I need to disable closing of my app (It works in full screen) by pressing alt + f4. I read other threads and wrote the following code:
override fun start(stage: Stage) {
Platform.setImplicitExit(false)
stage.addEventFilter(KeyEvent.KEY_PRESSED) {
if (it.isAltDown && it.code == KeyCode.F4) {
it.consume()
}
}
stage.setOnHiding { _ -> EventHandler<WindowEvent> { it.consume() } }
stage.setOnCloseRequest { _ -> EventHandler<WindowEvent> {
it.consume()
} }
stage.fullScreenExitKeyCombination = KeyCombination.NO_MATCH
stage.fullScreenExitHint = ""
stage.isFullScreen = true
super.start(stage)
}
The problem is that my app doesn't close but stage is being hidden. How can I disable the form from being hidden by pressing alf+f4?
I tried setting the onCloseRequest to the window not the stage. On OS X, after I hit cmd + q the window minimises, but does not close. I guess the best solution I was able to get was:
//Method in my view
override fun onDock() {
super.onDock()
currentWindow?.setOnCloseRequest {
it.consume()
currentStage?.isFullScreen = true //Set the window to fullscreen again
}
}
As I said, this behaviour is on OS X, I guess it will be different across various OSs...
Related
Am using DispatchTouchEvent and need to detect a single tap, for that considered the down point and up point, when both are same, detected the single tap.
But this works fine in some android devices, but in some device, there in a pixel difference in down and up point. How to overcome this?
public override bool DispatchTouchEvent(MotionEvent e)
{
if (e.Action == MotionEventActions.Down)
{
touchDownPoint = new Xamarin.Forms.Point(e.GetX(), e.GetY());
}
else if (e.Action == MotionEventActions.Up)
{
var position = new Xamarin.Forms.Point(e.GetX(), e.GetY());
if (touchDownPoint == position) // this condition not gets satisfied in some devices
{
// single tap detected.
}
}
else if (e.Action == MotionEventActions.Move)
{
}
return base.DispatchTouchEvent(e);
}
I was taking a look at this :
tornadofx
and tried to expand on it with database connection and little more options, (not all of them make sense, but its just playing in a sandbox).
Even though table can be directly edited and the data will persist in database, i did try to do edit through text fields too. actual table editing would happen through different view and not table itself, as i said its just example.
Database used is Jetbrains Exposed.
object Categories : IntIdTable() {
val name = varchar("name", 64).uniqueIndex()
val description = varchar("description", 128)
}
class Category(id: EntityID<Int>) : IntEntity(id) {
companion object : IntEntityClass<Category>(Categories)
var name by Categories.name
var description by Categories.description
override fun toString(): String {
return "Category(name=\"$name\", description=\"$description\")"
}
}
now controller looks something like this, functions are just rudimentary and picked as an example.
typealias ModelToDirtyState = Map.Entry<CategoryModel, TableColumnDirtyState<CategoryModel>>
class CategoryModel() : ItemViewModel<Category>() {
val name: SimpleStringProperty = bind(Category::name)
val description: SimpleStringProperty = bind(Category::description)
}
class DBController : Controller() {
val categories: ObservableList<CategoryModel> by lazy {
transaction {
SchemaUtils.create(Categories)
Category.all().map {
CategoryModel().apply {
item = it
}
}.observable()
}
}
init {
Database.connect(
"jdbc:mysql://localhost:3306/test", driver = "com.mysql.cj.jdbc.Driver",
user = "test", password = "test"
)
TransactionManager.manager.defaultIsolationLevel = Connection.TRANSACTION_SERIALIZABLE
}
fun deleteCategory(model: CategoryModel) {
runAsync {
transaction {
model.item.delete()
}
}
categories.remove(model)
}
fun updateCategory(model: CategoryModel) {
transaction {
Categories.update {
model.commit()
}
}
}
fun commitDirty(modelDirtyMappings: Sequence<ModelToDirtyState>) {
transaction {
modelDirtyMappings.filter { it.value.isDirty }.forEach {
it.key.commit()
println(it.key)// commit value to database
it.value.commit() // clear dirty state
}
}
}
Just to quickly comment on controller, delete method works as "intended" however the update one does not, it does not work in sense that after using delete item is remove both from database and tableview(underlying list) itself, and when i do update its not, now i know the reason, i call remove manually on both database and list, now for update perhaps i could do change listener, or maybe tornadofx can do this for me, i just cant set it up to do it. Following code will make things clearer i think.
