reactivate(self, activity_tool, activity=DEFAULT_ACTIVITY): # Reactivate the original object. obj = self._getObject(activity_tool) old_security_manager = getSecurityManager() try: # Change user if required (TO BE DONE) # We will change the user only in order to execute this method user = self.changeUser(self.user_name, activity_tool) active_obj = obj.activate(activity=activity, **self.activity_kw) getattr(active_obj, self.method_id)(*self.args, **self.kw) finally: # Use again the previous user setSecurityManager(old_security_manager) def setExecutionState(self, is_executed, exc_info=None, log=True, context=None): """ Set message execution state. is_executed can be one of MESSAGE_NOT_EXECUTED, MESSAGE_EXECUTED and MESSAGE_NOT_EXECUTABLE (variables defined above). exc_info must be - if given - similar to sys.exc_info() return value. log must be - if given - True or False. If True, a log line will be emited with failure details. This parameter should only be used when invoking this method on a list of messages to avoid log flood. It is caller's responsability to output a log line summing up all errors, and to store error in Zope's error_log. context must be - if given - an object wrapped in acquisition context. It is used to access Zope's error_log object. It is not used if log is False. If given state is not MESSAGE_EXECUTED, it will also store given exc_info. If not given, it will extract one using sys.exc_info(). If final exc_info does not contain any exception, current stack trace will be stored instead: it will hopefuly help understand why message is in an error state. """ assert is_executed in (MESSAGE_NOT_EXECUTED, MESSAGE_EXECUTED, MESSAGE_NOT_EXECUTABLE) self.is_executed = is_executed if is_executed == MESSAGE_NOT_EXECUTED: if not exc_info: exc_info = sys.exc_info() if self.on_error_callback is not None: self.exc_info = exc_info self.exc_type = exc_info[0] if exc_info[0] is None: # Raise a dummy exception, ignore it, fetch it and use it as if it was the error causing message non-execution. This will help identifyting the cause of this misbehaviour. try: raise Exception, 'Message execution failed, but there is no exception to explain it. This is a dummy exception so that one can track down why we end up here outside of an exception handling code path.' except Exception: exc_info = sys.exc_info() elif exc_info[0] is SkippedMessage: return if log: LOG('ActivityTool', WARNING, 'Could not call method %s on object %s. Activity created at:\n%s' % (self.method_id, self.object_path, self.call_traceback), error=exc_info) # push the error in ZODB error_log error_log = getattr(context, 'error_log', None) if error_log is not None: error_log.raising(exc_info) self.traceback = ''.join(ExceptionFormatter.format_exception(*exc_info)[1:]) def getExecutionState(self): return self.is_executed class GroupedMessage(object): __slots__ = 'object', '_message', 'result', 'exc_info' def __init__(self, object, message): self.object = object self._message = message args = property(lambda self: self._message.args) kw = property(lambda self: self._message.kw) def raised(self, exc_info=None): self.exc_info = exc_info or sys.exc_info() try: del self.result except AttributeError: pass # XXX: Allowing restricted code to implement a grouping method is questionable # but there already exist some. __parent__ = property(lambda self: self.object) # for object _guarded_writes = 1 # for result allow_class(GroupedMessage) # Activity Registration def activity_dict(): from Activity import SQLDict, SQLQueue, SQLJoblib return {k: getattr(v, k)() for k, v in locals().iteritems()} activity_dict = activity_dict() class Method(object): __slots__ = ( '_portal_activities', '_passive_url', '_passive_oid', '_activity', '_active_process', '_active_process_uid', '_kw', '_method_id', '_request', ) def __init__(self, portal_activities, passive_url, passive_oid, activity, active_process, active_process_uid, kw, method_id, request): self._portal_activities = portal_activities self._passive_url = passive_url self._passive_oid = passive_oid self._activity = activity self._active_process = active_process self._active_process_uid = active_process_uid self._kw = kw self._method_id = method_id self._request = request def __call__(self, *args, **kw): portal_activities = self._portal_activities m = Message( url=self._passive_url, oid=self._passive_oid, active_process=self._active_process, active_process_uid=self._active_process_uid, activity_kw=self._kw, method_id=self._method_id, args=args, kw=kw, request=self._request, portal_activities=portal_activities, ) portal_activities.getActivityBuffer().deferredQueueMessage( portal_activities, activity_dict[self._activity], m) if portal_activities.activity_tracking and m.is_registered: activity_tracking_logger.info('queuing message: activity=%s, object_path=%s, method_id=%s, args=%s, kw=%s, activity_kw=%s, user_name=%s' % (self._activity, '/'.join(m.object_path), m.method_id, m.args, m.kw, m.activity_kw, m.user_name)) allow_class(Method) class ActiveWrapper(object): __slots__ = ( '__portal_activities', '__passive_url', '__passive_oid', '__activity', '__active_process', '__active_process_uid', '__kw', '__request', ) # Shortcut security lookup (avoid calling __getattr__) __parent__ = None def __init__(self, portal_activities, url, oid, activity, active_process, active_process_uid, kw, request): # second parameter can be an object or an object's path self.