Using MonoMac, I have a NSDocument-based application, but I'm needing to create a new NSDocument object when a button is clicked.
For example. I have in another Window I have a NSWindowController and I can do
Controller c = new Controller ();
c.Window.MakeKeyAndOrderFront (this);
thus causing the Window to be loaded that is tied to the controller.
With the NSDocument I guess the controller is built in?
So I'm expecting something like
MyNSDocument doc = new MyNSDocument ("Some Value ");
doc.Window.MakeKeyAndOrderFront (this);
Of course this doesn't work.
Additional info, for example when in the Application if you hit Command + N, then a new Document Window is loaded. This is cool and I basically need the same thing to happen, but when a button is clicked.
Using "File" / "New" or Control + N invokes newDocument: on the application's shared document controller, which is the menu.xib's First Responder.
To do the same programmatically, use NSDocumentController.SharedDocumentController to get the application's shared document controller, then invoke NewDocument () on it (you can pass null as sender):
var controller = (NSDocumentController)NSDocumentController.SharedDocumentController;
controller.NewDocument (null);
Related
Like in a lot of apps, I have a list of items (populated by a Core Data fetch request), a sheet to create new items, and a sheet to edit an item when tapping on a row in my list. I'm trying to unify both forms to create and edit an update, and put the cancel / save logic in a superview of the form.
So I've something like this:
ListView: a list with row populated by a Core Data fetch request
AddView: a NavigationView with the FormView embed + cancel and save button
EditView: a NavigationView with the FormView embed + cancel and save button
FormView: a TextField to update the name of the item
In the init() for the AddView, I create a new NSManagedObject without any context (I do that because I don't want my ListView to be updated when I create a new item in the AddView, but only when I save this item -> alternative could be to use a child context, or filter the fetch request results based on the isInserted or objectID.isTemporaryID of the return objects). AddView contains a NavigationView with the FormView embed, a cancel button, and a save button. This save button is disabled based on a computed property on the managed object (name for the object can't be nil).
In the EditView, I pass the item that was tapped from the ListView. This item is an existing NSManagedObject attached to the main viewContext of the app (coming from the fetch request of the ListView). EditView contains a NavigationView with the FormView embed, a cancel button and a save button (exactly like the AddView). This save button is also disabled based on the same computed property.
My issue is that when I update the name of the item from the TextField in my FormView, the condition to enable / disable the save button is not working for the AddView (this AddView is actually not refreshed when I change the item name from the FormView) but working for the EditView (this EditView is refreshed when I change the item name from the FormView). If I attach a context to the new NSManagedObject in the init() of the AddView, the condition is working like in the EditView.
So it appears that a NSManagedObject without any context is not observed by SwiftUI? Am I missing anything or is that a bug?
I wouldn't be surprised (but haven't verified) if the change-notifying ability of a managed object depends on the presence of the context. I can't think of a situation where you'd want to create a managed object without a context.
You should use a child context. The context does a lot of work for you in Core Data (managing relationships, probably change notifying, validation etc), and offers a simple way to cancel / save changes - just save the child context and the data flows back up into the main context, or discard the context to abandon.
A workaround to get the change notifications is to add this override to the NSManagedObject subclass:
override public func willChangeValue(forKey key: String) {
super.willChangeValue(forKey: key)
self.objectWillChange.send()
}
NSManagedObject could be subclassed to add this override (more info here and here)
We can be more specific on the update value if we don't want to trigger the change for every key. This will also work for relationships (not the case for the above solution).
func setName(_ name: String) {
objectWillChange.send()
self.name = name
}
In this case, my AddView updates even if the observed object does not have a context (change notifications are probably trigger only when a context exists for the object). The save button is disabled / enabled based on the following computed property in my NSManagedObject subclass.
var canBeSaved: Bool {
if self.name.isEmpty {
return false
} else {
return true
}
}
I want to be able to just place a View component (plugin) into the page through code and have it appear at some X\Y on the page... but I'm a bit stumped.
Any attempt to add via page.content kinda adds it to the layout\render pass so it occupies space.
So this would get injected into "any" page at "any" time, I have no control over the markup this would be used in (know what I mean?) There is no XML for it and unfortunately the answer can't just be wrap everything in an AbsoluteLayout because one can't mandate that on users apps\layouts.
Thoughts, even possible?
Basically the simplest way to do this is to dynamically and be fully cross platform compatible is to create a AbsoluteLayout item in your JavaScript code, and dynamically insert your item and the AL into the page container.
Code would be something like this:
var AbsoluteLayout = require('ui/layouts/absolute-layout').AbsoluteLayout;
var myAL = new AbsoluteLayout();
var myItem = new myPluginItem();
// Set you left, right, top, bottom coords.
myItem.top = x;
// Add our item to the AbsoluteItem
myAL.addChild(myItem);
var frame = require('ui/frame');
var page = frame.topmost().currentPage;
var LayoutBase = require('ui/layouts/layout-base').LayoutBase;
page._eachChildView(function(view) {
if (view instanceof LayoutBase) {
view.addChild(myAL);
return false;
}
return true;
});
However, if you don't want to do this the really simple way; the only other way is to actually go a bit lower level. You can natively access the iOS view controller (page._ios.view) and the android view (page._nativeView), and then manually add it to the view by using things like addView (https://developer.android.com/reference/android/view/ViewManager.html) or addSubview (https://developer.apple.com/library/ios/documentation/UIKit/Reference/UIView_Class/).