class CategoryEditor : View("Categories") {
val categoryModel: CategoryModel by inject()
val dbController: DBController by inject()
var categoryTable: TableViewEditModel<CategoryModel> by singleAssign()
var categories: ObservableList<CategoryModel> by singleAssign()
override val root = borderpane {
categories = dbController.categories
center = vbox {
buttonbar {
button("Commit") {
action {
dbController.commitDirty(categoryTable.items.asSequence())
}
}
button("Roll;back") {
action {
categoryTable.rollback()
}
}
// This model only works when i use categorytable.tableview.selected item, if i use categoryModel, list gets updated but not the view itself
// Question #1 how to use just categoryModel variable without need to use categorytable.tableview.selecteditem
button("Delete ") {
action {
val model = categoryTable.tableView.selectedItem
when (model) {
null -> return#action
else -> dbController.deleteCategory(model)
}
}
}
//And here no matter what i did i could not make the view update
button("Update") {
action {
when (categoryModel) {
null -> return#action
else -> dbController.updateCategory(categoryModel)
}
categoryTable.tableView.refresh()
}
}
}
tableview<CategoryModel> {
categoryTable = editModel
items = categories
enableCellEditing()
enableDirtyTracking()
onUserSelect() {
//open a dialog
}
//DOES WORK
categoryModel.rebindOnChange(this) { selectedItem ->
item = selectedItem?.item ?: CategoryModel().item
}
// Question #2. why bindSelected does not work, and i have to do it like above
//DOES NOT WORK
// bindSelected(categoryModel)
//
column("Name", CategoryModel::name).makeEditable()
column("Description", CategoryModel::description).makeEditable()
}
}
right = form {
fieldset {
field("Name") {
textfield(categoryModel.name)
}
}
fieldset {
field("Description") {
textfield(categoryModel.description)
}
}
button("ADD CATEGORY") {
action {
dbController.addCategory(categoryModel.name.value, categoryModel.description.value)
}
}
}
}
}
I apologize for huge amount of code, also in last code snipped i left questions in form of comments where i fail to achive desired results.
I am sure i am not properly binding code, i just dont see why, also i sometimes use one variable to update data, my declared one "categoryModel" and sometimes i use tableview.selecteditem, it just seems hacky and i cant seem to grasp way.
Thank you!
I am trying to find a way to detect (or receive notification) that a Node has been added to a Scene and is visible.
I am creating Node objects off the main JavaFx thread and add them to the Stage and Scene using Platform.runLater(). However I would like the Node object to receive notification that is has been added to the Scene and is visible, for example I wish to trigger an animation to start.
I can't seem to find any property or method to add a listener to capture such an event. Any suggestions?
The third-party JavaFX library ReactFX has a mechanism for this, and this exact use case is cited in the blog. In short, you can do
Val<Boolean> showing = Val.flatMap(node.sceneProperty(), Scene::windowProperty)
.flatMap(Window::showingProperty);
and then of course
showing.addListener((obs, wasShowing, isNowShowing) -> {
if (isNowShowing) {
// node is showing
} else {
// node is not showing
}
});
The standard library has a version of this, but it is very badly written. (It is not typesafe, has no compile-time checking that the properties exist, and also pipes a lot of unnecessary warnings to standard error if any of the properties in the "chain" are null, even though the API docs indicate this is a supported use case.) If you want to do this with the standard JavaFX library, you can do
BooleanBinding showing = Bindings.selectBoolean(node.sceneProperty(), "window", "showing");
and then use the binding the same way as above.
Finally, you could do all this by hand, but it gets a bit ugly to manage the listeners properly:
BooleanProperty showing = new SimpleBooleanProperty();
ChangeListener<Window> windowListener = (obs, oldWindow, newWindow) -> {
showing.unbind();
if (newWindow != null) {
showing.bind(newWindow.showingProperty());
} else {
showing.set(false);
}
};
ChangeListener sceneListener = (obs, oldScene, newScene) -> {
showing.unbind();
if (oldScene != null) {
oldScene.windowProperty().removeListener(windowListener);
}
if (newScene == null) {
showing.set(false);
} else {
newScene.windowProperty().addListener(windowListener);
if (newScene.getWindow() == null) {
showing.set(false);
} else {
showing.bind(newScene.getWindow().showingProperty());
}
}
};
node.sceneProperty().addListener(sceneListener);
if (node.getScene() == null) {
showing.set(false);
} else {
node.getScene().windowProperty().add(windowListener);
if (node.getScene().getWindow() == null) {
showing.set(false);
} else {
showing.bind(node.getScene().getWindow().showingProperty());
}
}
You can add a listener to the children property of container node into which you are adding the new node.
