i'm implementing a bot for microsoft teams using com.microsoft.bot:bot-integration-core version 4.14
i'm subclassing TeamsActivityHandler and overriding methods like onMessageActivity and onAdaptiveCardInvoke
i also use BotFrameworkAdapter to add some middleware and onTurnError handler.
when i have an action, triggered from a button click for example, and this action raises an exception, then onTurnError is not called, since an error is handled in TeamsActivityHandler (https://github.com/microsoft/botbuilder-java/blob/main/libraries/bot-builder/src/main/java/com/microsoft/bot/builder/teams/TeamsActivityHandler.java#L181)
in case on an error in onMessageActivity, the onTurnError is called fine.
if i subclass from a parent of TeamsActivityHandler, which is ActivityHandler then i assume error handling will work fine.
so my question is:
how is onTurnError is supposed to be used in case of subclassing TeamsActivityHandler ?
Related
I have moved most of the core functionality of my non-document based macOS app to a custom, embedded framework.
The app code has a standard main storyboard with an initial window, and the window has a "window content" relationship/segue into a storyboard reference pointing to a storyboard inside the embedded framework. Therein lies a custom NSViewController subclass and a custom NSView subclass.
I want to group all the input event handling code inside the framework, which means implementing mouseDown(with:) on the custom NSView subclass, and --lo and behold-- it gets called when I click inside the app window. So far, so good.
Next, I implemented keyDown(with:) to similarly handle keyboard input. However, at runtime, it does not get called and instead, I hear the annoying beep (NSBeep).
I tried implementing keyDown(with:) on the view controller instead, but it's all the same.
Finally, I tried implementing the key handler on my NSWindowController subclass instead, and that does work.
So I could get around this by forwarding the event like so:
class WindowController: NSWindowController {
override func keyDown(with event: NSEvent) {
contentViewController?.view.keyDown(with: event)
}
}
, but it is very inelegant. I would prefer to not pollute the app code with input logic.
This doesn't seem to have anything to do with embedding frameworks, however. I put together a minimal project from the 'Cocoa App' template and confirmed that indeed keyDown(with:) only gets called if implemented on the window controller code, but not on the view or view controller side.
How can I get keyDown(with:) to be called on the view or view controller (not the window or window controller) in a storyboard-based app? (so I can move it from the main app to my embedded framework).
Edit: The question has been marked as duplicate. I tried the solutions pointed in answers to the other question (namely, override acceptsFirstResponder to return true). This solves the problem in my minimal demo project, but when I tried it on my full app, it still does not work (I did see that question and did try to override acceptsFirstResponder in my app before posting this question).
I will now try to modify my minimal poeject to see if I can reporduce the issue in the main app.
Edit 2: I have refactored the minimal project to:
Send the view controller to a separate storyboard,
Send the view controller's storyboard and represented classes (custom view, custom view controller) to a separate, embedded framework.
Now the basic setup mirrors that of my app in all that seems to matter, but still can not reproduce the issue in the minimal project. I will investiate further...
Edit 3: I haven't been able to reproduce the issue on the minimal project.
On my app's custom view, I implemented:
public override var acceptsFirstResponder: Bool {
return true
}
public override func performKeyEquivalent(with event: NSEvent) -> Bool {
let retVal = super.performKeyEquivalent(with: event)
return retVal
}
On startup, acceptsFirstResponder is called twice.
When hitting any key, performKeyEquivalent(with:) is called twice, too. Inspectig the intermediate variable retVal above reveals that the super class's implementation always returns false. After returning from this method, NSBeep() is called and keyDown(with:) isn't.
If instead of super.performKeyEquivalent(with:) I force-return true, I can avert the call to NSBeep() (but keyDown(with:) is still not called...)
Edit 4 (Final):
Out of desperation, I cleared the "Custom Class" field of the window controller's Identity Inspector in my app's main storyboard (to the default NSWindowController).
Suddenly, my custom view's keyDown(with:) starts getting called.
I reinstated the custom class to confirm.
It still works.
I clean the build folder and try again.
It still works.
Now I can no longer reproduce the issue even on my main app. I really don't know what to say...
I'm building an External Object for Android.
I have an asynchronous method and I guess I must to use an Event into the External Object to receive it's callback in Genexus.
But, can anyone show me how to trigger this event in GX from the android java class?
