I'm using the DI in the following way:
_actorSystem.ActorOf(
Supervise(PropsWithDI<MessagePublisherActor>()), MessagePublisher.ActorName);
private Props PropsWithDI<T>(params object[] args) where T : ActorBase
=> DependencyResolver.For(_actorSystem).Props<T>(args);
If somehow there is a Constructor parameter which isn't added to the DI container, it simply doesn't create the Actor, which is normal of course but I would expect an error or something.
Am I using it wrong or is this a missing functionality of Akka.NET Akka.DependencyInjection.
After reviewing the code I've found that this is actually a design choice.
The ActorSystem is designed with fault isolation in mind, it is designed so that any fault happening on the individual actor level would not propagate up to the application layer and crash the application, just like in any operating system, you don't want a fault in a single process to bring the whole operating system down.
To achieve this, actors are started asynchronously inside their own bubble to create a high degree of fault isolation. It is actually done in 2 steps:
An empty actor holder is returned by the ActorOf() method, this holder contains the MailBox and ActorCell. ActorCell contains the Props needed by the system to re/create the actor as needed.
A Create system message is enqueued into the Mailbox, triggering the actual Actor instantiation.
This is essentially a "bubble" where the actor code lives, any exception thrown by any code at step 2 onward are blocked from propagating to the application layer and instead propagated up the chain of actors above it, letting them decide on what would happen to the faulting actors. Since the DI exception happened on step 2 of the process, the exception would never reach the code that actually calls the ActorOf() method and instead be propagated up to the guardian actor and be logged as an error in the log.
Looks like indeed you can at least catch the error on runtime which helps a lot.
I ended up with something similar:
.WithSupervisorStrategy(new OneForOneStrategy(receivedException =>
{
_logger.LogError("Eror: {exception}", receivedException);
return Directive.Escalate;
}));
This article has more information about how to set up the supervisior:
https://getakka.net/articles/concepts/supervision.html
var childProps = Props.Create<EchoActor>();
var supervisor = BackoffSupervisor.Props(
Backoff.OnStop(
childProps,
childName: "myEcho",
minBackoff: TimeSpan.FromSeconds(3),
maxBackoff: TimeSpan.FromSeconds(30),
randomFactor: 0.2)
.WithAutoReset(TimeSpan.FromSeconds(10))
.WithSupervisorStrategy(new OneForOneStrategy(exception =>
{
if (exception is MyException)
return Directive.Restart;
return Directive.Escalate;
})));
system.ActorOf(supervisor, "echoSupervisor");
Related
I have started to implement a set of classes that provide a direct interface to MongoDB for persistence, similar in spirit to the now-unmaintained SDB persistor implementation for RDBMS.
I am using the time-honored technique of creating the necessary concrete classes from the interfaces and doing a println in each method, therein allowing me to trace the execution. I have gotten all the way to where the engine is calling out to my cursor set up:
public ExtendedIterator<Triple> find(Node s, Node p, Node o) {
System.out.println("+++ MongoGraph:extenditer:find(" + s + p + o + ")");
// TBD Need to turn s,p,o into a match expression! Easy!
MongoCursor cur = this.coll.find().iterator();
ExtendedIterator<Triple> curs = new JenaMongoCursorIterator(cur);
return curs;
}
Sadly, when I later call this:
while(rs.hasNext()) {
QuerySolution soln = rs.nextSolution() ;
System.out.println(soln);
}
It turns out rs.hasNext() is always false even though material is present in the MongoCursor (I can debug-print it in the find() method). Also, the trace print in the next() function in my concrete iterator JenaMongoCursorIterator (which extends NiceIterator which I believe is OK) is never hit. In short, the basic setup seems good but then the engine never cranks the iterator on find()
Trying to use SDB as a guide is completely overwhelming for someone not intimately familiar with the software architecture. It's fully factored and filled with interfaces and factories and although that is excellent, it is difficult to nav.
Has anyone tried to create their own persistor implementation and if so, what are the basic steps to getting a "hello world" running? Hello World in this case is ANY implementation, non-optimized, that can call next() on something to produce a Triple.