I would like to add you can set the Top and Left properties in TypeScript by importing AbsoluteLayout like so
import {AbsoluteLayout} from 'ui/layouts/absolute-layout';
and then using the exposed functions setLeft or setTop
AbsoluteLayout.setLeft(YourItem, LeftValue);
or
AbsoluteLayout.setTop(YourItem, TopValue);
I have a hierarchical relation inside an entity X, I Have parent lookup which allow to give parent to a record of this entity, and I have created a Subgrid attached to this lookup within the same form of the entity:
The problem is that the display of the button + is unstable in this subgrid, sometimes it appears sometimes no. I dont know if this problem is related to some setting or it is a bug of dynamics crm online last version?
For information, I don't have this problem with other sub-grids.
Thanks in advance,
if you want to add a custom button you may do this as follows
function CreateButton() {
var connectionSubGridPlusBtn = document.getElementById("Connections_addImageButton").parentNode.parentNode;
//Connections_addImageButton is the id of + button
if (connectionSubGridPlusBtn != null) {
//New Button
var div = document.createElement("div");
div.className = "ms-crm-contextButton";
div.innerHTML = "<button id='newButton' type='button' style='width:80px;cursor: pointer;padding:0px' >New Button</button>";
connectionSubGridPlusBtn.appendChild(addVendorDiv);
//Event and url for new
document.getElementById("newButton").onclick = function () {
//Write codefor the button click event
}
}
}
call this function on load of the form
The entity has to be created before you're able to add related entities. You can add disable all required fields, and perform a save in the onload, and you should always see the plus sign.
A slightly better solution is to override the create button for the entity, and rather than directing to the create form, perform a rest entity creation, then direct to that form. Then you don't have to perform a save in the on load.
Suppose I want to pass one object (reference) through several pages. I can navigate and pass parameters via Frame.Navigate(typeof(FirstPage), object). But how to pass the reference back on back press properly?
protected override void OnNavigatedTo(NavigationEventArgs e) {
if (e.Parameter is SomeClass) {
this.someObject = (SomeClass)e.Parameter;
}
else {
this.someObject = new SomeClass();
}
SystemNavigationManager.GetForCurrentView().AppViewBackButtonVisibility = AppViewBackButtonVisibility.Visible;
SystemNavigationManager.GetForCurrentView().BackRequested += OnHardwareButtonsBackPressed;
base.OnNavigatedTo(e);
}
private void OnHardwareButtonsBackPressed(object sender, BackRequestedEventArgs e) {
// This is the missing line!
Frame.Navigate(typeof(FirstPage), this.someObject);
}
But when I press back button it goes back to the FirstPage OnNavigatedTo with no parameter, and then back to the SecondPage OnHardwareButtonsBackPressed and then back to FirstPage OnNavigatedTo with filled parameter.
Could you please advice me some better approach?
In your back handler, don't navigate forwards again, just call GoBack -- and it's typically easier if you handle that at a global level rather than at a page level.
You can store your application state (the things you want to persist across page navigations) in global / static objects, or you could directly modify the object that was passed from the initial navigation (if the calling page still has a reference, it will be able to see the changes).
I would consider doing a search for "MVVM Windows Apps" and looking at some of the results to learn about a common way of building XAML apps.
In short:
I want to show a view or action sheet and only continue code execution after the user has dismissed the view / sheet. So: line one shows the view, line two reads some result variable.
In detail why I would need this:
I'm porting a Windows Forms application over to the iPad. The original implementation has a communication class which uses a web service to communicate with the server. It offers a couple of methods to get data. Conveniently it checks prior to each call if the user still has a valid connection or if he has to re-enter his password for security reasons.
If the password is required, the .NET class shows a modal dialog which blocks any further code executio and if the password was entered, retries the last call it has made before showing the dialog.
Now using CocoaTouch I'm facing a problem. I replaced the code that shows the dialog with a UIActionSheet. Works great but code execution continues immediately, whereas in Windows Forms it is blocked (the next line in Windows Forms after showing the dialogs is to read the entered password from the dialog) until the dialog has been closed.
I tried a Thread.Sleep() until the user dismisses the UIActionSheet but the Thread.Sleep() also blocks the main loop and my view won't even be drawn.
The alternative I currently see is to change all methods in the already working class and give them a return value: if password required, handle it, then retry.
But this means that all over my code I will have to add these checks because at any given moment the password might be needed. That's why it is nested in communication class in Windows Forms.
Any other ideas?
René
Yes, it is possible.
To do this, what you can do is to run the mainloop manually. I have not managed to stop the mainloop directly, so I instead run the mainloop for 0.5 seconds and wait until the user responds.
The following function shows how you could implement a modal query with the above approach:
int WaitForClick ()
{
int clicked = -1;
var x = new UIAlertView ("Title", "Message", null, "Cancel", "OK", "Perhaps");
x.Show ();
bool done = false;
x.Clicked += (sender, buttonArgs) => {
Console.WriteLine ("User clicked on {0}", buttonArgs.ButtonIndex);
clicked = buttonArgs.ButtonIndex;
};
while (clicked == -1){
NSRunLoop.Current.RunUntil (NSDate.FromTimeIntervalSinceNow (0.5));
Console.WriteLine ("Waiting for another 0.5 seconds");
}
Console.WriteLine ("The user clicked {0}", clicked);
return clicked;
}
I think this approach using async/await is much better, and doesn't suffer from freezing the app when rotating the device, or when the autoscrolling interferes and leaves you stuck in the RunUntil loop forever without the ability to click a button (at least these problems are easy to reproduce on iOS7).
Modal UIAlertView
Task<int> ShowModalAletViewAsync (string title, string message, params string[] buttons)
{
var alertView = new UIAlertView (title, message, null, null, buttons);
alertView.Show ();
var tsc = new TaskCompletionSource<int> ();
alertView.Clicked += (sender, buttonArgs) => {
Console.WriteLine ("User clicked on {0}", buttonArgs.ButtonIndex);
tsc.TrySetResult(buttonArgs.ButtonIndex);
};
return tsc.Task;
}