NativeActivity in XAML loaded COMPILED workflow throws Expression Activity type 'CSharpValue1' requires compilation in order to run - workflow-foundation-4

This is a know error when using C# expressions in windows workflow. The article at https://learn.microsoft.com/en-us/dotnet/framework/windows-workflow-foundation/csharp-expressions#CodeWorkflows explains the reason and how to fix it. It all works fine for me in standard workflows, but as soon as I add a custom NativeActivity to the WF, I get that same error again !
Below the code of how I load the XAML workflow and the simple NativeActivity (which is the ONLY activity in the test workflow and inside that activity is a simple assign expression).
Loading and invoking WF via XAML:
`XamlXmlReaderSettings settings = new XamlXmlReaderSettings()
{
LocalAssembly = GetContextAssembly()
};
XamlReader reader = reader = ActivityXamlServices.CreateReader(new XamlXmlReader(fileURL, settings));
ActivityXamlServicesSettings serviceSettings = new ActivityXamlServicesSettings
{
CompileExpressions = true
};
var activity = ActivityXamlServices.Load(reader, serviceSettings);
WorkflowInvoker.Invoke(activity);`
Doing it in code throws same Exception:
Variable<string> foo = new Variable<string>
{
Name = "Foo"
};
Activity activity = new Sequence
{
Variables = { foo },
Activities =
{
new TimeExecuteUntilAborted
{
Activities =
{
new Assign<string>
{
To = new CSharpReference<string>("Foo"),
Value = new CSharpValue<string>("new Random().Next(1, 101).ToString()")
}
}
}
}
};
CompileExpressions(activity);//the method from the article mentioned above
WorkflowInvoker.Invoke(activity);
The Native Activity:
[Designer("System.Activities.Core.Presentation.SequenceDesigner, System.Activities.Core.Presentation")]
public sealed class TimeExecuteUntilAborted : NativeActivity
{
private Sequence innerSequence = new Sequence();
[Browsable(false)]
public Collection<Activity> Activities
{
get
{
return innerSequence.Activities;
}
}
[Browsable(false)]
public Collection<Variable> Variables
{
get
{
return innerSequence.Variables;
}
}
protected override void CacheMetadata(NativeActivityMetadata metadata)
{
metadata.AddImplementationChild(innerSequence);
}
protected override void Execute(NativeActivityContext context)
{
context.ScheduleActivity(innerSequence);
}
}

Your TimeExecutedUntilAborted class seems to be the culprit. I was able to swap in one of my own template NativeActivities instead and your workflow executed fine with the expressions. I'm guessing that your class is causing an issue in the compiler method when it parses your code. I used this doc as an example for my NativeActivity: https://msdn.microsoft.com/en-us/library/system.activities.nativeactivity(v=vs.110).aspx.

Sizzle Finger's answer is no solution but pointed me into the right direction to simply check what is different. It came out that the simple call to the base class method was missing:
protected override void CacheMetadata(NativeActivityMetadata metadata)
{
base.CacheMetadata(metadata); // !! This needs to be added
metadata.AddImplementationChild(innerSequence);
}