TLDR: It is now working.
I was coding too eagerly and JenaMongoCursorIterator contained a method hasNexl which of course did not override hasNext (with a t ) in the default implementation of NiceIterator which returns false.
This is the sort of problem that eclipse and visual debugging and tracing makes a lot easier to resolve than regular jdb. jdb is fine if you know the software architecture pretty well but if you don't, the multiple open source files and being able to mouse over vars and such provides a tremendous boost in the amount of context that can be created to home in on the problem.
I am new to Android development with Xamarin.Android and I would like to understand how to have the next issue fixed.
Sometimes after restoring my Android application from background I was facing the next error:
Unable to find the default constructor on type MainMenuFragment. The MainMenuFragment is used by the application NavigationDrawerActivity to allow users to switch between different Fragments inside the app.
In order to solve it, I added a default constructor to the MainMenuFragment as described inside the next links:
Xamarin Limitations - 2.1. Missing constructors
Added a default constructor, should fix the issue.
public class MainMenuFragment : DialogFragment
{
readonly NavigationDrawerActivity navigationDrawer;
#region Constructors
public MainMenuFragment () {} // Default constructor...
public MainMenuFragment (NavigationDrawerActivity navigationDrawer, IMenuType launchMenu = null)
{
if (navigationDrawer == null)
throw new ArgumentNullException ("navigationDrawer");
this.navigationDrawer = navigationDrawer;
...
Fragment UpdateTopFragmentForCurrentMenu (Fragment newMenuRootFragment = null)
{
Fragment currentMenuRootFragment = navigationDrawer.CurrentFragment; // issued line.
But now sometime in the future, the MainMenuFragment gets initialized using its default constructor and at the first time it tries to access its navigationDrawer it throws a System.NullReferenceException:
System.NullReferenceException: Object reference not set to an instance of an object
at MainMenuFragment.UpdateTopFragmentForCurrentMenu (Android.App.Fragment) <0x00018>
at MainMenuFragment.OpenMenu (IMenuType,bool) <0x0006b>
at MainMenuFragment.OnCreate (Android.OS.Bundle) <0x00053>
at Android.App.Fragment.n_OnCreate_Landroid_os_Bundle_ (intptr,intptr,intptr) <0x0005b>
at (wrapper dynamic-method) object.3919a6ec-60c1-49fd-b101-86191363dc45 (intptr,intptr,intptr) <0x00043>
How can I have a default constructor implemented without facing this null reference exception?
You're programming like a C# developer, thats what the problem is :) I faced these same hurdles learning monodroid.
Take a look at the examples out there, in java, you'll see almost all the time they initialize using a static method like object.NewInstance() which returns object. This is how they initialize their views/receivers/fragments. At that point they populate the Arguments property and store that in the fragment. You need to remove all your constructors EXCEPT the empty ones and use arguments to pass your data around. If you try to do this using constructors and regular oo concepts you'll be in for a world of hurt. Arguments.putExtra and all those methods are there. It makes things a little verbose but once you get the hang of it you'll start creating some helper methods etc.
Once you get that sorted, you'll need to figure out if you need to recreate your fragments everytime the activity is resumed and if not, mark them as RetainInstance = true as well as get them onto a fragmentmanager which will help you retain all your state.
If you haven't built on android before it's weird and certainly not what I expected. But it's reeaaallly cool, much more awesome than I expected too. And same with Xamarin.
Great similar question: Best practice for instantiating a new Android Fragment
As far as I looked, there's no answered question about QTP's Ext. sdk on stackoverflow (and almost anywhere else on the net; there isn't even a appropriated tag for it...), so I'm aware it's unlikely I get my problem solved by asking , but whatever, it worth trying.
Anyway, before I lose the attention of anyone who never heard or used the Ext. sdk, maybe I will have more luck asking you to help me figure out how to locate the error log file QTP produces at run-time. I know such a file exists in the new UFT 11.5 version, but I couldn't locate it in QTP 10 or 11 (For the record, I don't talk about QTP's Log Tracking feature, but about the "meta" error log of errors/exceptions produced by QTP itself).