__portal_activities = portal_activities self.__passive_url = url self.__passive_oid = oid self.__activity = activity self.__active_process = active_process self.__active_process_uid = active_process_uid self.__kw = kw self.__request = request def __getattr__(self, name): return Method( self.__portal_activities, self.__passive_url, self.__passive_oid, self.__activity, self.__active_process, self.__active_process_uid, self.__kw, name, self.__request, ) def __repr__(self): return '<%s at 0x%x to %s>' % (self.__class__.__name__, id(self), self.__passive_url) # True when activities cannot be executing any more. has_processed_shutdown = False def cancelProcessShutdown(): """ This method reverts the effect of calling "process_shutdown" on activity tool. """ global has_processed_shutdown is_running_lock.release() has_processed_shutdown = False class ActivityTool (BaseTool): """ ActivityTool is the central point for activity management. Improvement to consider to reduce locks: Idea 1: create an SQL tool which accumulate queries and executes them at the end of a transaction, thus allowing all SQL transaction to happen in a very short time (this would also be a great way of using MyISAM tables) Idea 2: do the same at the level of ActivityTool Idea 3: do the same at the level of each activity (ie. queueMessage accumulates and fires messages at the end of the transactino) """ id = 'portal_activities' meta_type = 'CMF Activity Tool' portal_type = 'Activity Tool' allowed_types = ( 'CMF Active Process', ) security = ClassSecurityInfo() manage_options = tuple( [ { 'label' : 'Overview', 'action' : 'manage_overview' } , { 'label' : 'Activities', 'action' : 'manageActivities' } , { 'label' : 'LoadBalancing', 'action' : 'manageLoadBalancing'} , { 'label' : 'Advanced', 'action' : 'manageActivitiesAdvanced' } , ] + list(BaseTool.manage_options)) security.declareProtected( CMFCorePermissions.ManagePortal , 'manageActivities' ) manageActivities = DTMLFile( 'dtml/manageActivities', globals() ) security.declareProtected( CMFCorePermissions.ManagePortal , 'manageActivitiesAdvanced' ) manageActivitiesAdvanced = DTMLFile( 'dtml/manageActivitiesAdvanced', globals() ) security.declareProtected( CMFCorePermissions.ManagePortal , 'manage_overview' ) manage_overview = DTMLFile( 'dtml/explainActivityTool', globals() ) security.declareProtected( CMFCorePermissions.ManagePortal , 'manageLoadBalancing' ) manageLoadBalancing = DTMLFile( 'dtml/manageLoadBalancing', globals() ) distributingNode = '' _nodes = () activity_creation_trace = False activity_tracking = False activity_timing_log = False cancel_and_invoke_links_hidden = False # Filter content (ZMI)) def filtered_meta_types(self, user=None): # Filters the list of available meta types. all = BaseTool.filtered_meta_types(self) meta_types = [] for meta_type in self.all_meta_types(): if meta_type['name'] in self.allowed_types: meta_types.append(meta_type) return meta_types def getSQLConnection(self): return self.aq_inner.aq_parent.cmf_activity_sql_connection() def maybeMigrateConnectionClass(self): connection_id = 'cmf_activity_sql_connection' sql_connection = getattr(self, connection_id, None) if (sql_connection is not None and not isinstance(sql_connection, ActivityConnection)): # SQL Connection migration is needed LOG('ActivityTool', WARNING, "Migrating MySQL Connection class") parent = aq_parent(aq_inner(sql_connection)) parent._delObject(sql_connection.getId()) new_sql_connection = ActivityConnection(connection_id, sql_connection.title, sql_connection.connection_string) parent._setObject(connection_id, new_sql_connection) security.declarePrivate('initialize') def initialize(self): self.maybeMigrateConnectionClass() for activity in activity_dict.itervalues(): activity.initialize(self, clear=False) def _callSafeFunction(self, batch_function): return batch_function() security.declareProtected(Permissions.manage_properties, 'isSubscribed') def isSubscribed(self): """ return True, if we are subscribed to TimerService. Otherwise return False. """ service = getTimerService(self) if service: path = '/'.join(self.getPhysicalPath()) return path in service.lisSubscriptions() LOG('ActivityTool', INFO, 'TimerService not available') return False security.declareProtected(Permissions.manage_properties, 'subscribe') def