I have a UISplitViewController which has a UINavigationController in the master and a UIViewController in the detail. When the device is orientated into landscape mode I want the normal behaviour to be preserved. I.e. The master gets shown in landscape and hidden in portrait.
However depending what the user clicks in the master depends on which UIViewController is loaded into the detail part of the UISplitViewController. What I would like is for the master to be hidden in landscape mode when a user clicks on a button in the detail UIViewController. The problem is I can't get this to work.
My delegate looks like this (have removed some lines for simpler viewing):
public class SplitControllerDelegate : UISplitViewControllerDelegate {
SplitViewController incomingController;
private bool hideMaster = false;
public override bool ShouldHideViewController (UISplitViewController svc,
UIViewController viewController,
UIInterfaceOrientation inOrientation) {
return hideMaster;
}
public void SetHideMaster(bool value) {
hideMaster = value;
}
}
I then call it from the detail UIViewController like
splitControllerDelegate.SetHideMaster(value);
However nothing changes. I'm unsure of how to make it perform the change? Should the master disappear immediately? What causes the WillHideViewController to fire?
Thanks
Mike
What you're trying to do cannot be done officially. ShouldHideViewController() is called only upon device rotation. So unless you rotate forth and back, your controller won't disappear.
You have various options:
Don't use UIListViewController but some other third party replacement
Use UIViewController containment feature of iOS5 and implement your own split view
Apply a hack to UISplitViewController
About the last point. You should be able to force ShouldHideViewController() being called if you set the Delegate property to NULL and then assign a new delegate. Afterwards, call the WillRotate() method of the split view controller using the current orientation.
I'd go for the 2nd option.
By design you cannot do much with the standard UISplitView, try that third party control :
https://github.com/mattgemmell/MGSplitViewController
Related
I want to prevent the user that he will loose his changes in an EditView when changing the view to another.
I use MVP4G in my project and the project is divided as mvp's structure (one package for the template another one for views ..) is there any solution to detect the EditView in the eventBus. or detect the current View shown to user
Thanks in advance
Thanks to the Navigation Event feature in mvp4g, the presenter will get control before the view changes. At this point the presenter can decide if the navigation will be done or not. This is the correct place in a mvp4g application to save your data.
First zu have to mark all events in the eventbus that will change your view with:
#Event(..., navigationEvent = true)
void goToPage1();
Next your presenters have to implement the NavigationConfirmationInterface and the requires confirm-method:
public class Presenter extends ... implements NavigationConfirmationInterface {
public void confirm(NavigationEventCommand event) {
//pseudo method to verify if the view has changed
if (isViewModified(){
//Window shouldn't be used inside a presenter
//this is just to give a simple example
if (Window.confirm("Are you sure you want to leave?")){
event.fireEvent();
}
} else {
event.fireEvent();
}
}
}
And the last thing to do, is to set the presenter of the current view to the confirmation presenter by calling:
event.fireEvent(false);
This is usually done when the presenter gets control.
You will find the documentation here:
https://github.com/FrankHossfeld/mvp4g/wiki/03.-Defining-EventBus#navigation-event
Thanks to MVP4G's team including El Hoss who gives me a hint to check the MVP4G's blog.. I've solved my problem by following this example
http://mvp4g.blogspot.com/2011/06/navigation-control.html
We're using MVVMCross within our application and I've come up against something that I'm not sure I've solved in the best way possible.
One of our ViewModels contains 3 other view models - a dashboard and 2 lists. In iOS this is presented using a MvxTabBarViewController which works great. Android and WP present this view in a similar manner. An example of the object model is below:
public class ProjectViewModel : MvxViewModel
{
public DashboardViewModel Dashboard {get;set;}
public FirstListViewModel FirstList {get;set;}
public SecondListViewModel SecondList {get;set;}
}
We're now in the situation where if a certain action happens within the DashboardViewModel we would like to instruct the navigation to change the tab in iOS and the same thing to happen on the other platforms.
The only way I've been able to get the tab to change on iOS is to use this.SelectedIndex = 1; from within the iOS ProjectView.
At the moment also the only way I've managed to trigger this change is to fire an event from the DashboardViewModel and then the ProjectViewModel subscribes to this and fires another event which is subscribed to by the ProjectView to instruct it to change the tab in whatever device specific way it needs to. I can't help but think there is a better way to do this.