This method:
mCoordinator.runAction("yourEvent", null);
And the mCoordinator you can get it at the constructor:
public YourUC(Context context, Coordinator coordinator, LayoutItemDefinition definition){
}
Is there any event function in WinAPI that allows catching window registration moment by it's class name? For exaple trigger event when application starts or create new window.
I've found EVENT_OBJECT_CREATE but it doesn't seem to catch proper event. Any ideas?
Does anyone know a way of finding out precisely when either a Ti.Platform.openUrl or Ti.App.fireEvent finishes? Something like a callback or an completion event? Thanks.
Nathan
App-level events are global to your app. They are accessible in all contexts, functional scopes, CommonJS modules, and so forth. You can fire them and listen for them via the Ti.App module as follows.
//Creating an app level custom event
Ti.App.addEventListener('myAppEvent', myCallBack);
//Fire the event like
Ti.App.fireEvent('myAppEvent');
//Callback function of your custom event
function myCallBack(e){
alert("In call back. App level event get fired!!!");
}
Please refer Event handling in Titanium for more details
Is there any way to set a custom error page in IIS 7 without creating a web.config?
Unfortunately researching this particular topic has been very difficult because there are SO many articles on how to do it with a web.config. What I'm looking for is either buried beneath the 8 million results I don't want or it's not possible.
Yes, there is. It involves either subscribing to the Application_Error event in Global.asax or by writing a custom ErrorHandlerAttribute.
Darin already gave the correct answer, but I want to go into a little more depth.
In any ASP.NET application, given it is Web Forms, MVC, or raw ASP.NET, you can always use Application_Error Global.asax. If your ASP.NET application does not have a Global.asax, all you need to do is right-click your project in Solution Explorer, Add New Item, and choose Global Application Class. You should only have this option available if you don't already have one.
In your Global.asax, if you don't already see it, you can add Application_Error as shown below:
protected void Application_Error(object sender, EventArgs e) {
}
This will be called automatically by ASP.NET whenever there is an error. But as stated here, this is not perfect. Specifically:
An error handler that is defined in the Global.asax file will only
catch errors that occur during processing of requests by the ASP.NET
runtime. For example, it will catch the error if a user requests an
.aspx file that does not occur in your application. However, it does
not catch the error if a user requests a nonexistent .htm file. For
non-ASP.NET errors, you can create a custom handler in Internet
Information Services (IIS). The custom handler will also not be called
for server-level errors.
In Application_Error you can process the uncaught exception with Server.GetLastError(). This will provide you the Exception that was thrown, or null. I am not sure why this handler would be called if an exception didn't occur, but I believe that it is possible.
To redirect the user, use Response.Redirect(). Whatever you pass for the url is going to be sent directly to the browser without any further processing, so you can't use application-relative paths. To do that I would use this method in combination with VirtualPathUtility.ToAbsolute(). For example:
Response.Redirect( VirtualPathUtility.ToAbsolute( "~/Error.aspx" ) );
This redirect will be a 302 (temporary redirect) rather than a 301 (permanent), which is what you want in the case of handling errors. It's worth noting that this overload of Response.Redirect is the same as calling the overload Response.Redirect(url, endResponse: true). This method works by throwing an exception, which is not ideal in terms of performance. Instead, call Response.Redirect(url, false) immediately followed by Response.Complete​Request().
If you're using ASP.NET MVC, [HandleError] is also an option. Place this attribute on your Controller or on an Action within a controller. When this attribute is present, MVC will display the Error view, found in the ~/Views/Shared folder.
But you can make this even easier for yourself. You can automatically add this attribute to call Controllers in your project by creating a FilterConfig class in your project. Example:
public class FilterConfig {
public static void RegisterGlobalFilters(GlobalFilterCollection filters) {
filters.Add(new HandleErrorAttribute());
}
}
And then add FilterConfig.RegisterGlobalFilters( GlobalFilters.Filters ); to your Application_Start() in Global.asax.
You can read more about the HandleErrorAttribute at https://msdn.microsoft.com/en-us/library/system.web.mvc.handleerrorattribute(v=vs.118).aspx.
But as stated above, both of these methods will never cover absolutely all errors that can occur during the processing of your application. It's not possible to provide the best user experience for all possible errors without using Web.config or configuring IIS manually.