subscribe(self, REQUEST=None, RESPONSE=None): """ subscribe to the global Timer Service """ service = getTimerService(self) url = '%s/manageLoadBalancing?manage_tabs_message=' %self.absolute_url() if not service: LOG('ActivityTool', INFO, 'TimerService not available') url += urllib.quote('TimerService not available') else: service.subscribe(self) url += urllib.quote("Subscribed to Timer Service") if RESPONSE is not None: RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'unsubscribe') def unsubscribe(self, REQUEST=None, RESPONSE=None): """ unsubscribe from the global Timer Service """ service = getTimerService(self) url = '%s/manageLoadBalancing?manage_tabs_message=' %self.absolute_url() if not service: LOG('ActivityTool', INFO, 'TimerService not available') url += urllib.quote('TimerService not available') else: service.unsubscribe(self) url += urllib.quote("Unsubscribed from Timer Service") if RESPONSE is not None: RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'isActivityTrackingEnabled') def isActivityTrackingEnabled(self): return self.activity_tracking security.declareProtected(Permissions.manage_properties, 'manage_enableActivityTracking') def manage_enableActivityTracking(self, REQUEST=None, RESPONSE=None): """ Enable activity tracing. """ self.activity_tracking = True if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Tracking log enabled') RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'manage_disableActivityTracking') def manage_disableActivityTracking(self, REQUEST=None, RESPONSE=None): """ Disable activity tracing. """ self.activity_tracking = False if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Tracking log disabled') RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'isActivityTimingLoggingEnabled') def isActivityTimingLoggingEnabled(self): return self.activity_timing_log security.declareProtected(Permissions.manage_properties, 'manage_enableActivityTimingLogging') def manage_enableActivityTimingLogging(self, REQUEST=None, RESPONSE=None): """ Enable activity timing logging. """ self.activity_timing_log = True if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Timing log enabled') RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'manage_disableActivityTimingLogging') def manage_disableActivityTimingLogging(self, REQUEST=None, RESPONSE=None): """ Disable activity timing logging. """ self.activity_timing_log = False if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Timing log disabled') RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'isActivityCreationTraceEnabled') def isActivityCreationTraceEnabled(self): return self.activity_creation_trace security.declareProtected(Permissions.manage_properties, 'manage_enableActivityCreationTrace') def manage_enableActivityCreationTrace(self, REQUEST=None, RESPONSE=None): """ Enable activity creation trace. """ self.activity_creation_trace = True if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Activity creation trace enabled') RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'manage_disableActivityCreationTrace') def manage_disableActivityCreationTrace(self, REQUEST=None, RESPONSE=None): """ Disable activity creation trace. """ self.activity_creation_trace = False if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Activity creation trace disabled') RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'isCancelAndInvokeLinksHidden') def isCancelAndInvokeLinksHidden(self): return self.cancel_and_invoke_links_hidden security.declareProtected(Permissions.manage_properties, 'manage_hideCancelAndInvokeLinks') def manage_hideCancelAndInvokeLinks(self, REQUEST=None, RESPONSE=None): """ """ self.cancel_and_invoke_links_hidden = True if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Cancel and invoke links hidden') RESPONSE.redirect(url) security.declareProtected(Permissions.manage_properties, 'manage_showCancelAndInvokeLinks') def manage_showCancelAndInvokeLinks(self, REQUEST=None, RESPONSE=None): """ """ self.cancel_and_invoke_links_hidden = False if RESPONSE is not None: url = '%s/manageActivitiesAdvanced?manage_tabs_message=' % self.absolute_url() url += urllib.quote('Cancel and invoke links visible') RESPONSE.redirect(url) security.declarePrivate('manage_beforeDelete') def manage_beforeDelete(self, item, container): self.unsubscribe() BaseTool.inheritedAttribute('manage_beforeDelete')(self, item, container) security.declarePrivate('manage_afterAdd') def manage_afterAdd(self, item, container): self.subscribe() BaseTool.inheritedAttribute('manage_afterAdd')(self, item, container) security.declareProtected(CMFCorePermissions.ManagePortal, 'getServerAddress') def getServerAddress(self): """ Backward-compatibility code only. """ warnings.warn( '"getServerAddress" class method is deprecated, use "getServerAddress" module-level function instead.', DeprecationWarning, stacklevel=2, ) return