Related

Re-running complete class and not just #Test in TestNG

I have been browsing stackoverflow for some days, trying to find how to re-run a whole test class, and not just an #Test step. Many say that this is not supported with TestNG and IRetryAnalyzer, whereas some have posted workarounds, that don't really work.
Has anyone manage to do it?
And just to clarify the reasons for this, in order to avoid answers that say that is not supported in purpose: TestNG is a tool not only for developers. Meaning that is also used from sw testers for e2e testing. E2e tests can have steps that depend each from the previous one. So yes it's valid to re-run whole test class, rather than simple #Test, which is easily can be done via IRetryAnalyzer.
An example of what I want to achieve would be:
public class DemoTest extends TestBase {
#Test(alwaysRun = true, description = "Do this")
public void testStep_1() {
driver.navigate().to("http://www.stackoverflow.com");
Assert.assertEquals(driver.getCurrentUrl().contains("stackoverflow)"));
}
#Test(alwaysRun = true, dependsOnMethods = "testStep_1", description = "Do that")
public void testStep_2() {
driver.press("button");
Assert.assertEquals(true, driver.elementIsVisible("button"));
}
#Test(alwaysRun = true, dependsOnMethods = "testStep_2", description = "Do something else")
public void testStep_3() {
driver.press("button2");
Assert.assertEquals(true, driver.elementIsVisible("button"));
}
}
Let's say that testStep_2 fails, I want to rerun class DemoTest and not just testStep_2
Okay, I know that you probably want some easy property you can specify in your #BeforeClass or something like that, but we might need to wait for that to be implemented. At least I couldn't find it either.
The following is ugly as hell but I think it does the job, at least in a small scale, it is left to see how it behaves in more complex scenarios. Maybe with more time, this can be improved into something better.
Okay, so I created a Test Class similar to yours:
public class RetryTest extends TestConfig {
public class RetryTest extends TestConfig {
Assertion assertion = new Assertion();
#Test( enabled = true,
groups = { "retryTest" },
retryAnalyzer = TestRetry.class,
ignoreMissingDependencies = false)
public void testStep_1() {
}
#Test( enabled = true,
groups = { "retryTest" },
retryAnalyzer = TestRetry.class,
dependsOnMethods = "testStep_1",
ignoreMissingDependencies = false)
public void testStep_2() {
if (fail) assertion.fail("This will fail the first time and not the second.");
}
#Test( enabled = true,
groups = { "retryTest" },
retryAnalyzer = TestRetry.class,
dependsOnMethods = "testStep_2",
ignoreMissingDependencies = false)
public void testStep_3() {
}
#Test( enabled = true)
public void testStep_4() {
assertion.fail("This should leave a failure in the end.");
}
}
I have the Listener in the super class just in the case I'd like to extend this to other classes, but you can as well set the listener in your test class.
#Listeners(TestListener.class)
public class TestConfig {
protected static boolean retrySuccessful = false;
protected static boolean fail = true;
}
Three of the 4 methods above have a RetryAnalyzer. I left the testStep_4 without it to make sure that what I'm doing next doesn't mess with the rest of the execution. Said RetryAnalyzer won't actually retry (note that the method returns false), but it will do the following:
public class TestRetry implements IRetryAnalyzer {
public static TestNG retryTestNG = null;
#Override
public boolean retry(ITestResult result) {
Class[] classes = {CreateBookingTest.class};
TestNG retryTestNG = new TestNG();
retryTestNG.setDefaultTestName("RETRY TEST");
retryTestNG.setTestClasses(classes);
retryTestNG.setGroups("retryTest");
retryTestNG.addListener(new RetryAnnotationTransformer());
retryTestNG.addListener(new TestListenerRetry());
retryTestNG.run();
return false;
}
}
This will create an execution inside of your execution. It won't mess with the report, and as soon as it finishes, it will continue with your main execution. But it will "retry" the methods within that group.
Yes, I know, I know. This means that you are going to execute your test suite forever in an eternal loop. That's why the RetryAnnotationTransformer. In it, we will remove the RetryAnalyzer from the second execution of those tests:
public class RetryAnnotationTransformer extends TestConfig implements IAnnotationTransformer {
#SuppressWarnings("rawtypes")
#Override
public void transform(ITestAnnotation annotation, Class testClass, Constructor testConstructor, Method testMethod) {
fail = false; // This is just for debugging. Will make testStep_2 pass in the second run.
annotation.setRetryAnalyzer(null);
}
}
Now we have the last of our problems. Our original test suite knows nothing about that "retry" execution there. This is where it gets really ugly. We need to tell our Reporter what just happened. And this is the part that I encourage you to improve. I'm lacking the time to do something nicer, but if I can, I will edit it at some point.
First, we need to know if the retryTestNG execution was successful. There's probably a million ways to do this better, but for now this works. I set up a listener just for the retrying execution. You can see it in TestRetry above, and it consists of the following:
public class TestListenerRetry extends TestConfig implements ITestListener {
(...)
#Override
public void onFinish(ITestContext context) {
if (context.getFailedTests().size()==0 && context.getSkippedTests().size()==0) {
successful = true;
}
}
}
Now the Listener of the main suite, the one you saw above in the super class TestConfig will see if it the run happened and if it went well and will update the report:
public class TestListener extends TestConfig implements ITestListener , ISuiteListener {
(...)
#Override
public void onFinish(ISuite suite) {
if (TestRetry.retryTestNG != null) {
for (ITestNGMethod iTestNGMethod : suite.getMethodsByGroups().get("retryTest")) {
Collection<ISuiteResult> iSuiteResultList = suite.getResults().values();
for (ISuiteResult iSuiteResult : iSuiteResultList) {
ITestContext iTestContext = iSuiteResult.getTestContext();
List<ITestResult> unsuccessfulMethods = new ArrayList<ITestResult>();
for (ITestResult iTestResult : iTestContext.getFailedTests().getAllResults()) {
if (iTestResult.getMethod().equals(iTestNGMethod)) {
iTestContext.getFailedTests().removeResult(iTestResult);
unsuccessfulMethods.add(iTestResult);
}
}
for (ITestResult iTestResult : iTestContext.getSkippedTests().getAllResults()) {
if (iTestResult.getMethod().equals(iTestNGMethod)) {
iTestContext.getSkippedTests().removeResult(iTestResult);
unsuccessfulMethods.add(iTestResult);
}
}
for (ITestResult iTestResult : unsuccessfulMethods) {
iTestResult.setStatus(1);
iTestContext.getPassedTests().addResult(iTestResult, iTestResult.getMethod());
}
}
}
}
}
}
The report should show now 3 tests passed (as they were retried) and one that failed because it wasn't part of the other 3 tests:
I know it's not what you are looking for, but I help it serves you until they add the functionality to TestNG.