grid.getChildren().addListener((ListChangeListener<? super Node>) change -> {
System.out.println(change.getList().get(0).getTypeSelector());
});
change.getList().get(0) returns the first node that is added to grid object.
After James's comment, I have looked up and yes, it is possible to do it from node's perspective as well. You can listen to parentProeprty's changes on the node. Following snippet shows the way to do it.
Button b = new Button("Test");
b.parentProperty().addListener((observable, oldValue, newValue) -> {
System.out.println("added to a container " + newValue);
});
answerPane.getChildren().add(b);
Let me explain my current issue right now:
I have a webapp located at domain A. Let's call it A-App. I open an iframe from A-App that points to a Flex app on domain B. We'll call it B-FlexApp. B-FlexApp wants to post some data to another app located on the same domain, we'll call it B-App. The problem is that in IE the communication breaks somewhere between B-FlexApp and B-App while B-FlexApp is opened in the iframe. This only happens in IE.
However when opening B-FlexApp in a new window, posting the data to B-App works just fine. How to overcome this? Dropping the iframe is not possible.
There´s a issue with AS3 navigateToURL and IE. You can try calling javascript to navigate: I have a little utility class to handle this:
//class URLUtil
package com
{
import flash.external.*;
import flash.net.*;
public class URLUtil extends Object
{
protected static const WINDOW_OPEN_FUNCTION:String="window.open";
public function URLUtil()
{
super();
return;
}
public static function openWindow(arg1:String = "", arg2:String="_blank", arg3:String=""):void
{
var browserName:String = getBrowserName();
switch (browserName)
{
case "Firefox":
{
flash.external.ExternalInterface.call(WINDOW_OPEN_FUNCTION, arg1, arg2, arg3);
break;
}
case "IE":
{
flash.external.ExternalInterface.call("function setWMWindow() {window.open(\'" + arg1 + "\');}");
break;
}
case "Safari":
case "Opera":
{
flash.net.navigateToURL(new URLRequest(arg1), arg2);
break;
}
default:
{
flash.net.navigateToURL(new URLRequest(arg1), arg2);
break;
}
}
return;
}
private static function getBrowserName():String
{
var str:String="";
var browserName:String = ExternalInterface.call("function getBrowser(){return navigator.userAgent;}");
if (!(browserName == null) && browserName.indexOf("Firefox") >= 0)
{
str = "Firefox";
}
else
{
if (!(browserName == null) && browserName.indexOf("Safari") >= 0)
{
str = "Safari";
}
else
{
if (!(browserName == null) && browserName.indexOf("MSIE") >= 0)
{
str = "IE";
}
else
{
if (!(browserName == null) && browserName.indexOf("Opera") >= 0)
{
str = "Opera";
}
else
{
str = "Undefined";
}
}
}
}
trace("Browser: \t" + str);
return str;
}
}
}
and you call it like:
btn.addEventListener(MouseEvent.CLICK, onBTNClick);
function onBTNClick(evt:MouseEvent):void
{
URLUtil.openWindow(YOUR_URL_STRING);
}
Hope it helps!
It is better to let the browser actually does the "navigate to URL" function instead of Flex.
For example, in the page that contains the Flex app, the page would contain a Javascript function call handleNavigationRequest(pageName, target). In the Flex application, you may utilize ExternalInterface, and call the handleNavigationRequest.
By using this paradigm, the Flex application would not have to figure the details as to how the external implementations such as frame setup, etc, and you end up having a cleaner and less-coupled design.
I've found out that i can use swfObject to embed the flash object thus the iframe implementation is completely useless. Embedding the flash component in the overlay, instead of opening it in an iframe, makes IE behave properly.
I had the same problem and I solved it simply passing the second argument (browser window) to the function:
navigateToUrl(url,"_blank"); , in my case I use "_blank".
It works with IE8 and IE9.