getServerAddress() security.declareProtected(CMFCorePermissions.ManagePortal, 'getCurrentNode') def getCurrentNode(self): """ Backward-compatibility code only. """ warnings.warn( '"getCurrentNode" class method is deprecated, use "getCurrentNode" module-level function instead.', DeprecationWarning, stacklevel=2, ) return getCurrentNode() security.declareProtected(CMFCorePermissions.ManagePortal, 'getDistributingNode') def getDistributingNode(self): """ Return the distributingNode """ return self.distributingNode def getNodeList(self, role=None): node_dict = self.getNodeDict() if role is None: result = [x for x in node_dict.keys()] else: result = [node_id for node_id, node_role in node_dict.items() if node_role == role] result.sort() return result def getNodeDict(self): nodes = self._nodes if isinstance(nodes, tuple): new_nodes = OIBTree() new_nodes.update([(x, ROLE_PROCESSING) for x in nodes]) self._nodes = nodes = new_nodes return nodes def registerNode(self, node): node_dict = self.getNodeDict() if node not in node_dict: if node_dict: # BBB: check if our node was known by address (processing and/or # distribution), and migrate it. server_address = getServerAddress() role = node_dict.pop(server_address, ROLE_IDLE) if self.distributingNode == server_address: self.distributingNode = node else: # We are registering the first node, make # it both the distributing node and a processing node. role = ROLE_PROCESSING self.distributingNode = node self.updateNode(node, role) def updateNode(self, node, role): node_dict = self.getNodeDict() node_dict[node] = role security.declareProtected(CMFCorePermissions.ManagePortal, 'getProcessingNodeList') def getProcessingNodeList(self): return self.getNodeList(role=ROLE_PROCESSING) security.declareProtected(CMFCorePermissions.ManagePortal, 'getIdleNodeList') def getIdleNodeList(self): return self.getNodeList(role=ROLE_IDLE) def _isValidNodeName(self, node_name) : """Check we have been provided a good node name""" return isinstance(node_name, str) security.declareProtected(CMFCorePermissions.ManagePortal, 'manage_setDistributingNode') def manage_setDistributingNode(self, distributingNode, REQUEST=None): """ set the distributing node """ if not distributingNode or self._isValidNodeName(distributingNode): self.distributingNode = distributingNode if REQUEST is not None: REQUEST.RESPONSE.redirect( REQUEST.URL1 + '/manageLoadBalancing?manage_tabs_message=' + urllib.quote("Distributing Node successfully changed.")) else : if REQUEST is not None: REQUEST.RESPONSE.redirect( REQUEST.URL1 + '/manageLoadBalancing?manage_tabs_message=' + urllib.quote("Malformed Distributing Node.")) security.declareProtected(CMFCorePermissions.ManagePortal, 'manage_delNode') def manage_delNode(self, unused_node_list=None, REQUEST=None): """ delete selected unused nodes """ processing_node = self.getDistributingNode() updated_processing_node = False if unused_node_list is not None: node_dict = self.getNodeDict() for node in unused_node_list: if node in node_dict: del node_dict[node] if node == processing_node: self.processing_node = '' updated_processing_node = True if REQUEST is not None: if unused_node_list is None: message = "No unused node selected, nothing deleted." else: message = "Deleted nodes %r." % (unused_node_list, ) if updated_processing_node: message += "Disabled distributing node because it was deleted." REQUEST.RESPONSE.redirect( REQUEST.URL1 + '/manageLoadBalancing?manage_tabs_message=' + urllib.quote(message)) security.declareProtected(CMFCorePermissions.ManagePortal, 'manage_addToProcessingList') def manage_addToProcessingList(self, unused_node_list=None, REQUEST=None): """ Change one or more idle nodes into processing nodes """ if unused_node_list is not None: for node in unused_node_list: self.updateNode(node, ROLE_PROCESSING) if REQUEST is not None: if unused_node_list is None: message = "No unused node selected, nothing done." else: message = "Nodes now procesing: %r." % (unused_node_list, ) REQUEST.RESPONSE.redirect( REQUEST.URL1 + '/manageLoadBalancing?manage_tabs_message=' + urllib.quote(message)) security.declareProtected(CMFCorePermissions.ManagePortal, 'manage_removeFromProcessingList') def manage_removeFromProcessingList(self, processing_node_list=None, REQUEST=None): """ Change one or more procesing nodes into idle nodes """ if processing_node_list is not None: for node in processing_node_list: self.updateNode(node, ROLE_IDLE) if REQUEST is not None: if processing_node_list is None: message = "No used node selected, nothing done." else: message = "Nodes now unused %r." % (processing_node_list, ) REQUEST.RESPONSE.redirect( REQUEST.URL1 + '/manageLoadBalancing?manage_tabs_message=' + urllib.quote(message)) security.declarePrivate('process_shutdown') def process_shutdown(self, phase, time_in_phase): """ Prevent shutdown from happening while an activity