Haxe: Binding pattern with abstract fields access methods

I'd like to make wrapper to implement simple data binding pattern -- while some data have been modified all registered handlers are got notified. I have started with this (for js target):
class Main {
public static function main() {
var target = new Some();
var binding = new Bindable(target);
binding.one = 5;
// binding.two = 0.12; // intentionally unset field
binding.three = []; // wrong type
binding.four = 'str'; // no such field in wrapped class
trace(binding.one, binding.two, binding.three, binding.four, binding.five);
// outputs: 5, null, [], str, null
trace(target.one, target.two, target.three);
// outputs: 5, null, []
}
}
class Some {
public var one:Int;
public var two:Float;
public var three:Bool;
public function new() {}
}
abstract Bindable<TClass>(TClass) {
public inline function new(source) { this = source; }
#:op(a.b) public function setField<T>(name:String, value:T) {
Reflect.setField(this, name, value);
// TODO notify handlers
return value;
}
#:op(a.b) public function getField<T>(name:String):T {
return cast Reflect.field(this, name);
}
}
So I have some frustrating issues: interface of wrapped object doesn't expose to wrapper, so there's no auto completion or strict type checking, some necessary attributes can be easily omitted or even misspelled.
Is it possible to fix my solution or should I better move to the macros?
I almost suggested here to open an issue regarding this problem. Because some time ago, there was a #:followWithAbstracts meta available for abstracts, which could be (or maybe was?) used to forward fields and call #:op(a.b) at the same time. But that's not really necessary, Haxe is powerful enough already.
abstract Binding<TClass>(TClass) {
public function new(source:TClass) { this = source; }
#:op(a.b) public function setField<T>(name:String, value:T) {
Reflect.setField(this, name, value);
// TODO notify handlers
trace("set: $name -> $value");
return value;
}
#:op(a.b) public function getField<T>(name:String):T {
trace("get: $name");
return cast Reflect.field(this, name);
}
}
#:forward
#:multiType
abstract Bindable<TClass>(TClass) {
public function new(source:TClass);
#:to function to(t:TClass) return new Binding(t);
}
We use here multiType abstract to forward fields, but resolved type is actually regular abstract. In effect, you have completion working and #:op(a.b) called at the same time.
You need #:forward meta on your abstract. However, this will not make auto-completion working unless you remove #:op(A.B) because it shadows forwarded fields.
EDIT: it seems that shadowing happened first time I added #:forward to your abstract, afterwards auto-completion worked just fine.

