I would like to get notification in main Qt app in case any control in QML (loaded via QQuickWidget) changes its value. There are CheckBox'es, ComboBox'es, SpinBox'es and TextEditor's.
My current approach is to declare a slot for every control (onCheckedChanged, onCurrentIndexChanged, onValueChanged and onTextChanged respectively) and call mainApp.notifyMe() from them, there mainApp is a link to parent Qt Widget propagated to QML with help of
QQmlEngine* engine = new QQmlEngine(this);
if (QQmlContext* cntx = engine->rootContext()) {
cntx->setContextProperty("mainApp", this);
}
and notifyMe() is a slot in it on C++ side.
But this requires dozens of functions with the same code bcs I have dozens of controls. It would be ideal to have one or 4 functions with notifyMe() in QML which could be connected to all controls value changes.
Is there a way in QML to create such slot and connect it to multiple objects property changes?
I've end up with following in my root item:
MySettingsForm {
signal notify()
onNotify: {
mainApp.notifyMe();
}
Component.onCompleted: {
var cbs = [Combobox1, Combobox2, Combobox3];
for (var i in cbs) {
cbs[i].checkedChanged.connect(notify);
}
var sbs = [SpinBox1, SpinBox2, SpinBox3];
for (i in sbs) {
sbs[i].valueChanged.connect(notify);
}
var tes = [TextField1, TextFiel2, TextField3];
for (i in tes) {
tes[i].textChanged.connect(notify);
}
var cxs = [ComboBox1, ComboBox2, ComboBox3];
for (i in cxs) {
cxs[i].currentIndexChanged.connect(notify);
}
}
...
}
The trick is in creation of custom signal notify and slot attached to it.
Related
I am creating some automated GUI tests in my application using QTest.
I can access the widgets from my application using the command:
savePushButton = mainWindow->findChild<QPushButton *>("savePushButton");
It is working fine, but now I have to click on the OK button of a QMessageBox.
I created the QMessageBox in my application like this:
if( something_wrong )
{
QMessageBox::warning(new Widget(), "Title", "Something wrong!");
}
How can I have access to this QMessageBox, and its buttons?
I found a solution on the following link: http://www.qtcentre.org/threads/31239-Testing-modal-dialogs-with-QTestLib .
It uses the command QApplication::topLevelWidgets(); to get a widget list. Then it searches for the message box widget and simulates a key enter (QTest::keyClick(mb, Qt::Key_Enter);) which closes the message box.
Example:
void MyTest::testDialog()
{
QTimer::singleShot(500, this, SLOT(timeOut()));
QVERIFY(functionThatProducesMessageBox());
}
void MyTest::timeOut()
{
QWidgetList allToplevelWidgets = QApplication::topLevelWidgets();
foreach (QWidget *w, allToplevelWidgets) {
if (w->inherits("QMessageBox")) {
QMessageBox *mb = qobject_cast<QMessageBox *>(w);
QTest::keyClick(mb, Qt::Key_Enter);
}
}
}
The header file must contain the Q_OBJECT macro to use the signals and slots mechanism.
Example:
class MyClass: public QWidget
{
Q_OBJECT
public:
...
It worked well for me since the UI (thread) is blocked when the message box appears.
Note: remember to rebuild the project when you add the Q_OBJECT macro.
It often helps to look to Qt's auto tests:
void ExecCloseHelper::timerEvent(QTimerEvent *te)
{
if (te->timerId() != m_timerId)
return;
QWidget *modalWidget = QApplication::activeModalWidget();
if (!m_testCandidate && modalWidget)
m_testCandidate = modalWidget;
if (m_testCandidate && m_testCandidate == modalWidget) {
if (m_key == CloseWindow) {
m_testCandidate->close();
} else {
QKeyEvent *ke = new QKeyEvent(QEvent::KeyPress, m_key, Qt::NoModifier);
QCoreApplication::postEvent(m_testCandidate, ke);
}
m_testCandidate = Q_NULLPTR;
killTimer(m_timerId);
m_timerId = m_key = 0;
}
}
Judging from that code, you can get the message box via QApplication::activeModalWidget(). Testing native (I'm assuming they're native) widgets is difficult, which is likely why they chose to send key events, as you don't need to know e.g. the location of the buttons for those, as you would with a mouse click.