Now for the question:
I'm developing an extension for QTP to support native record and run tests on my application.
I'm currently able to import an object repository, and write test steps using The COM object testing agent I developed.
Problem started when I was trying to implement the IRecordable interface; I'm getting the IRecorder object from qtp, and even able to use it as ISuppressor object to exclude redundant steps from being recorded, but all my attempts to record a step (that is, to add new recorded objects to the repository, and add steps to the test) simply failed.
This is the code that I'm using:
public class MyTestingAgent :
AutInterface.ITestable,
AutInterface.IRecordable
{
QTPInterface.IRecorder recorder;
...
public void AutInterface.IRecordable.BeginRecording(object recorder)
{
IRecordSuppressor recordSuppressor = recorder as IRecordSuppressor;
recordSuppressor.Suppress(MyTestingAgentId,
"<Suppress><Item type=\"HWND\" value=\"[#HWND]\" /></Suppress>".Replace("[#HWND]", getMyAppHWND().ToString()));
this.recorder = recorder as QTPInterface.IRecorder;
...
}
public void recordNewObjStep(string parentName, string objName, string method, Object[] arguments)
{
object[] objectHyrarchy = new object[] { findObjectId(objName), findObjectId(parentName) };
string externalParent = null;
string appDescriptionXml = getDescriptionXml(parentName, objName);
try
{
recorder.Record(MyTestingAgentId, objectHyrarchy , appDescriptionXml, externalParent, method, arguments);
Trace.TraceInformation("Record successfully done.");
}
catch (Exception e)
{
Trace.TraceError("TEAAgent.recordSTElement: " + e.ToString());
}
}
...
}
I'm pretty sure all the arguments I send with the call to Record() are accurate. getDescriptionXml() and findObjectId() are used in different cases in the code, and works fine, the method name and argument are correct.
The annoying thing is that the call to Record doesn't even throw exception, and I get "Record successfully done." in the trace log. Needless to say no new object is created in the repository, and no step is added to the test.
As I can't debug QTP, I'm pretty much in the dark with what I'm doing wrong. That's why I'm asking for help with finding QTP's log file, or any other approach that might shed some light on the subject.
For QTP 11 you can turn on the logs by going to QTP's bin directory and running ClientLogs.exe.
Specifically for TEA extensibility do the following.
select the QTP node from the list on the left
find the LogCatPackTEA from the Available Categories list
Click the > button to move it to Selected Categories
Change TEAs level to Debug2 by selecting the category and changing the level
Click OK and run QTP
The logs will show up as QTP.log in the diretory specified in Path:
I'm curious on what the problem you're facing is, please update if you find the cause.
When I try to register my class with autofac I get the following error: "The instance registration 'GetAllDivisionsCommand' can support SingleInstance() sharing only".
I don't understand why I'm getting this error, but assume it's something to do with the class having static member variables used for caching as that's the only thing that's different about this class. I haven't had any trouble registering any other classes as either SingleInstance or InstancePerDependency.
Essentially, the class is used to retrieve a rarely changing list of divisions from the database, and caches the result. Each time the command is run, it first checks for changes on the database and then re-runs the query if changes are detected; if not, it returns the cached list.
So I am trying to register GetAllDivisionsCommand with Autofac as an IGetAllDivisionsCommand. IGetAllDivisionsCommand itself implements IInjectableCommand, which is just a marker interface, and ICachedListCommand. The concrete command class inherits from the abstract base class CachedListCommand which is where the static member variables live.
Does anyone know what would cause this error message? SingleInstance won't work for me as I can't keep reusing the same session.
Code:
Type commandType = typeof(IInjectedCommand);
Type aCommandType = typeof(GetAllDivisions);
var commands =
from t in aCommandType.Assembly.GetExportedTypes()
where t.Namespace == aCommandType.Namespace
&& t.IsClass
&& !t.IsAbstract
&& (commandType.IsAssignableFrom(t))
let iface = t.GetInterfaces().FirstOrDefault(x => "I" + t.Name == x.Name)
select new { Command = t, Interface = iface };
foreach (var cmd in commands)
{
builder.RegisterInstance(cmd.Command).As(cmd.Interface).InstancePerLifetimeScope();
}
RegisterInstace as the name implies is for registering instances not types.
