I would like to track a global position of an object (or relative to one of it's ancestors) and bind it to some other item's position.
I was thinking about using mapFromItem as follows:
SomeObject {
x: ancestor.mapFromItem(trackedObject, trackedObject.x, 0).x
y: ancestor.mapFromItem(trackedObject, 0, trackedObject.y).y
}
The problem with this approach is that the mapFromItem is evaluated once and doesn't update as one of it's arguments gets updated. Moreover the mapping sometimes returns the new position altered by an offset I'm unable to track in the code (but that's not the matter at hand).
My second idea was to calculate the global position by implementing a function that would recursively sum the offsets, stopping at the provided ancestor (something like calculateOffsetFrom(ancestor)). Still this is just a function and as far as I'm concerned it won't get re-evaluated as one of the ancestors position changes (unless, in that function, I'll bind calling it to the onXChanged signal for each one of the ancestors along the way, which seems like a dirty solution).
So in the end I've added properties to the object I intend to track and then I bind to them:
TrackedObject {
property real offsetX: x + parent.x + parent.parent.x + parent.parent.parent.x ...
property real offsetY: y + parent.y + parent.parent.y + parent.parent.parent.y ...
}
SomeObject {
x: trackedObject.globalX
y: trackedObject.globalY
}
But well... yeah... this one doesn't scale at all and is as ugly as it gets.
Does anyone have any idea how this problem might be solved in a cleaner way?
Edit:
As far as I'm concerned I can't use anchors in this case. The SomeObject component is a custom component drawing a bezier curve from one point to another (it will connect two TrackedObjects). For that I need the difference between the coordinates. If I'm correct anchors don't provide any way of calculating the distance between them.
This is a hard point, but here is the hack i used in one of my projects : to make blue rect which is in another parent than green rect move, to stay aligned with it, when green rect moves but also when yellow rect (green rect parent) moves :
import QtQuick 2.0;
Rectangle {
id: window;
width: 800;
height: 480;
property bool globalBit : true;
function updatePos (item_orig, item_dest, bit) {
var pos_abs = window.mapFromItem (item_orig.parent, item_orig.x, item_orig.y);
return window.mapToItem (item_dest.parent, pos_abs.x, pos_abs.y);
}
Rectangle {
id: rectYellow;
width: 400;
height: 300;
x: 300;
y: 200;
color: "yellow";
onXChanged: { globalBit = !globalBit; }
onYChanged: { globalBit = !globalBit; }
MouseArea {
drag {
target: rectYellow;
minimumX: 0;
minimumY: 0;
maximumX: (rectYellow.parent.width - rectYellow.width);
maximumY: (rectYellow.parent.height - rectYellow.height);
}
anchors.fill: parent;
}
Rectangle {
id: rectGreen;
x: 100;
y: 100;
width: 50;
height: 50;
color: "green";
MouseArea {
drag {
target: rectGreen;
minimumX: 0;
minimumY: 0;
maximumX: (rectGreen.parent.width - rectGreen.width);
maximumY: (rectGreen.parent.height - rectGreen.height);
}
anchors.fill: parent;
}
}
}
Rectangle {
id: rectBlue;
x: pos.x + 50;
y: pos.y + 50;
width: 50;
height: 50;
color: "blue";
property var pos : updatePos (rectGreen, rectBlue, globalBit);
}
}
The trick is to bring all coordinates back to the first common ancestor, using both mapfromItem and mapToItem, and to force the function to be re-evaluated, just put a global boolean flag that you pass to the computing function, and that you invert each time a movable element on your map moves... You don't have to put it every where, just on parents of items that can move and are inside the ancestor item.
So it works, your positions will always be right, and it's quite scalable and doesn't add much code.
The solution below will trigger the qmlElementToTrack.onPropertyNameXChanged() and qmlElementToTrack.onPropertyNameYChanged() events each time one of its parents 'x' or 'y' values change.
It does this by attaching to each parent's onXChanged() and onYChanged() signals. When one of those values changes, it recalculates the propertyNameX or propertyNameY values by traversing all of qmlElementToTrack's parents.
To make it 'relative' positioned (instead of 'absolute'), add current !== qmlElementToStopAt to each while() condition.
ElementToTrack {
id: qmlElementToTrack
property real propertyNameX: 0
property real propertyNameY: 0
}
setPositionChangedToParents(qmlElementToTrack);
/**
Connect to each parent's 'onXChanged' and 'onYChanged' signals.