I've tried taking a look at a custom ViewPresenter for iOS and calling ShowViewModel FirstListViewModel from within the DashboardViewModel but the presenter doesn't appear to be getting used so we just transition normally. My idea was I could get in the middle, cancel the navigation request and then flip the active tab on the ProjectView.
Any suggestions would be appreciated on how we could do this in a better cross platform way using MVVMCross to handle the change if at all possible.
You should be able to do this in any of several ways:
using a custom presenter with overridden Show as you suggest
using a custom presenter with overridden ChangePresentation - and using a custom hint
using a custom binding or a binding to a property within the ProjectView to drive the transition
using a custom IMvxInteraction property
using a custom event from VM to View
using a messenger to send a message from the ViewModels to the Views.
Ultimately lots of these could work and which of these I might choose would depend on which one worked and which one the team are happy with - shipping the working app is always the ultimate goal.
Given where I am with MvvmCross experience, I'd probably opt today for trying the approach of trying a custom IMvxInteraction property. But this might not be for everyone... it certainly might be overkill for this sample...
However, to do this, I would try:
add a public enum Display { Dash, First, Second } to the Core project
add a ProjectViewModel property:
private MvxInteraction<Display> _display = new MvxInteraction< Display >();
public IMvxInteraction<Display> DisplayChange { get { return _display; } }
whenever this ViewModel wants to fire the change it can fire it using e.g. _display.Raise(Display.First)
the ProjectView could then bind Display to its own property which might be implemented like:
private IDisposable _subscription;
private IMvxInteraction<Display> _displayInteraction;
public IMvxInteraction<Display> ChangeDisplay
{
get { return _displayInteraction; }
set
{
if (_subscription != null)
{
_subscription.Dispose();
_subscription = null;
}
_displayInteraction = value;
if (_displayInteraction != null)
{
_subscription = _displayInteraction.WeakSubscribe(DoDisplayChange);
}
}
}
private void DoDisplayChange(Display which)
{
// change the tab display here
}
the binding would be added in ViewDidLoad like:
set.Bind(this).For(v => v.ChangeDisplay).To(vm => vm.DisplayChange);
I want to create a JavaFX table that, in the cells of one column, allows the user to edit XHTML text. I only need very basic formatting capabilities like bold, italic, striketrough.
I have already managed to implement this by using my own subclass of TableCell and using a WebView for each cell (HTMLEditor would of course have been another choice, but my guess is that for my requirements, WebView should be sufficient).
However, to make editing comfortable for the user, I need the following features:
1. The cell height needs to resize if the user enters multi-line text.
2. A context menu (or, if not possible, some other menu or button) should allow formatting parts of the text in a cell as described above (bold, italic..)
Has anybody been successful in implementing something similar ? I have seen suggestions on the web, but they rarely included code samples.
I have succedded doing something similar.
I figured I can share some of the basic clues that allowed me to achieve it.
Resize the whole WebView. For that, the whole WebView must be an editable html page. You achive that by setting contenteditable to true:
<body contenteditable='true' id='content'></body>
You can have a context menu over a webview. But it is something tricky, as you must first disable the original context menu associated to it.
WebView editView;
...
EventDispatcher originalDispatcher = editView.getEventDispatcher();
editView.setEventDispatcher(new WebmenuEventDispatcher(originalDispatcher));
And this is the event dispatcher class:
public class WebmenuEventDispatcher implements EventDispatcher {
private EventDispatcher originalDispatcher;
public WebmenuEventDispatcher(EventDispatcher originalDispatcher) {
this.originalDispatcher = originalDispatcher;
}
#Override
public Event dispatchEvent(Event event, EventDispatchChain tail) {
if (event instanceof MouseEvent) {
MouseEvent mouseEvent = (MouseEvent) event;
if (MouseButton.SECONDARY == mouseEvent.getButton()) {
mouseEvent.consume();
// Show our own menu
cmEdit.show(editView.getScene().getWindow(), mouseEvent.getScreenX(), mouseEvent.getScreenY());
}
}
return originalDispatcher.dispatchEvent(event, tail);
}
}
Now, for setting the font from within that menu, you need a bidirectional Java<->javascript bridge and use some javascript in the webview side.
I am developing a control for an IPAD application (My first time doing Apple development). Its a simple control that mimics a grid - consists of a collection of UIViews (each of which represents a cell) all added to a parent UIView (in a grid like fashion).