What you need is RegisterType:
foreach (var cmd in commands)
{
builder.RegisterType(cmd.Command).As(cmd.Interface).InstancePerLifetimeScope();
}
And by the way with the Autofac scanning feature your registration code is roughly equivalent:
builder
.RegisterAssemblyTypes(aCommandType.Assembly)
.AssignableTo<IInjectedCommand>()
.InNamespace(aCommandType.Namespace)
.AsImplementedInterfaces()
.InstancePerLifetimeScope();
In my case, I did want RegisterInstance because I actually have an instance in hand that I wanted to register.
I had
builder.RegisterInstance(myInstance).InstancePerDependency();
The documentation for InstancePerDependency reads:
Configure the component so that every dependent component or call to
Resolve() gets a new, unique instance (default.)
On closer inspection, it makes sense that registering an instance with "instance per dependency" is not possible, since it is not possible for Autofac to give back a new instance each time Resolve is called if there is in fact 1 instance registered.
So, in my case, the solution was this.
builder.RegisterInstance(myInstance).SingleInstance();
The Autofac exception could possibly have been worded more clearly to explain this problem.
I'm evaluating Amazon SWF as an option to build a distributed workflow system. The main language will be Java, so the Flow framework is an obvious choice. There's just one thing that keeps puzzling me and I would get some other opinions before I can recommend it as a key component in our architecture:
The examples are all about tasks that produce a result after a deterministic, relatively short period of time (i.e. after some minutes). In our real-life business workflow, the matter looks different, here we have tasks that could take potentially weeks to complete. I checked the calculator already, having workflows that live 30 days or so do not lead to a cost explosion, so it seems they already counted in that possibility.
Did anybody use SWF for some scenario like this and can share any experience? Are there any recommendations, best practices how to design a workflow like this? Is Flow the right choice here?
It seems to me Activity implementations are expected to eventually return a value synchronously, however, for long running transactions we would rather use messages to send back worker results asynchronously.
Any helpful feedback is appreciated.
From the Amazon Simple Workflow Service point of view an activity execution is a pair of API calls: PollForActivityTask and RespondActivityTaskCompleted that share a task token. There is no requirement for those calls coming from the same thread, process or even host.
By default AWS Flow Framework executes an activity synchronously. Use #ManualActivityCompletion annotation to indicate that activity is not complete upon return of the activity method. Such activity should be explicitly completed (or failed) using provided ManualActivityCompletionClient.
Here is an example taken from the AWS Flow Framework Developer Guide:
#ManualActivityCompletion
public String getName() {
ActivityExecutionContext executionContext = contextProvider.getActivityExecutionContext();
String taskToken = executionContext.getTaskToken();
sendEmail("abc#xyz.com",
"Please provide a name for the greeting message and close task with token: " + taskToken);
return "This will not be returned to the caller";
}
public class CompleteActivityTask {
public void completeGetNameActivity(String taskToken) {
AmazonSimpleWorkflow swfClient = new AmazonSimpleWorkflowClient(…); //pass in user credentials
ManualActivityCompletionClientFactory manualCompletionClientFactory = new ManualActivityCompletionClientFactoryImpl(swfClient);
ManualActivityCompletionClient manualCompletionClient
= manualCompletionClientFactory.getClient(taskToken);
String result = "Hello World!";
manualCompletionClient.complete(result);
}
public void failGetNameActivity(String taskToken, Throwable failure) {
AmazonSimpleWorkflow swfClient
= new AmazonSimpleWorkflowClient(…); //pass in user credentials
ManualActivityCompletionClientFactory manualCompletionClientFactory
= new ManualActivityCompletionClientFactoryImpl(swfClient);
ManualActivityCompletionClient manualCompletionClient
= manualCompletionClientFactory.getClient(taskToken);
manualCompletionClient.fail(failure);
}
}
That an activity is implemented using #ManualActivityCompletion is an implementation detail. Workflow code calls it through the same interface and doesn't treat any differently than any activity implemented synchronously.