*/
setPositionChangedToParents = function(current) {
while (current && current.parent) {
current.onXChanged.connect(calculatePropertyNameX);
current.onYChanged.connect(calculatePropertyNameY);
current = current.parent;
}
};
/**
Disconnects the signals set to all parents.
*/
removePositionChangedFromParents = function(current) {
while (current && current.parent) {
current.onXChanged.disconnect(calculatePropertyNameX);
current.onYChanged.disconnect(calculatePropertyNameY);
current = current.parent;
}
};
/**
When any parent's 'x' changes, recalculate the 'x' value for the 'property name'.
*/
calculatePropertyNameX = function() {
var calculatedX, current;
calculatedX = 0;
current = qmlElementToTrack;
while (current && current.parent) {
calculatedX += current.x;
current = current.parent;
}
propertyNameX = calculatedX;
};
/**
When any parent's 'y' changes, recalculate the 'y' value for the 'property name'.
*/
calculatePropertyNameY = function() {
var calculatedY, current;
calculatedY = 0;
current = qmlElementToTrack;
while (current && current.parent) {
calculatedY += current.y;
current = current.parent;
}
propertyNameY = calculatedY;
};
I don't know whether this will help, but for the above yellow rect,blue rect and green rect problem mentioned by TheBootroo, I used the below code to solve the problem
import QtQuick 2.0;
Rectangle {
id: window;
width: 800;
height: 480;
Rectangle {
id: rectYellow;
width: 400;
height: 300;
x: 300;
y: 200;
color: "yellow";
MouseArea {
drag {
target: rectYellow;
minimumX: 0;
minimumY: 0;
maximumX: (rectYellow.parent.width - rectYellow.width);
maximumY: (rectYellow.parent.height - rectYellow.height);
}
anchors.fill: parent;
}
Rectangle {
id: rectGreen;
x: 100;
y: 100;
width: 50;
height: 50;
color: "green";
MouseArea {
drag {
target: rectGreen;
minimumX: 0;
minimumY: 0;
maximumX: (rectGreen.parent.width - rectGreen.width);
maximumY: (rectGreen.parent.height - rectGreen.height);
}
anchors.fill: parent;
}
}
}
Rectangle {
id: rectBlue;
//Need to acheive the below behvior(commented)
//x: window.x+rectYellow.x+rectGreen.x+50
//y: window.y + rectYellow.y +rectGreen.y+50
width: 50;
height: 50;
color: "blue";
}
Component.onCompleted: {
rectBlue.x =Qt.binding(
function()
{
//Returns window.x+rectYellow.x+rectGreen.x+rectGreen.width
var docRoot = null;
var x=rectGreen.x;
if(!docRoot)
{
docRoot = rectGreen.parent;
x+=docRoot.x;
while(docRoot.parent)
{
docRoot = docRoot.parent;
x+=docRoot.x
}
}
return x+rectGreen.width;
}
)
rectBlue.y = Qt.binding(
function()
{
//Returns window.y+rectYellow.y+rectGreen.y+rectGreen.height
var docRoot = null;
var y=rectGreen.y
if(!docRoot)
{
docRoot = rectGreen.parent;
y+=docRoot.y;
while(docRoot.parent)
{
docRoot = docRoot.parent;
y+=docRoot.y
}
}
return y+rectGreen.height;
}
)
}
}
The idea is to calculate the position of blue rectangle relative to the green rectangle, by
calculating the position of green rectangle and its visual ancestors.
The inspiration behind this solution is -> http://developer.nokia.com/Community/Wiki/How_to_create_a_Context_Menu_with_QML
Tracking certain Item's global positions seems like an important problem if developing some complex graphics interaction. I came up with a relatively simple & graceful solution. Here is my core codes:
Item{
id: globalRoot
signal globalPositionChanged(Item item, real newX, real newY);
function tracking(item){
var obj = item;
var objN;
function onGlobalXYChanged(){
var pt = mapFromItem(item, item.x, item.y);
globalRoot.globalPositionChanged(item, pt.x, pt.y);
}
do{
objN = obj.objectName;
obj.xChanged.connect(onGlobalXYChanged);
obj.yChanged.connect(onGlobalXYChanged);
obj = obj.parent;
}while(objN !== "furthestAncestorObjectName");
}
}
The core idea is: what essentially makes an Item's global position change? It maybe itself, its parent or its parent's parent etc. So make a traverse back to its furthest parent and connect each of its ancestor's x/y change signal to a function, within which we get the item's global position and broadcast outside.
