I'm using OpenSceneGraph and Qt to develop a simulation software. Simulations can involve multiple bots in the same virtual world. My requirements for views are as follows:
Ability to show static world cameras
Ability to show bot cameras
Ability to move these views around and/or stack them
To accomplish these tasks, I have made an 'OSGWidget' that uses an Osg Viewer to render the scene inside a QGLWidget. This OSGWidget is in turn put inside a QDockWidget that can be moved around and/or stacked, fulfilling that requirement.
The problem is that when using multiple singular viewers to render the same scene in different widgets, I get strange render behavior. Namely, textures do not display properly or sometimes even at all.
I have looked around SO and the OSG forums and while people have had similar problems, the only responses I have seen have suggested switching to an Osg CompositeViewer. For my purposes, I would like to avoid using that as it breaks my desired requirement of movable and stackable widgets rendering the same scene.
Is this an intractable situation that just isn't easily handled by Osg? I have seen several posts that say this is not how OSG is 'supposed to work' but they haven't really provided facts to support that claim. Has anyone done something similar or have any ideas/insight? I can provide code snippets if needed, but as this might just be a contradiction to Osg's ideology I will wait to get some responses.
Thanks to some help from the comments and from a couple other sites, I was able to successfully get the behavior I wanted from CompositeViewer.
Basically, all OSG Views go through my "WidgetDriver" which contains a CompositeViewer.
class OsgWidgetDriver {
public:
void init() {
compositeViewer = new osgViewer::CompositeViewer;
compositeViewer->setThreadingModel(osgViewer::Viewer::SingleThreaded);
compositeViewer->setReleaseContextAtEndOfFrameHint(false);
}
void start() {
initialized = true;
}
void stop() {
compositeViewer->stopThreading();
compositeViewer->setDone(true);
}
void updateFrame() {
if (initialized)
compositeViewer->frame();
}
void addView(osgViewer::View *view) { compositeViewer->addView(view); }
bool isInitialized() { return initialized; }
protected:
bool initialized;
osgViewer::CompositeViewer *compositeViewer;
};
Then, whenever I make a new Qt OSG Widget, I hand the osg::View off to the driver. I let the driver update the render window, while Qt can still update the QWidget accordingly. It even allows me to place the widget in a QDockWidget so I can move them around and stack them as needed.
Some final notes on the process if anyone else wants to do this:
You will very likely run into weird texture display problems when multiple OSG Viewers are looking at one scene. If that happens, use the osgUtil::Optimizer::TextureVisitor to set all textures to "UnrefImageAfterApply = false". This will allow proper texture displays across multiple osg::View instances.
The CompositeViewer options I set above, threading model to SingleThreaded and releaseContextAtEndOfFrame to false, are necessary if you want a single thread acting upon multiple views. If you use a widget driver like I did, you will want to do this.
Related
I have an app with four main pages, switched through a tab bar (no "back" button).
One page has a lot of content (ScrollView) and takes quite a few seconds until it's rendered. I handle that by showing a "loading" overlay while the work is done. But for that specific page I'd like to keep the view alive, so that when the user switches to another page and comes back later, the page is ready without loading everything again.
I'm not sure how to do that in MvvmCross, though.
I did read the documentation and from what I understood the View Presenter would be the right way to do it, since the docs say:
"Another kind of presentation changes your app can request through
hints includes clearing / modifying the BackStack, changing a root
while maintaining the existent views, … possibilities are really
endless. Once again your app is king here!"
I guess I would need to create a custom MvxPresentationHint for that, but I don't quite get it :(
How or rather where would I access and store/load the View?
I'm generally still quite unfamiliar with MvvmCross (how it works under the hood) and especially customization of Mvx classes, even though I've been using it for a while.
Any explanation and preferably code examples beyond what's written in the documentation would be extremely appreciated!
It isn't meaningful to attempt to "store" a view in MVVM. The XF view is a representation of what will be created with native (e.g. "Android" or "iOS") widgets. Creating and measuring/laying out those native widgets is what is slow. MVVM View Presenter won't speed up that logic.
Instead of "store", you need "keep alive":
For a ContentPage called MyPage, when you create it, store it in a static variable. Then re-use that variable. If you never need more than one of these, you can store it in the class itself.