In QML it is impossible to call .disconnect() without arguments for a signal:
file:mainwindow.qml:107: Error: Function.prototype.disconnect: no arguments given
So how can I disconnect ALL slots without specifying each of them?
Or maybe it is possible by passing signal object to C++ and disconnect it somehow there?
Or maybe any workaround exists?
The goal I want to reach is to change behavior of an object by connecting different slots to it's signal. For example:
object.disconnect() // disconnect all slots
object.connect(one_super_slot)
object.disconnect() // disconnect all slots
object.connect(another_super_slot)
No. I looked at the source code in qv4objectwrapper.cpp, and you can see this code:
void QObjectWrapper::initializeBindings(ExecutionEngine *engine)
{
engine->functionClass->prototype->defineDefaultProperty(QStringLiteral("connect"), method_connect);
engine->functionClass->prototype->defineDefaultProperty(QStringLiteral("disconnect"), method_disconnect);
}
Those are the only two methods that are added. If you look at the source code for method_disconnect() you can see that it always requires one or two parameters, including the name of the slot to disconnect.
There is no disconnectAll() unfortunately.
Okay, 5 minutes after my question I've made a workaround: connect only once to one signal that calls jsobject from inside:
Item {
property var fire
// Any qml object. In this example it is ActionExecutor which emits actionRequest
ActionExecutor {
//signal actionRequest(int actionType)
onActionRequest: fire(actionType)
}
Action {
shortcut: "Ctrl+S"
text: "One action"
onTriggered: {
parent.fire = function(actionType) {
console.log('one slot')
}
}
}
Action {
shortcut: "Ctrl+X"
text: "Another action"
onTriggered: {
parent.fire = function(actionType) {
console.log('Another slot')
}
}
}
}
So that js object can be reassigned many times as you want so you may change your behavior by reassigning this object. If you want to disconnect all simple assign undefined to fire. Also you can make a chain of "slots" by modifying code to something like:
Item {
property var fire
property var slots: [
function(actionType) {
console.log('1: ' + actionType)
},
function() {
console.log('2: ' + actionType)
},
function() {
console.log('3: ' + actionType)
}
]
// Any qml object. In this example it is ActionExecutor which emits actionRequest
ActionExecutor {
//signal actionRequest(int actionType)
onActionRequest: fire(actionType)
}
Action {
shortcut: "Ctrl+S"
text: "One action"
onTriggered: {
parent.fire = function(actionType) {
console.log('calling all custom JS-slots')
for (var i in slots) {
slots[i](actionType)
}
}
}
}
}
So anyone can implement own signal-slot architecture in qml as a simple javascript observer pattern.
Enjoy.
I'm trying to use QDBusPendingCallWatcher to watch an async call. Some sample code like this:
{
// interface = new QDBusInterface(...);
QDBusPendingCall pcall = interface->asyncCall("query");
QDBusPendingCallWatcher *watcher = new QDBusPendingCallWatcher(pcall, this);
QObject::connect(watcher, SIGNAL(finished(QDBusPendingCallWatcher*)), this, SLOT(handler(QDBusPendingCallWatcher*)));
}
and handler function:
void Client::handler(QDBusPendingCallWatcher* call)
{
QDBusPendingReply<QString> reply = *call;
// do something
}
My questions are:
It looks like QDBusPendingCallWatcher uses shared data pointer inside, is it safe to not manually delete the watcher pointer? Just leave the scope and forget it?