I have tried to improve on #Shubhanga's answer a bit by moving the code into its own ItemPositionTracker.qml file:
import QtQuick 2.3
Item {
id: root
property Item trackedItem
property Item movedItem
Component.onCompleted: {
movedItem.x =
Qt.binding(
function()
{
if (trackedItem === null) return 0;
var docRoot = trackedItem;
var x = trackedItem.x;
while(docRoot.parent)
{
docRoot = docRoot.parent;
x += docRoot.x
}
return x;
}
)
movedItem.y =
Qt.binding(
function()
{
if (trackedItem === null) return 0;
var docRoot = trackedItem;
var y = trackedItem.y
while(docRoot.parent)
{
docRoot = docRoot.parent;
y += docRoot.y
}
return y;
}
)
}
}
The code can now be added to any QML object like this:
ItemPositionTracker {
trackedItem: rectGreen
movedItem: rectBlue
}
Which makes rectBlue follow rectGreen.
Related
If I replace component with Rectangle, then the scrollview works.....
Any idea? How to include dynamic parts in ScrollView?
Thanks in advance
ScrollView {
id:template1
contentHeight: 2000
....
Component {
id: routeComp
CruiseRouteButton {
}
}
function createTab() {
}
Component.onCompleted: {
for(var i = 0; i < mainWnd.config.routes.length; i++) {
....
let t = routeComp.createObject(template1, cfg);
routes[i] = t
}
createTab()
}
This component is used to dynamically set some routes for user to choose
I have a ListView that displays a list of all notifications for a user. Right now since I'm using Component.onCompleted, if the list updates, the new list is not displayed, but the one which existed during instantiation. How could we solve this? Would using a Loader with a separate component instead help?
property int numNotifications: backend_service.num_notifications
property var notifications: []
onNumNotificationsChanged: {
for(var x=0; x<numNotifications; x++) {
var notif = backend_service.notifications.get(x);
notifications.push(notif)
}
}
Rectangle {
anchors.horizontalCenter: parent.horizontalCenter
color: "black"
height: 500
width: 0.95 * parent.width
ListView {
anchors.fill: parent
model: notifModel
delegate: notifDelegate
}
}
ListModel {
id: notifModel
Component.onCompleted: {
for(var i in notifications) {
notifModel.append({"name": notifications[i]})
}
}
}
Component {
id: notifDelegate
Row {
spacing: 10
Text { text: name; color: "white" }
}
}
Component.onCompleted only runs when the object is built and never again. So using that method to add items to the model is useless, instead you should use the function that reports the new data:
onNumNotificationsChanged: {
for(var x=0; x<numNotifications; x++) {
var notif = backend_service.notifications.get(x);
notifModel.append({"name": notif})
}
}
I am unsure what the markdirty() function does in QML. The QT documentation of that function does not seem clear to me.
My interpretation of this is that it allows us to track changes to a small subset of the canvas, so that any changes to that part redraws everything only in that part, using paint()
Doing requestPaint() on the other hand would be far more inefficient because it would redraw the whole canvas.
Is this correct? Some simple example codes would be quite helpful in understanding the usecase of markDirty()
That's a widely used term in programming especially GUI. Using that you can mark part of the canvas as in need of updating. So if this part is visible the render engine will fire paint(rect region) as soon as possible. In the onPaint handler you should repaint only items inside this region. requestPaint() does almost the same but for all the visible region.
Check the output in the example below:
import QtQuick 2.9
import QtQuick.Window 2.2
Window {
visible: true
width: 800
height: 800
title: qsTr("QML Test")
Canvas {
property var shapes: [20,20,220,20,20,220,220,220]
id: canvas
width: 400
height: 400
anchors.centerIn: parent
onPaint: {
console.log(region);
var ctx = getContext("2d");
ctx.beginPath();
ctx.fillStyle = Qt.rgba(Math.random(),Math.random(),Math.random(),1);
// draw 4 circles
for(var i = 0;i < canvas.shapes.length;i +=2) {
var x = canvas.shapes[i];
var y = canvas.shapes[i + 1];
var width = 160;
var height = 160;
// check the circle is inside the region
if(!( (x + width) < region.x || x > (region.x + region.width) || (y + height) < region.y || y > (region.y + region.height) ) ) {
ctx.ellipse(x, y, width, height);
}
}
ctx.fill();
}
}
Timer {
id: timer
property int step: 0
interval: 2000
repeat: true
running: true
onTriggered: {
switch(step++) {
case 0: canvas.markDirty(Qt.rect(0, 0, 200, 200)); break;
case 1: canvas.requestPaint(); break;
case 2: timer.stop(); break;
}
}
}
}
I have 2 QQuickItems like below which I can fetch on C++ side using the QMLEngine like this.