queue is processing a batch. """ global has_processed_shutdown if phase == 3 and not has_processed_shutdown: has_processed_shutdown = True LOG('CMFActivity', INFO, "Shutdown: Waiting for activities to finish.") is_running_lock.acquire() LOG('CMFActivity', INFO, "Shutdown: Activities finished.") security.declareProtected(CMFCorePermissions.ManagePortal, 'process_timer') def process_timer(self, tick, interval, prev="", next=""): """ Call distribute() if we are the Distributing Node and call tic() with our node number. This method is called by TimerService in the interval given in zope.conf. The Default is every 5 seconds. """ # Prevent TimerService from starting multiple threads in parallel if timerservice_lock.acquire(0): try: # make sure our skin is set-up. On CMF 1.5 it's setup by acquisition, # but on 2.2 it's by traversal, and our site probably wasn't traversed # by the timerserver request, which goes into the Zope Control_Panel # calling it a second time is a harmless and cheap no-op. # both setupCurrentSkin and REQUEST are acquired from containers. self.setupCurrentSkin(self.REQUEST) old_sm = getSecurityManager() try: # get owner of portal_catalog, so normally we should be able to # have the permission to invoke all activities user = self.portal_catalog.getWrappedOwner() newSecurityManager(self.REQUEST, user) currentNode = getCurrentNode() self.registerNode(currentNode) processing_node_list = self.getNodeList(role=ROLE_PROCESSING) # only distribute when we are the distributingNode if self.getDistributingNode() == currentNode: self.distribute(len(processing_node_list)) # SkinsTool uses a REQUEST cache to store skin objects, as # with TimerService we have the same REQUEST over multiple # portals, we clear this cache to make sure the cache doesn't # contains skins from another portal. try: self.getPortalObject().portal_skins.changeSkin(None) except AttributeError: pass # call tic for the current processing_node # the processing_node numbers are the indices of the elements # in the node tuple +1 because processing_node starts form 1 if currentNode in processing_node_list: self.tic(processing_node_list.index(currentNode) + 1) except: # Catch ALL exception to avoid killing timerserver. LOG('ActivityTool', ERROR, 'process_timer received an exception', error=sys.exc_info()) finally: setSecurityManager(old_sm) finally: timerservice_lock.release() security.declarePublic('distribute') def distribute(self, node_count=1): """ Distribute load """ # Call distribute on each queue for activity in activity_dict.itervalues(): activity.distribute(aq_inner(self), node_count) security.declarePublic('tic') def tic(self, processing_node=1, force=0): """ Starts again an activity processing_node starts from 1 (there is not node 0) """ global active_threads # return if the number of threads is too high # else, increase the number of active_threads and continue tic_lock.acquire() too_many_threads = (active_threads >= max_active_threads) if not too_many_threads or force: active_threads += 1 else: tic_lock.release() raise RuntimeError, 'Too many threads' tic_lock.release() inner_self = aq_inner(self) try: # Loop as long as there are activities. Always process the queue with # "highest" priority. If several queues have same highest priority, # use a round-robin algorithm. # XXX: We always finish by iterating over all queues, in case that # getPriority does not see messages dequeueMessage would process. activity_list = activity_dict.values() def sort_key(activity): return activity.getPriority(self) while is_running_lock.acquire(0): try: activity_list.sort(key=sort_key) # stable sort for i, activity in enumerate(activity_list): # Transaction processing is the responsability of the activity if not activity.dequeueMessage(inner_self, processing_node): activity_list.append(activity_list.pop(i)) break else: break finally: is_running_lock.release() finally: # decrease the number of active_threads tic_lock.acquire() active_threads -= 1 tic_lock.release() def hasActivity(self, *args, **kw): # Check in each queue if the object has deferred tasks # if not argument is provided, then check on self if args: obj, = args else: obj = self path = None if obj is None else '/'.join(obj.getPhysicalPath()) db = self.getSQLConnection() quote = db.string_literal return bool(db.query("(%s)" % ") UNION ALL (".join( activity.hasActivitySQL(quote, path=path, **kw) for activity in activity_dict.itervalues()))[1]) security.declarePrivate('getActivityBuffer') def getActivityBuffer(self, create_if_not_found=True): """ Get activtity buffer for this thread for this activity tool. If no activity buffer is found at lowest level and create_if_not_found is True, create one. Intermediate level is unconditionaly created if non existant because chances are it will be used in the instance life. """ # XXX: using a volatile attribute to cache