If I can let the smart pointer of pendingcall to do all the tricks, can I use just one QDBusPendingCallWatcher pointer in my class to watch all the async calls? Like this:
{
QDBusPendingCall pcall = interface->asyncCall("query");
watcher = new QDBusPendingCallWatcher(pcall, this);
QObject::connect(watcher, SIGNAL(finished(QDBusPendingCallWatcher*)), this, SLOT(handleOne(QDBusPendingCallWatcher*)));
pcall = interface->asyncCall("anotherQuery");
watcher = new QDBusPendingCallWatcher(pcall, this);
QObject::connect(watcher, SIGNAL(finished(QDBusPendingCallWatcher*)), this, SLOT(handleTwo(QDBusPendingCallWatcher*)));
}
Will this makes a disaster? Or should I use multiple pointers for each call?
Thanks!
Take a closer look at the QDBusPendingCallWatcher documentation:
The slot connected to by the above code could be something similar to the following:
void MyClass::callFinishedSlot(QDBusPendingCallWatcher *call)
{
QDBusPendingReply<QString, QByteArray> reply = *call;
if (reply.isError()) {
showError();
} else {
QString text = reply.argumentAt<0>();
QByteArray data = reply.argumentAt<1>();
showReply(text, data);
}
call->deleteLater();
}
The call of QObject::deleteLater is the key: This means Qt will delete the Object as soon as execution returns to the event loop.
As long as you call deleteLater inside Client::handler(...), you don't need to - more precisely you musn't - call delete watcher; anywhere. The only thing you have to ensure is that noone uses the object behind call after the slot returns.
When I instantiate a new StageVideo instsance with stage.stageVideos[0] everything works great, but when i leave my view that's displaying the video it sticks on the stage. So when i goto another view it's still showing on the stage in the background. I tried setting sv = null, removing event listeners...etc.
I've created a StageVideoDisplay class which is instantiated in mxml like: and on view initialization i call a play() method:
if ( _path )
{
//...
// Connections
_nc = new NetConnection();
_nc.connect(null);
_ns = new NetStream(_nc);
_ns.addEventListener(NetStatusEvent.NET_STATUS, onNetStatus);
_ns.client = this;
// Screen
_video = new Video();
_video.smoothing = true;
// Video Events
// the StageVideoEvent.STAGE_VIDEO_STATE informs you whether
// StageVideo is available
stage.addEventListener(StageVideoAvailabilityEvent.STAGE_VIDEO_AVAILABILITY,
onStageVideoState);
// in case of fallback to Video, listen to the VideoEvent.RENDER_STATE
// event to handle resize properly and know about the acceleration mode running
_video.addEventListener(VideoEvent.RENDER_STATE, videoStateChange);
//...
}
On video stage event:
if ( stageVideoInUse ) {
try {
_rc = new Rectangle(0,0,this.width,this.height);
_sv.viewPort = _rc;
} catch (e:Error) {}
} else {
try {
_video.width = this.width;
_video.height = this.height;
} catch (e:Error) {}
}
And on stage video availability event:
protected function toggleStageVideo(on:Boolean):void
{
// To choose StageVideo attach the NetStream to StageVideo
if (on)
{
stageVideoInUse = true;
if ( _sv == null )
{
try {
_sv = stage.stageVideos[0];
_sv.addEventListener(StageVideoEvent.RENDER_STATE, stageVideoStateChange);
_sv.attachNetStream(_ns);
_sv.depth = 1;
} catch (e:Error) {}
}
if (classicVideoInUse)
{
// If you use StageVideo, remove from the display list the
// Video object to avoid covering the StageVideo object
// (which is always in the background)
stage.removeChild ( _video );
classicVideoInUse = false;
}
} else
{
// Otherwise attach it to a Video object
if (stageVideoInUse)
stageVideoInUse = false;
classicVideoInUse = true;
try {
_video.attachNetStream(_ns);
stage.addChildAt(_video, 0);
} catch (e:Error) {}
}
if ( !played )
{
played = true;
_ns.play(path);
}
}
What happens is in the view when i navigator.popView(), the stageVideo remains on the stage, even in other views, and when returning to that view to play a different stream the same stream is kind of "burned" on top. I can not figure out a way to get rid of it! Thanks in advance!