QQuickItem * quick_item_1 = m_qml_engine->rootObjects()[0]->findChild<QQuickItem*>("quickitem1");
QQuickItem * quick_item_2 = m_qml_engine->rootObjects()[0]->findChild<QQuickItem*>("quickitem2");
Note: quick_item_1's immediate parent is different & quick_item_2's immediate parent is also different. But they both are drawn on the same application window into different immediate parents.
I am drawing both of them offscreen on a different QQuickItem. Let's call it new_parent_surface. I draw both these items on new_parent_surface by changing their parent to new_parent_surface like this.
quick_item_1->setParentItem(new_parent_surface);
quick_item_2->setParentItem(new_parent_surface);
This works fine for the objective of drawing them on a new parent QQuickItem. I get the both quick_item_1 & quick_item_2 drawn on new_parent_surface. Even though new_parent_surface is not drawn on UI, but if I take a snapshot using grabToImage of new_parent_surface, I can see the 2 items drawn on them. Fine till here.
However the positioning of quick_item_1 & quick_item_2 is not correct. I want to position them similar to the way they were positioned their original parent item. I can do some percentage math & try positioning them the same way as they were drawn on their original parent but isn't there a QQQuickItem or Qt API to translate this positioning to a new parent?
I tried to look into QQuickItem's mapping APIs like mapToItem & trying them out like this.
quick_item_2->mapToItem(new_parent_surface, quick_item_2->position());
But the positioning is still not correct.
So, how can I map a QQuickItem's position into its new parent QQuickItem after doing a setParentItem?
Items position is always relative to its parent. And position of the parent is relative to its parent and and so on. But you always can get both relative or global position. QML has lots of coordination translation function. Here is small example that could explain the issue:
import QtQuick 2.7
import QtQuick.Window 2.0
import QtQuick.Controls 2.1
Window {
id:container
width: 800
height: 800
visible: true
Component {
id: rect
Rectangle {
property bool imParent: false
x: 50 + Math.round(Math.random() * 550)
y: 50 + Math.round(Math.random() * 550)
width: 100 + Math.round(Math.random() * 100)
height: 100 + Math.round(Math.random() * 100)
color: Qt.rgba(Math.random(),Math.random(),Math.random(),1);
Drag.active: dragArea.drag.active
MouseArea {
id: dragArea
anchors.fill: parent
drag.target: parent
}
Text {
anchors.centerIn: parent
text: imParent ? "I'm parent" : "Drag me"
color: "white"
}
}
}
Rectangle {
id: blk
x: 10
y: 10
z: 100
parent: null
height: 50
width: 50
radius: 5
border.color: "white"
color: "black"
}
Repeater {
id: reptr
model: 5
property int pos: 0
Loader {
id: loader
sourceComponent: rect
onLoaded: {
if(blk.parent == null) {
blk.parent = loader.item;
loader.item.imParent = true;
}
}
}
}
Row {
anchors.horizontalCenter: container.contentItem.horizontalCenter
spacing: 2
Button {
text: "Reparent relative to parent"
onClicked: {
reptr.pos ++;
if(reptr.pos >= reptr.model) {
reptr.pos = 0;
}
var item = reptr.itemAt(reptr.pos).item;
blk.parent.imParent = false;
blk.parent = item;
blk.parent.imParent = true;
}
}
Button {
text: "Reparent relative to scene"
onClicked: {
reptr.pos ++;
if(reptr.pos >= reptr.model) {
reptr.pos = 0;
}
var item = reptr.itemAt(reptr.pos).item;
var coord = blk.mapToGlobal(blk.x, blk.y);
blk.parent.imParent = false;
blk.parent = item;
blk.parent.imParent = true;
coord = blk.mapFromGlobal(coord.x, coord.y);
blk.x = coord.x;
blk.y = coord.y;
}
}
}
}
In MouseArea.onEntered, can I detect if the cause for the event firing is only that the MouseArea moved and came to be under the cursor, rather than the other way round?