getPhysicalPath result. # This cache may need invalidation if all the following is # simultaneously true: # - ActivityTool instances can be moved in object tree # - moved instance is used to get access to its activity buffer # - another instance is put in the place of the original, and used to # access its activity buffer # ...which seems currently unlikely, and as such is left out. try: my_instance_key = self._v_physical_path except AttributeError: # Safeguard: make sure we are wrapped in acquisition context before # using our path as an activity tool instance-wide identifier. assert getattr(self, 'aq_self', None) is not None self._v_physical_path = my_instance_key = self.getPhysicalPath() thread_activity_buffer = global_activity_buffer[my_instance_key] my_thread_key = get_ident() try: return thread_activity_buffer[my_thread_key] except KeyError: if create_if_not_found: buffer = ActivityBuffer() else: buffer = None thread_activity_buffer[my_thread_key] = buffer return buffer def activateObject(self, object, activity=DEFAULT_ACTIVITY, active_process=None, **kw): if active_process is None: active_process_uid = None elif isinstance(active_process, str): # TODO: deprecate active_process_uid = self.unrestrictedTraverse(active_process).getUid() else: active_process_uid = active_process.getUid() active_process = active_process.getPhysicalPath() if isinstance(object, str): oid = None url = tuple(object.split('/')) else: try: oid = aq_base(object)._p_oid # Note that it's too early to get the OID of a newly created object, # so at this point, self.oid may still be None. except AttributeError: pass url = object.getPhysicalPath() if kw.get('serialization_tag', False) is None: del kw['serialization_tag'] return ActiveWrapper(self, url, oid, activity, active_process, active_process_uid, kw, getattr(self, 'REQUEST', None)) def getRegisteredMessageList(self, activity): activity_buffer = self.getActivityBuffer(create_if_not_found=False) if activity_buffer is not None: #activity_buffer._register() # This is required if flush flush is called outside activate return activity.getRegisteredMessageList(activity_buffer, aq_inner(self)) else: return [] def unregisterMessage(self, activity, message): activity_buffer = self.getActivityBuffer() #activity_buffer._register() return activity.unregisterMessage(activity_buffer, aq_inner(self), message) def flush(self, obj, invoke=0, **kw): self.getActivityBuffer() if isinstance(obj, tuple): object_path = obj else: object_path = obj.getPhysicalPath() for activity in activity_dict.itervalues(): activity.flush(aq_inner(self), object_path, invoke=invoke, **kw) def invoke(self, message): if self.activity_tracking: activity_tracking_logger.info('invoking message: object_path=%s, method_id=%s, args=%r, kw=%r, activity_kw=%r, user_name=%s' % ('/'.join(message.object_path), message.method_id, message.args, message.kw, message.activity_kw, message.user_name)) old_localizer_context = False if getattr(self, 'aq_chain', None) is not None: # Grab existing acquisition chain and extrach base objects. base_chain = [aq_base(x) for x in self.aq_chain] # Grab existig request (last chain item) and create a copy. request_container = base_chain.pop() request = request_container.REQUEST # Generate PARENTS value. Sadly, we cannot reuse base_chain since # PARENTS items must be wrapped in acquisition parents = [] application = self.getPhysicalRoot().aq_base for parent in self.aq_chain: if parent.aq_base is application: break parents.append(parent) # XXX: REQUEST.clone() requires PARENTS to be set, and it's not when # runing unit tests. Recreate it if it does not exist. if getattr(request.other, 'PARENTS', None) is None: request.other['PARENTS'] = parents # XXX: PATH_INFO might not be set when runing unit tests. if request.environ.get('PATH_INFO') is None: request.environ['PATH_INFO'] = '/Control_Panel/timer_service/process_timer' # restore request information new_request = request.clone() request_info = message.request_info # PARENTS is truncated by clone new_request.other['PARENTS'] = parents if '_script' in request_info: new_request._script = request_info['_script'] if 'SERVER_URL' in request_info: new_request.other['SERVER_URL'] = request_info['SERVER_URL'] if 'VirtualRootPhysicalPath' in request_info: new_request.other['VirtualRootPhysicalPath'] = request_info['VirtualRootPhysicalPath'] if 'HTTP_ACCEPT_LANGUAGE' in request_info: new_request.environ['HTTP_ACCEPT_LANGUAGE'] = request_info['HTTP_ACCEPT_LANGUAGE'] # Replace Localizer/iHotfix Context, saving existing one localizer_context = LocalizerContext(new_request) id = get_ident() localizer_lock.acquire() try: old_localizer_context = localizer_contexts.get(id) localizer_contexts[id] = localizer_context finally: localizer_lock.release() # Execute Localizer/iHotfix "patch 2" new_request.processInputs() new_request_container = request_container.