The guard is already checked once using isWorkflowMethodSupported, and

notify* methods are never called if the guard evaluates to False. This
check looks costly and useless.


The only possibility (?) would be that one of your Interaction workflow guards
checks kw['workflow_method_args'] and returns a different result depending
on when _checkTransitionGuard is called.
That seems like a tricky (wrong? unlikely?) usecase.
If you are in this case, let me know, and if needed I'll engineer a fast guard
check that is only ever checked once.


git-svn-id: https://svn.erp5.org/repos/public/erp5/trunk@42725 20353a03-c40f-0410-a6d1-a30d3c3de9de

product/ERP5/InteractionWorkflow.py

@@ -220,8 +220,7 @@ class InteractionWorkflowDefinition (DCWorkflowDefinition, ActiveObject):
       tdef = self.interactions[t_id]
       assert tdef.trigger_type == TRIGGER_WORKFLOW_METHOD
       if (tdef.portal_type_filter is None or \
-          ob.getPortalType() in tdef.portal_type_filter) and \
-          self._checkTransitionGuard(tdef, ob, **kw):
+          ob.getPortalType() in tdef.portal_type_filter):
         filtered_transition_list.append(tdef.id)
         former_status = self._getStatusOf(ob)
         # Execute the "before" script.