Davide
I have a VideoDisplay instance playing some video. When I click on the video slider (also my component) the property videoDisplay.playheadTime is set and the videoDisplay.state goes from 'playing' into a 'seeking' state for a brief moment (the videoDisplay seeks for a new position and then plays the video again). Intended bevaiour.
But if I'm (or any random user) fast enough, I can set the playheadTime again while the player is still in 'seeking' state. When repeated several times every click is enqueued and the videoDisplay jump on every place of the video I have clicked(this is happening in an interval about 10-15 second after my last click). When I use live dragging the videoDisplay, overwhelmed by seekings, goes into 'error' state.
My question is - is there any way to cancel seeking state of the VideoDisplay class? For example player is in 'seeking' state, I set playheadTime, and the player forgets about last seeking and try to find the new place of the video.
I will downvote pointless answers like 'Use the Flex4 VideoPlayer class'!
One possible way is wrap the video display in a component and manage the seek a little better yourself. So if someone calls seek, make sure that the video is not currently seeking, if so, then wait till the current operation is complete before proceeding to the new one. If the user tries to seek again, discard all currently pending operations and make the latest one the next operation. Working on this exact problem right now.... Here's the code:
public function Seek(nSeconds:Number, bPlayAfter:Boolean):void
{
trace("Player Seek: "+ nSeconds);
var objSeekComand:VideoPlayerSeekCommand = new VideoPlayerSeekCommand(ucPlayer, nSeconds, bPlayAfter);
ProcessCommand(objSeekComand);
}
protected function ProcessCommand(objCommand:ICommand):void
{
if(_objCurrentCommand != null)
{
_objCurrentCommand.Abort();
}
_objCurrentCommand = objCommand
objCommand.SignalCommandComplete.add(OnCommandComplete);
objCommand.Execute();
}
Here's the Command
public class VideoPlayerSeekCommand extends CommandBase
{
private var _ucVideoDisplay:VideoDisplay;
private var _nSeekPoint:Number;
private var _bPlayAfterSeek:Boolean;
private var _bIsExecuting:Boolean;
public function VideoPlayerSeekCommand(ucVideoDisplay:VideoDisplay, nSeekPointInSeconds:Number, bPlayAfterSeek:Boolean, fAutoAttachSignalHandler:Function = null)
{
_ucVideoDisplay = ucVideoDisplay;
_nSeekPoint = nSeekPointInSeconds;
_bPlayAfterSeek = bPlayAfterSeek;
super(fAutoAttachSignalHandler);
}
override public function Execute():void
{
//First check if we are playing, and puase if needed
_bIsExecuting = true;
if(_ucVideoDisplay.playing == true)
{
_ucVideoDisplay.addEventListener(MediaPlayerStateChangeEvent.MEDIA_PLAYER_STATE_CHANGE, OnPlayerStateChangedFromPlay, false, 0, true);
_ucVideoDisplay.pause();
}
else
{
DoSeek();
}
}
protected function OnPlayerStateChangedFromPlay(event:MediaPlayerStateChangeEvent):void
{
_ucVideoDisplay.removeEventListener(MediaPlayerStateChangeEvent.MEDIA_PLAYER_STATE_CHANGE, OnPlayerStateChangedFromPlay);
if(_bIsExecuting == true)
{
if(_ucVideoDisplay.playing == false)
{
DoSeek();
}
else
{
throw new Error("VideoPlayerSeekAndPlayCommand - OnPlayerStateChangedFromPlay error");
}
}
}
private function DoSeek():void
{
if(_bIsExecuting == true)
{
_ucVideoDisplay.seek(_nSeekPoint);
CheckSeekComplete();
}
}
private function CheckSeekComplete():void
{
if(_bIsExecuting == true)
{
if (Math.abs( _ucVideoDisplay.currentTime - _nSeekPoint) < 2)
{
if(_bPlayAfterSeek == true)
{
_ucVideoDisplay.play();
}
DispatchAndDestroy();
}
else
{
CoreUtils.CallLater(CheckSeekComplete, .07);
}
}
}
override public function Abort():void
{
_bIsExecuting = false;
SignalCommandComplete.removeAll();
}
}
Im Using AS3 Signals here instead of events, and the CoreUtils.Call later you can use setInterval, or a Timer. But the idea is to not call seek until the video is paused, and to keep track of when the seek is complete.