Prism6 shared service and dependency property

I use Prism6 + Unity container for desktop application developing.
This is a long-read, sorry. So I ask at top: Prism SetProperty() function is not rising property changed event if input value is Unity singleton. And I understand why: because input value and save value have same reference to singleton instance. RaisePropertyChanged() don't help in this situation.
Long-read is statring...
So, I have a dependency property in my UserControl component:
public static readonly DependencyProperty WorksheetDataProperty =
DependencyProperty.Register("WorksheetData", typeof(WorksheetDataModel), typeof(SheetUserControl),
new PropertyMetadata(new WorksheetDataModel(), WorksheetDataPropertyChanged));
public WorksheetDataModel WorksheetData {
get { return (WorksheetDataModel)GetValue(WorksheetDataProperty); }
set { SetValue(WorksheetDataProperty, value); }
}
private void WorksheetDataPropertyChanged(WorksheetDataModel worksheetData) {
if (worksheetData == null)
return;
SheetGrid.Model.ActiveGridView.BeginInit();
this.ClearWorksheetModel();
this.ResizeWorksheetModel();
SheetGrid.Model.ActiveGridView.EndInit();
}
private static void WorksheetDataPropertyChanged(
DependencyObject d, DependencyPropertyChangedEventArgs e) {
((SheetUserControl)d).WorksheetDataPropertyChanged((WorksheetDataModel)e.NewValue);
}
It's important for me to invoke actions from WorksheetDataPropertyChanged() function.
And scheme without shared service (singleton) is working well: this function is called.
But now I want to share data between several modules. How I see it: I have some "parent" module, which load\save data from storage and shared this data with several other modules, which can modificate shared data, but can't save it.
And EventAggregator is not convenient for me: I don't want to create copies of data and then collect it again after modifications.
So I register my "shared service" as singleton:
_container.RegisterInstance(new WorksheetDataModel());
Now I can load data from database in "parent" viewmodel to singleton object created in previous step:
var data = _container.Resolve<WorksheetDataModel>();
data.Header = args.Header;
data.User = args.User;
data.RowHeader = new WorksheetRowHeader(_model.ReadRowHeader(data.Header.WshCode));
data.ColHeader = new WorksheetColHeader(_model.ReadColHeader(data.Header.WshCode));
data.Cells = _model.ReadCells(data.Header.WshCode);
Further, I notify child viewmodels about new data in singleton:
data.OnDataChanged?.Invoke();
And now most important code from child viewmodel.
In delegate handler I "apply" new value:
WorksheetData = _container.Resolve<WorksheetDataModel>();
WorksheetData is:
private WorksheetDataModel _worksheetData;
public WorksheetDataModel WorksheetData {
get { return _worksheetData; }
set { SetProperty(ref _worksheetData, value); }
}
And problem in this line:
set { SetProperty(ref _worksheetData, value); }
It works only once at first call, because _worksheetData is null. But then refernce of _worksheetData (pointer) setted to singleton and in all next call value and _worksheetData are identical for SetProperty() and, as result, it just quit.
I tried next code:
set {
SetProperty(ref _worksheetData, value);
RaisePropertyChanged("WorksheetData")
}
But no effect. WorksheetDataPropertyChanged() callback in UserControl component is not calling.
So, I don't know now how to better share some data between several modules.
Thanks for any advice.
WorksheetData does not change, the contents of the WorksheetDataModel instance change.
So to update your bindings,
either WorksheetDataModel implements INotifyPropertyChanged and/or uses INotifyCollectionChanged-implementing collections
or you let the view model listen to WorksheetDataModel.OnDataChanged and raise its own PropertyChanged to update all bindings to WorksheetData.
Example:
private WorksheetDataModel _worksheetData;
public WorksheetDataModel WorksheetData
{
get { return _worksheetData; }
set
{
if (_worksheetData != null)
_worksheetData.OnDataChanged -= DataChangedHandler;
SetProperty(ref _worksheetData, value);
if (_worksheetData != null)
_worksheetData.OnDataChanged += DataChangedHandler;
}
}
private void DataChangedHandler( object sender, DataChangedEventArgs args )
{
RaisePropertyChanged( nameof( WorksheetData ) );
}