One of the requirements is to implement a swipe gesture - the users swipe across the grid to activate/inactivate the cell - this corresponds to a 1/0 in the database.
I create a UISwipeGesture and added it to each of my UIView which represents a cell. That appears to be an incorrect approach as it fires the event for the UIView in which the swipe originated but not across all the UIViews.
My understanding would be that i need to implement the SwipeGesture across the parent UIView which contains all these children UIView. However if i do that how will i know which child UIView has been swiped over? Or any other approach which would make sense?
I know this thread is fairly old, but I created a Swipe extension method that might have helped.
View.Swipe(UISwipeGestureRecognizerDirection.Right).Event += Swipe_Event;
void Swipe_Event(ViewExtensions.SwipeClass sender, UISwipeGestureRecognizer recognizer)
{
View view = sender.View; // do something with view that was swiped.
}
This may not answer your question, but I can speak to the approach I've taken here with a similar use case:
1) I would abandon UIScrollView and use UITableView. You'll notice that UITableView inherits from UIScrollView and has all the performance benefits of virtualization and cell / view re-use. Which you'll find terribly useful as you work towards optimizing your app for performance on device.
2) Utilize the UITableViewCell's ContentView to create custom "Grid" cells. Or better yet, utilize MonoTouch.Dialog if you're not required to create Grid rows ad-hoc.
3) Use this awesome class (props to #praeclarum) to setup gestures in MonoTouch. You essentially provide a UIGestureRecognizer as a generic argument. You can then utilize the LocationInView method to grab the point in the UITableView where the gesture occurred
public void HandleSwipe(UISwipeGestureRecognizer recognizer)
{
if(recognizer.State == UIGestureRecognizerState.Ended) {
var point = recognizer.LocationInView(myTableView);
var indexPath = myTableView.IndexPathForRowAtPoint(point);
// do associated calculations here
}
}
I think you're correct that the gesture recognizer has to be attached to the parent view. In the action method associated with the gesture recognizer I think you can use the Monotouch equivalent of CGRectContainsPoint() to determine whether the swipe occurred in a particular subview. I imagine you would have to iterate through the subviews until you found the one in which the swipe occurred. I'm not aware of a method that would immediately identify the swiped subview.
According to the documentation, Display.setCurrent doesn't work if the current displayable is an alert. This is a problem as I would like to pop up another alert when the user selects a command. Does anyone know how to work around this so that we can go from one alert to another? I am using CLDC 1.0 and MIDP 2.0.
Additional Information
The spec does allow us to edit an alert while it is on screen, but some Nokia phones don't handle it well at all. So I am now trying to go from the alert to a blank canvas, then back to the alert. Of course I don't want the user to interact with the previous canvas, so it seems that I am forced to create a new blank canvas. As a sidenote, this has the slight disadvantage of looking worse on phones which still have the previous screen when an alert is shown.
The bigger problem is how to transition from the blank canvas back to an alert once the canvas is loaded. Testing on the Motorola emulator revealed that showNotify is not called after returning from an alert to the previous screen. I guess I could create the next alert in the paint method, but this seems like a ugly hack.
OK, so your problem is that you can't set it up to do:
Display.setCurrent(alert1, alert2);
and
Display.setCurrent(alert2);
is also not possible if the current Displayable is already alert1.
So how about put an intermediate Displayable item that is blank and that immediately changes to the next alert? Assuming the current Displayable is alert1, like this in your alert1's command block:
Display.setCurrent(blankForm);
Display.setCurrent(alert2);
That should work assuming you are not using the default 'Dismiss' command. So basically it goes from alert1->(blankForm->alert2).
I couldn't find a way around this, so I just used the paint hack.
public class AlertPage extends Canvas{
MIDlet midlet;
Alert alert;
private AlertPage(MIDlet midlet){
this.midlet=midlet;
}
protected void paint(Graphics arg0){
//Yep, this is a hack, but showNotify doesn't seem to work well for Motorola
if(alert!=null){
Display d=Display.getDisplay(midlet);
d.setCurrent(alert);
alert=null;
}
}
public static void showAlert(MIDlet m, Alert a){
AlertPage page=new AlertPage(m);
Display d=Display.getDisplay(m);
page.alert=a;
d.setCurrent(page);
}
}