In Flash Player 11, Adobe added the dispose() method to the NetStream class.
This is useful to clear the Video or StageVideo object when you're done with it.
When you call the dispose() method at runtime, you may get an exception indicating that there is no property named dispose on the NetStream object.
This occurs because Flash Builder is not compiling the app with the proper SWF version. To fix that, just add this to your compiler directives:
-swf-version=13
In the new Flash Builder 4.7, we hopefully won't have to specify the SWF version to use the newer Flash Player features.
This seems to be the best solution, but if you can't use Flash 11 (or the latest Adobe AIR), some other work arounds would be:
set the viewPort of stage video object to a rectangle that has width=0 and height=0
since stage video appears underneath your app, you can always cover the viewport (with a background color or some DisplayObject)
Ok the issue was actually that, even though it seemed like stage video was in use since i got the "Accelerated" message in status, that it was actually the video render even running and classic video was actually in use. The only thing I needed to do was add stage.removeChild( _video ) to the close() event in the class!!
I try to load to plugins. If they are loaded with success , then i must connect returned widgets.
With one plugin i create an action and add it to a menu , with another plugin i create a label and add it to window. Even if i get this error during runtime (when app loads plugins) , these two widgets are created and are visible. But there is no connection between them.
This is how i try to connect widgets
QObject *plugin = pluginLoader.instance();
if (plugin) {
myAction = qobject_cast<ActionInterface *>(plugin);
if (myAction) {
pluginMenu->addAction(myAction->newAction());
verify ++;
}
myLabel = qobject_cast<LabelInterface *>(plugin);
if (myLabel) {
layout->addWidget(myLabel->newLabel());
verify++;
}
if (verify == 2)
connect(myAction, SIGNAL(pushMyAction()),
myLabel, SLOT(setTextforLabel()));
}
...
}
Error message is :
QObject::connect: Cannot connect (null)::pushMyAction() to LabelPlugin::setTextforLabel()
You have two different plugins. Apparently one can be cast to an ActionInterface but not a LabelInterface, and the other can be cast to a LabelInterface but not an ActionInterface.
Your idea here seems to be that once you have both plugins loaded (and a verify count of 2) then it's safe to make a connect call between these plugins. However you appear to be trying to cast the second loaded plugin to serve as both the signal and the slot. This is because each time you run the code you overwrite both myAction and myLabel. So at minimum:
QObject* plugin = pluginLoader.instance();
if (plugin) {
ActionInterface* myActionTemp = qobject_cast<ActionInterface*>(plugin);
if (myActionTemp) {
myAction = myActionTemp;
pluginMenu->addAction(myAction->newAction());
verify++;
}
LabelInterface* myLabelTemp = qobject_cast<LabelInterface*>(plugin);
if (myLabelTemp) {
myLabel = myLabelTemp;
layout->addWidget(myLabel->newLabel());
verify++;
}
/* if (myAction and myLabel) would be less convoluted... */
if (verify == 2) {
connect(myAction, SIGNAL(pushMyAction()),
myLabel, SLOT(setTextforLabel()));
}
...
}
Still, this looks like a fairly brittle design that could use some rethinking...!
QObject *plugin = pluginLoader.instance();
if (plugin) {
if (plugin->inherits("ActionInterface")) {
myAction = qobject_cast<ActionInterface *>(plugin);
pluginMenu->addAction(myAction->newAction());
}
if (plugin->inherits("LabelInterface")) {
myLabel = qobject_cast<LabelInterface *>(plugin);
layout->addWidget(myLabel->newLabel());
}
if (myLabel && myAction)
connect(myAction, SIGNAL(pushMyAction()),
myLabel, SLOT(setTextforLabel()));
}
...
}
Remove this "ugly" verify counter. Remember to initialize myLabel and myAction with NULL