how to update Visual Studio UI when using DynamicItemStart inside a vsix package

I'm implementing a DynamicItemStart button inside a Menu Controller. I'm loading the dynamic items for this button when Visual Studio starts. Everything is loaded correctly so the initialize method is called an I see all the new items in this Dynamic button. After the package is completely loaded I want to add more items to this Dynamic button, but since the package is already loaded the initialize method is not called again and I cannot see the new items in this Dynamic button. I only see the ones that were loaded when VS started.
Is there any way that I can force the update of this Dynamic button so it shows the new items?. I want to be able to update the VS UI after I added more items but outside the Initialize method.
The implementation I did is very similar to the one showed on this msdn example:
http://msdn.microsoft.com/en-us/library/bb166492.aspx
Does anyone know if an Update of the UI can be done by demand?
Any hints are greatly appreciated.
I finally got this working. The main thing is the implementation of a derived class of OleMenuCommand that implements a new constructor with a Predicate. This predicate is used to check if a new command is a match within the DynamicItemStart button.
public class DynamicItemMenuCommand : OleMenuCommand
{
private Predicate<int> matches;
public DynamicItemMenuCommand(CommandID rootId, Predicate<int> matches, EventHandler invokeHandler, EventHandler beforeQueryStatusHandler)
: base(invokeHandler, null, beforeQueryStatusHandler, rootId)
{
if (matches == null)
{
throw new ArgumentNullException("Matches predicate cannot be null.");
}
this.matches = matches;
}
public override bool DynamicItemMatch(int cmdId)
{
if (this.matches(cmdId))
{
this.MatchedCommandId = cmdId;
return true;
}
this.MatchedCommandId = 0;
return false;
}
}
The above class should be used when adding the commands on execution time. Here's the code that creates the commands
public class ListMenu
{
private int _baselistID = (int)PkgCmdIDList.cmdidMRUList;
private List<IVsDataExplorerConnection> _connectionsList;
public ListMenu(ref OleMenuCommandService mcs)
{
InitMRUMenu(ref mcs);
}
internal void InitMRUMenu(ref OleMenuCommandService mcs)
{
if (mcs != null)
{
//_baselistID has the guid value of the DynamicStartItem
CommandID dynamicItemRootId = new CommandID(GuidList.guidIDEToolbarCmdSet, _baselistID);
DynamicItemMenuCommand dynamicMenuCommand = new DynamicItemMenuCommand(dynamicItemRootId, isValidDynamicItem, OnInvokedDynamicItem, OnBeforeQueryStatusDynamicItem);
mcs.AddCommand(dynamicMenuCommand);
}
}
private bool IsValidDynamicItem(int commandId)
{
return ((commandId - _baselistID) < connectionsCount); // here is the place to put the criteria to add a new command to the dynamic button
}
private void OnInvokedDynamicItem(object sender, EventArgs args)
{
DynamicItemMenuCommand invokedCommand = (DynamicItemMenuCommand)sender;
if (null != invokedCommand)
{
.....
}
}
private void OnBeforeQueryStatusDynamicItem(object sender, EventArgs args)
{
DynamicItemMenuCommand matchedCommand = (DynamicItemMenuCommand)sender;
bool isRootItem = (matchedCommand.MatchedCommandId == 0);
matchedCommand.Enabled = true;
matchedCommand.Visible = true;
int indexForDisplay = (isRootItem ? 0 : (matchedCommand.MatchedCommandId - _baselistID));
matchedCommand.Text = "Text for the command";
matchedCommand.MatchedCommandId = 0;
}
}
I had to review a lot of documentation since it was not very clear how the commands can be added on execution time. So I hope this save some time whoever has to implement anything similar.
The missing piece for me was figuring out how to control the addition of new items.
It took me some time to figure out that the matches predicate (the IsValidDynamicItem method in the sample) controls how many items get added - as long as it returns true, the OnBeforeQueryStatusDynamicItem gets invoked and can set the details (Enabled/Visible/Checked/Text etc.) of the match to be added to the menu.

How can I enable logging in NVelocity?

Any idea how to do what the title says? Only thing I found was on the original Velocity site, and I don't think
ve.setProperty( RuntimeConstants.RUNTIME_LOG_LOGSYSTEM_CLASS,
"org.apache.velocity.runtime.log.Log4JLogChute" );
ve.setProperty("runtime.log.logsystem.log4j.logger",
LOGGER_NAME);
will work wonderfully well on .NET. I am using log4net, which should make it quite easy, but the documentation on NVelocity is really a mess.
Implement NVelocity.Runtime.Log.ILogSystem (you could write a simple implementation that bridges to log4net) and set this impl type in the property RuntimeConstants.RUNTIME_LOG_LOGSYSTEM_CLASS
How I got this information:
Get the code.
Search for "log" in the codebase
Discover the classes in NVelocity.Runtime.Log.
Read those classes' source, they're very simple and thoroughly documented.
Update:
Currently, NVelocity does not support logging. The initializeLogger() and Log() methods in RuntimeInstance Class are commented out.
If you need to log, uncomment the two methods, add a private ILogSystem logSystem; property
Here's our on-the-fly implementation:
public class RuntimeInstance : IRuntimeServices
{
private ILogSystem logSystem;
...
...
private void initializeLogger()
{
logSystem = LogManager.CreateLogSystem(this);
}
...
...
private void Log(LogLevel level, Object message)
{
String output = message.ToString();
logSystem.LogVelocityMessage(level, output);
}
...
}
Then, we implemented ILogSystem for log4net
using log4net;
using NVelocity.Runtime;
using NVelocity.Runtime.Log;
namespace Services.Templates
{
public class Log4NetILogSystem : ILogSystem
{
private readonly ILog _log;
public Log4NetILogSystem(ILog log )
{
_log = log;
}
public void Init(IRuntimeServices rs)
{
}
public void LogVelocityMessage(LogLevel level, string message)
{
switch (level)
{
case LogLevel.Debug:
_log.Debug(message);
break;
case LogLevel.Info:
_log.Info(message);
break;
case LogLevel.Warn:
_log.Warn(message);
break;
case LogLevel.Error:
_log.Error(message);
break;
}
}
}
}
Then, when creating the engine:
var engine = new VelocityEngine();
var props = new ExtendedProperties();
props.SetProperty(RuntimeConstants.RUNTIME_LOG_LOGSYSTEM,
new Log4NetILogSystem(LogManager.GetLogger(typeof(NVelocityEngine))));
engine.Init(props);

Resources