__class__(REQUEST=new_request) # Recreate acquisition chain. my_self = new_request_container base_chain.reverse() for item in base_chain: my_self = item.__of__(my_self) else: my_self = self LOG('CMFActivity.ActivityTool.invoke', INFO, 'Strange: invoke is called outside of acquisition context.') try: message(my_self) finally: if my_self is not self: # We rewrapped self # Restore default skin selection skinnable = self.getPortalObject() skinnable.changeSkin(skinnable.getSkinNameFromRequest(request)) if old_localizer_context is not False: # Restore Localizer/iHotfix context id = get_ident() localizer_lock.acquire() try: if old_localizer_context is None: del localizer_contexts[id] else: localizer_contexts[id] = old_localizer_context finally: localizer_lock.release() if self.activity_tracking: activity_tracking_logger.info('invoked message') if my_self is not self: # We rewrapped self for held in my_self.REQUEST._held: self.REQUEST._hold(held) def invokeGroup(self, method_id, message_list, activity, merge_duplicate): if self.activity_tracking: activity_tracking_logger.info( 'invoking group messages: method_id=%s, paths=%s' % (method_id, ['/'.join(m.object_path) for m in message_list])) # Invoke a group method. message_dict = {} path_set = set() # Filter the list of messages. If an object is not available, mark its # message as non-executable. In addition, expand an object if necessary, # and make sure that no duplication happens. for m in message_list: # alternate method is used to segregate objects which cannot be grouped. alternate_method_id = m.activity_kw.get('alternate_method_id') try: object_list = m.getObjectList(self) if object_list is None: continue message_dict[m] = expanded_object_list = [] for subobj in object_list: if merge_duplicate: path = subobj.getPath() if path in path_set: continue path_set.add(path) if alternate_method_id is not None \ and hasattr(aq_base(subobj), alternate_method_id): # if this object is alternated, # generate a new single active object activity_kw = m.activity_kw.copy() activity_kw.pop('group_method_id', None) activity_kw.pop('group_id', None) active_obj = subobj.activate(activity=activity, **activity_kw) getattr(active_obj, alternate_method_id)(*m.args, **m.kw) else: expanded_object_list.append(GroupedMessage(subobj, m)) except: m.setExecutionState(MESSAGE_NOT_EXECUTED, context=self) expanded_object_list = sum(message_dict.itervalues(), []) try: if expanded_object_list: # Store site info setSite(self.getParentValue()) traverse = self.getPortalObject().unrestrictedTraverse # FIXME: how to apply security here? # NOTE: The callee must update each processed item of # expanded_object_list, by setting: # - 'exc_info' in case of error # - 'result' otherwise, with None or the result to post # on the active process # Skipped item must not be touched. traverse(method_id)(expanded_object_list) except: # In this case, the group method completely failed. exc_info = sys.exc_info() for m in message_dict: m.setExecutionState(MESSAGE_NOT_EXECUTED, exc_info, log=False) LOG('WARNING ActivityTool', 0, 'Could not call method %s on objects %s' % (method_id, [x.object for x in expanded_object_list]), error=exc_info) error_log = getattr(self, 'error_log', None) if error_log is not None: error_log.raising(exc_info) else: # Note there can be partial failures. for m, expanded_object_list in message_dict.iteritems(): result_list = [] for result in expanded_object_list: try: if result.result is not None: result_list.append(result) except AttributeError: exc_info = getattr(result, "exc_info", (SkippedMessage,)) break # failed or skipped message else: try: if result_list and m.active_process: active_process = traverse(m.active_process) for result in result_list: m.activateResult(active_process, result.result, result.object) except: exc_info = None else: m.setExecutionState(MESSAGE_EXECUTED, context=self) continue m.setExecutionState(MESSAGE_NOT_EXECUTED, exc_info, context=self) if self.activity_tracking: activity_tracking_logger.info('invoked group messages') security.declarePrivate('dummyGroupMethod') class dummyGroupMethod(object): def __bobo_traverse__(self, REQUEST, method_id): def group_method(message_list): user_name = None sm = getSecurityManager() try: for m in message_list: message = m._message if user_name != message.user_name: user_name = message.user_name message.changeUser(user_name, m.object) m.result = getattr(m.object, method_id)(*m.args, **m.kw) except Exception: m.raised() finally: setSecurityManager(sm) return group_method dummyGroupMethod = dummyGroupMethod() def newMessage(self, activity, path, active_process, activity_kw, method_id, *args, **kw): # Some Security Cheking should be made here XXX self.getActivityBuffer() activity_dict[activity].queueMessage(aq_inner(self), Message(path, active_process, activity_kw, method_id, args, kw, portal_activities=self)) security.declareProtected( CMFCorePermissions.ManagePortal, 