I am trying to construct a workflow that can be assigned to arbitrary containers and will have the following behavior. When the container makes the transition from state A -> A', the contents of the container should be checked for their state and moved to a new state that depends on their present state. In other words:
obj1: in state 'a' would be transitioned to 'a*'
obj2: in state 'b' would be transitioned to 'b*'
and so on...
obj1 and obj2 are following the same workflow that branches to a number of final states (approved, denied, alternate, etc...)
I know how to do this, if I can trigger my python code as trusted code. Unfortunately, I have not been able to figure out how or where to put the code in my product so this will work. I have found references to using "external methods", however that seems to be going away. Also I want the code to reside in my project.
I think this is probably simple and I am overlooking something. Help with how to put this in my project or another route to achieving the same goal would be welcome.
The scripts support for DCWorkflow only supports through-the-web-addable objects, which limits you to External Methods there.
A better bet is to use workflow events instead. For each workflow transition, two events are fired:
Products.DCWorkflow.interfaces.IAfterTransitionEvent
Products.DCWorkflow.interfaces.IBeforeTransitionEvent
If you subscribe to either of those events, you can then filter on the correct workflow and transistion to react to transitions from trusted code.
Each event fired has the following attributes:
object: the workflowed object
workflow: the current applicable workflow object
old_state: a Products.DCWorkflow.States.StateDefinition object
new_state: another Products.DCWorkflow.States.StateDefinition object
transition: a Products.DCWorkflow.Transititions.TransitionDefinition object
status: a dictionary with the current workflow variables
kwargs: a dictionary of extra arguments passed in to the change-transition call.
Register a subscriber using ZCML:
<subscriber
provides="zope.component.testfiles.adapter.IS"
factory=".youreventsmodule.aftertransition_handler"
for="Products.DCWorkflow.interfaces.IAfterTransitionEvent"
/>
or, because the transition events are object events, listen only to the transition event if it applies to your objects:
<subscriber
provides="zope.component.testfiles.adapter.IS"
factory=".youreventsmodule.container_aftertransition_handler"
for=".interfaces.IMyContainerType
Products.DCWorkflow.interfaces.IAfterTransitionEvent"
/>
which registers your handler only for transition events on objects with the IMyContainerType interface only.
The handler would be:
def aftertransition_handler(event):
# handle all transition events, ever
or, when limiting it to one object interface:
def aftertransition_handler(obj, event):
# handle all transition events for our container interface
Related
How is the PropertyChanged event used?
I'm wanting to evaluated the next state of a state machine whenever any property of the class is changed.
Regards Steve
PropertyChanged is part of the standard IPropertyNotifyChanged signaling.
If you want to create side effects if any change is done - consider setting "HasUserCode" on each attribute, Generate code, now you will find a partial method available for you to fill for Set and Get of attributes.
In a setter you can have side effects like running a trigger to step a state machine.
I Have a variable that updates every time i move my cube in the level blueprint , now i want to access this variable from multiple class blueprints , what do I do , I tried casting to gamestate but didn't succeed , I am really new to ue4 if you could explain in details please
edit: sorry for not adding details ,
The var I want to access is an integer named cube_side that tells me what side the cube is on every time I move , all of this happens in the level bp , I want to access this variable to see what side the cube is on from other class blueprints ->
here are some details in a picture
I know it's not good to code everything in the level blueprint , but it's too late now , I only need to transfer the var cube_side to other class blueprints so the other object can change depending what side the cube is on.
Create an Actor Class for your logic/functionality. Then use Get all actors of class (choose your class) -> Get a copy -> get variable
Communication with the level blueprint is rather tricky in UE4, since they are not as persistent as e.g. the GameMode, and therefore shouldn't be accessed directly (Imagine older games like Final Fantasy where a new level was loaded every time you stepped outside a boundary, so relying on it could potentially break your actors or crash the game due to nullptrs).
It's a little hacky, but works:
Move the variable inside the cube-blueprint. Add an event dispatcher to the cube, if it is moved, call it and pass the variable in.
Select the cube in the editor, open the level blueprint, right-click, "add reference to selected actor" (the cube must be part of a blueprintclass, not only a static mesh dragged in, though), and bind the event dispatcher inside the Level BP.
Create a function inside every blueprint that needs access to the variable, which does whatever it should do, depending on the variable.