I thought of doing something like this: (pseudocode)
MouseArea {
// ...
property bool positionDirty = false
Connections {
target: window
onAfterRendering: {
positionDirty = false;
}
}
onMouseAreaPosChanged: {
positionDirty = true;
}
onEntered: {
if(positionDirty) {
positionDirty = false;
return;
}
// handle event here
}
}
But this makes the assumption that entered will be fired after mouseAreaPosChanged, and before window.afterRendering. And I'm not confident in that assumption.
Also, it doesn't work when an ancestor of the MouseArea moves, or when the MouseArea is positioned/sized via anchoring.
Assumptions:
This only affects the edge case, that both, the cursor and the MouseArea are moving.
My Assumption here is, that the movement of the cursor is handled before the movement of the MouseArea. I don't have any definite proof for this. Only my test with the solution below, suggests that.
Solution
The first challenge is to detect movement of the MouseArea. It might be that it moves, without its own x and y-values changing, e.g. if its parent is moving.
To solve this, I'll introduce two properties globalX and globalX. Then I use the trick from this answer on how to track a gobal position of an object.
Now I'll have two signals to handle: globalXChanged and globalYChanged.
According to my assumption, they are fired, after the mouseXChanged and mouseYChanged. I will use a flag isEntered to make sure, I only handle one of them, by setting it to true, if the first of them is fired.
I will use the cursor position on a globalMouseArea to determine, whether the cursor is within bounds of the MouseArea. This requires, the cursor is not in some other MouseArea at that time, or at least I know of it
-> With this I already succeeded in detecting the entrance.
The second challenge is to detect the exit. Here we have 4 cases to distinguish:
Cursor enters and leaves the MouseArea because of it's movement.
Cursor enters and leaves the MouseArea because of the movement of the MouseArea
Cursor enters because the MouseArea moves, and leaves, because the cursor moves
Cursor enters because it moves, and leaves as the MouseArea moves away.
The first would be easy to handle. After it enters we handle entered and when it leaves we handle exited. But after the fix, mentioned by Mitch we can't rely on this anymore.
So we will not set hoverEnabled: true and map the position of the cursor to the targetMouseArea whenever either the cursor moves, or the targetMouseArea moves, and act accordingly.
import QtQuick 2.7
import QtQuick.Controls 2.0
ApplicationWindow {
id: root
visible: true
width: 400; height: 450
MouseArea {
id: globalMouseArea
anchors.fill: parent
hoverEnabled: true
onClicked: ani.restart()
}
Rectangle {
x: 300
y: 300
width: 50
height: 50
color: 'green'
}
Rectangle {
id: rect
width: 50
height: 50
color: 'red'
Text {
text: targetMouseArea.isEntered.toString()
}
MouseArea {
id: targetMouseArea
anchors.fill: parent
signal enteredBySelfMovement
signal enteredByMouseMovement
onEnteredByMouseMovement: console.log('Cause: Mouse')
onEnteredBySelfMovement: console.log('Cause: Self')
property point globalPos: {
var c = Qt.point(0, 0)
var itm = this
for (; itm.parent !== null; itm = itm.parent) {
c.x += itm.x
c.y += itm.y
}
return c
}
property bool isEntered: false
function checkCollision(sig) {
if ((globalPos.y < globalMouseArea.mouseY)
&& (globalPos.y + height > globalMouseArea.mouseY)
&& (globalPos.x < globalMouseArea.mouseX)
&& (globalPos.x + width > globalMouseArea.mouseX)) {
if (!isEntered) {
isEntered = true
sig()
}
}
else if (isEntered && !containsMouse) {
console.log(isEntered = false)
}
}
onGlobalPosChanged: {
checkCollision(enteredBySelfMovement)
}
Connections {
target: globalMouseArea
onPositionChanged: {
targetMouseArea.checkCollision(targetMouseArea.enteredByMouseMovement)
}
}
}
}
NumberAnimation {
id: ani
target: rect
properties: 'x,y'
from: 0
to: 300
running: true
duration: 10000
}
}
Problems left: When we clicked within the targetMouseArea, as long as a button is pressed, we won't detect the leaving.
You can check whether mouseX, mouseY were changed since last event.
property int previousMouseX = mouseX; // or use other values to init
property int previousMouseY = mouseY; // e.g., 0, x, parent.x,
// or set it from extern
onEntered() {
if (mouseX != previousMouseX || mouseY != previousMouseY) {
// TODO do something
previousMouseX = mouseX;
previousMouseY = mouseY;
}
}
In case mouseX, mouseY are relative to the mouse area 0,0 you can use mapFromItem(null, 0, 0) to get the absolute values.