'manageInvoke' ) def manageInvoke(self, object_path, method_id, REQUEST=None): """ Invokes all methods for object "object_path" """ if type(object_path) is type(''): object_path = tuple(object_path.split('/')) self.flush(object_path,method_id=method_id,invoke=1) if REQUEST is not None: return REQUEST.RESPONSE.redirect('%s/%s' % (self.absolute_url(), 'manageActivities')) security.declareProtected( CMFCorePermissions.ManagePortal, 'manageRestart') def manageRestart(self, message_uid_list, activity, REQUEST=None): """ Restart one or several messages """ if not(isinstance(message_uid_list, list)): message_uid_list = [message_uid_list] if message_uid_list: activity_dict[activity].unreserveMessageList(self.getSQLConnection(), 0, message_uid_list) if REQUEST is not None: return REQUEST.RESPONSE.redirect('%s/%s' % ( self.absolute_url(), 'view')) security.declareProtected( CMFCorePermissions.ManagePortal, 'manageCancel' ) def manageCancel(self, object_path, method_id, REQUEST=None): """ Cancel all methods for object "object_path" """ LOG('ActivityTool', WARNING, '"manageCancel" method is deprecated, use "manageDelete" instead.') if type(object_path) is type(''): object_path = tuple(object_path.split('/')) self.flush(object_path,method_id=method_id,invoke=0) if REQUEST is not None: return REQUEST.RESPONSE.redirect('%s/%s' % ( self.absolute_url(), 'manageActivities')) security.declareProtected( CMFCorePermissions.ManagePortal, 'manageDelete' ) def manageDelete(self, message_uid_list, activity, REQUEST=None): """ Delete one or several messages """ if not(isinstance(message_uid_list, list)): message_uid_list = [message_uid_list] activity_dict[activity].deleteMessageList( self.getSQLConnection(), message_uid_list) if REQUEST is not None: return REQUEST.RESPONSE.redirect('%s/%s' % ( self.absolute_url(), 'view')) security.declareProtected( CMFCorePermissions.ManagePortal, 'manageClearActivities' ) def manageClearActivities(self, keep=1, RESPONSE=None): """ Recreate tables, clearing all activities """ for activity in activity_dict.itervalues(): activity.initialize(self, clear=True) if RESPONSE is not None: return RESPONSE.redirect(self.absolute_url_path() + '/manageActivitiesAdvanced?manage_tabs_message=Activities%20Cleared') security.declarePublic('getMessageTempObjectList') def getMessageTempObjectList(self, **kw): """ Get object list of messages waiting in queues """ message_list = self.getMessageList(**kw) object_list = [] for sql_message in message_list: message = self.newContent(temp_object=1) message.__dict__.update(**sql_message.__dict__) object_list.append(message) return object_list security.declarePublic('getMessageList') def getMessageList(self, activity=None, **kw): """ List messages waiting in queues """ if activity: return activity_dict[activity].getMessageList(aq_inner(self), **kw) message_list = [] for activity in activity_dict.itervalues(): try: message_list += activity.getMessageList(aq_inner(self), **kw) except AttributeError: LOG('getMessageList, could not get message from Activity:',0,activity) return message_list security.declarePublic('countMessageWithTag') def countMessageWithTag(self, value): """ Return the number of messages which match the given tag. """ return self.countMessage(tag=value) security.declarePublic('countMessage') def countMessage(self, **kw): """ Return the number of messages which match the given parameter. Parameters allowed: method_id : the id of the method path : for activities on a particular object tag : activities with a particular tag message_uid : activities with a particular uid """ db = self.getSQLConnection() quote = db.string_literal return sum(x for x, in db.query("(%s)" % ") UNION ALL (".join( activity.countMessageSQL(quote, **kw) for activity in activity_dict.itervalues()))[1]) security.declareProtected( CMFCorePermissions.ManagePortal , 'newActiveProcess' ) def newActiveProcess(self, REQUEST=None, **kw): # note: if one wants to create an Actice Process without ERP5 products, # she can call ActiveProcess.addActiveProcess obj = self.newContent(portal_type="Active Process", **kw) if REQUEST is not None: REQUEST['RESPONSE'].redirect( 'manage_main' ) return obj security.declarePrivate('getDependentMessageList') def getDependentMessageList(self, message): message_list = [] for validator_id, validation_value in message.activity_kw.iteritems(): method_id = "_validate_" + validator_id for activity in activity_dict.itervalues(): method = getattr(activity, method_id, None) if method is not None: result = method(self, message, validation_value) if result: message_list += [(activity, m) for m in result] return message_list # Required for tests (time shift) def timeShift(self, delay): for activity in activity_dict.itervalues(): activity.timeShift(aq_inner(self), delay) InitializeClass(ActivityTool)