The custom event of the Level Bp (that was bound to the Event Dispatcher of the cube), needs references to all actors that have to work, when the variable changes and call each Actors function (you can get the references like you got the one from the cube)
Then, every time the variable changes, the Level BP is notified, the custom Event is executed and this custom event calls all the actor's functions.
Event Dispatchers explained
This is a huge wastage of functions/code, since you only need it for this one level and may never use it again. Avoid this in the future, by not relying on the level BP so much.
You can use GameStateBP to create and store all variables that you need in game, in GameModeBP create functions to get and set this variables via Get Game State function and then function Cast To GameState and then logic. After that from any blueprint access it using Get Game Mode -> Cast to you GameMode -> use your function to set or get data from GameState.
Is there any idea wich allows me using layouts declared in MyApplication.java from other classes and functions.
I tried put them in parameters it works but it becomes very complicated
For example xhen callin a function named Y in function X I have to pass all layouts on parameters like this:
X(layout1,layout2,layout3,layout4)
{
Y(a,b,c,layout1,layout2,layout3,layout4)
}
I tried to use a class named uiHelper but it didn't works
You can take a look at Blackboard addon for vaadin.
https://vaadin.com/addon/blackboard
From that page:
Sometimes, having a deep component hierarchy poses a problem, when you need to inform a component high up in the tree that something happened deep down below. You normally have one of two choices - either pass the listener all the way down the hierarchy, leading to more coupled code, or let each component in between be a listener/notifier, passing the event all the way back up. With the Blackboard, you can register any listener to listen for any event, and when that event is fired, all the listeners for that event are triggered. This keeps your components clean and rid of unnecessary boilerplate code.
For your example, you can create a LayoutChangeListener and LayoutChangeEvent.
MyApplication can then implements LayoutChangeListener and when a LayoutChangeEvent is fired, you can change your layout without passing it around.
I have a repository named MemberRepository which is used by the MemberListViewModel and the MemberEditViewModel. In the repository I have my GetMember(), CreateMember, SaveMember, Rollback(), and DeleteMember().
Throughout my app I have numerous lookup tables that populate comboboxes. Example Status, Trade Code, Agent, User etc. Where do I place the GetUsers, GetStatuses(), GetTradeCodes() methods? In the appropriate repositories in which they are used?
For example, the member view model needs a list of statuses, trade codes, classes, etc. So I would put the gets for these in the MemberRepository?
Bill
Ok, so your MemberListViewModel is for a listing page and your MemberEditViewModel is an editing page.
What I would do, in my model, is expose the following classes:
ListMembersTask
EditMemberTask
You inject these with all the repositories each one would need, and they expose the properties and methods needed by some abstract thing that's performing each task. For instance, ListMembersTask might have a method called CreateMember() that returns a new EditMemberTask initialized with an empty Member object.
Your ViewModel then gets injected with the appropriate Task (so MemberListViewModel gets injected with ListMembersTask, etc.). Your MemberListViewModel would have a RelayCommand that called CreateMember() and took the returned EditMemberTask, injected it into a MemberEditViewModel and passed the new MemberEditViewModel to the presenter.
If you follow that approach, then the Repositories are only responsible for persistence. A task wraps up the state of the business logic during a session, and a view-model just makes a task bindable.
The next step I've been working on is to dispense with the Task-specific view-models, and I now hand a raw Task to the Presenter, and it inspects the object and builds a view-model hierarchy for the task out of elementary view-model elements (like EditTextViewModel, ChooseOneViewModel, DockingLayoutViewModel, etc.).
I'm implementing a Custom Workflow and Activities to be reused in multiple projects and trying to get them to be as easy to use as possible. In this workflow I have a Property whose name is 'UserID' which I'd like to bind to a dependencyproperty in one of my activities. I can currently bind it at design time searching explicitly for the property each time I add one of these activities to the workflow, but I'd like for this activity to be binded automatically.
As far as i know (correct me if I'm wrong), to bind a dependency property at design time I need to specify a string of the form "Activity=NameOfWorkflow, Path=UserID" to the DefaultBindingProperty metadata tag, and I'd like the name of the workflow to be completed in some way. Any way of doing this?
Thanks
I finally managed to achieve this by attaching an ActivityToolboxItem to the Activity, and overriding a method in it that creates the instance shown in the designer. I used an ActivityBind object to bind the dependencyproperty to the workflow's property. To get the instance of the workflow, I just searched for an ancestor to my activity by calling act.Parent until the activity had no parent (and thus was the StateMachineWorkflowActivity itself)