Modify the "code behind", MyPage.xaml.cs:
public partial class MyPage : ContentPage
{
// Singleton Pattern.
private static MyPage _it;
public static MyPage It {
get {
if (_it == null)
_it = new MyPage();
return _it;
}
}
// "private", because calling this directly defeats the purpose. Instead, use `MyPage.It`.
private MyPage()
{
InitializeComponent();
}
}
To create it, whereever you would put:
new MyPage()
instead put this:
MyPage.It
For instance, you might do PushAsync(MyPage.It);
This will always return the SAME INSTANCE of MyPage. So once it has been created, it keeps its state.
IMPORTANT: Note that the constructor is only called ONCE. Any code that needs to be done each time the page appears, put in override .. OnAppearing() method.
LIMITATION: Views "expect" to be part of the visual hierarchy when they are manipulated. If you attempt to alter the page or its view model while it is not on the screen, you may encounter problems. Those are beyond the scope of this answer - create a new StackOverflow question with the details of any problem you encounter.
I am currently working on somebody else code and I need to fix a bug linked with dynamic translation.
When the language is changed, the Loader is reloaded, it works but it generates unwanted effects (including the bug mentioned above).
So I tried to look for a way to dynamically change the translation without reloading everything.
I added m_engine->retranslate() in my switchLanguage function and this works perfectly, but only for texts directly defined in QML files. The thing is there is also a lot of text defined with setContextProperty in the C++ main controller class, and for them, it doesn't work at all (which seems pretty normal since m_engine is a QQmlApplicationEngine).
I don't see how I can simply force these texts to retranslate too. I have them in pretty much every controller function and they are used by different QML files. I am afraid that there will be no other choice but to change completely the way translation is managed. I hope advanced programmers can help me with this.
Other information:
I work with 5.13.0 version of Qt.
I don't use Designer and cannot use ui.retranslateUi().
It's hard to tell how your main controller class looks like, so here is a short general answer.
You can install an eventFilter and listen for LanguageChange.
In constructor of "main controller class", add this:
auto *core = QCoreApplication::instance();
if(core != nullptr)
{
core->installEventFilter(this);
}
Then add a function to your class:
bool MainControllerClass::eventFilter(QObject *watched, QEvent *event)
{
Q_UNUSED(watched);
if(event->type() == QEvent::LanguageChange)
{
//set properties again or emit property changed signals
}
}
I am currently developing a video player.
The GUI as the topmost layer is written in QML. It should be transparent to lower layers. It contains control elements, some Lists etc., It's displayed using a QDeclarativeView.
Description
QDeclarativeView *upperLayer = new QDeclarativeView(this);
upperLayer->setSource(QUrl("/home/projects/QtVideo/qml/videoControl.qml"));
upperLayer->setStyleSheet(QString("background: transparent");
upperLayer->setResizeMode(QDeclarativeView::SizeRootObjectToView);
uperLayer->showFullScreen();
The layer underneath is a QWidget: I use the libvlc to display the video content
in this widget.
Reason: I am receiving MPEG-TS, which can not be decoded by phonon, afaik. Therefore I need the libvlc to decode the incoming *.ts stream and put the output onto the display.
QWidget *lowerLayer = new QWidget(this);
lowerLayer.setGeometry(QString("background: red"));
QUrl* url = new QUrl("file:///home/projects/QtVideo/video.ts");
libvlc_instance_t*vlcObject;
libvlc_media_t*vlcMedia;
libvlc_media_player_t*vlcPlayer;
vlcPlayer = NULL;
if(vlcObject = libvlc_new(argc, argv)) == NULL)
{
printf("Not able to initialize";
exit(1);
}
if(vlcPlayer && libvlc_media_player_is_playing(vlcPlayer))
{
libvlc_media_player_stop(vlcPlayer);
}
vlcPlayer = libvlc_media_player_new(vlcObject);
vlcMedia = libvlc_media_new_location(vlcObject, url.toString().toUtf8().constData());
libvlc_media_player_set_media(vlcPlayer, vlcMedia);
#if defined(Q_OS_MAC)
libvlc_media_player_set_nsobject(vlcPlayer, lowerLayer->winId());
#elif defined(Q_OS_UNIX)
libvlc_media_player_set_x_window(vlcPlayer, lowerLayer->winId());
#elif defined(Q_OS_WIN)
libvlc_media_player_set_hwnd(vlcPlayer, lowerLayer->winId());
#endif
libvlc_media_player_play(vlc_player);
Both Elements, the QDeclarativeView and the QWidget
are embedded in a QMainWindow, lowerLayer created before the upperLayer,
upperLayer Transparent to the lowerLayer.
The Problem:
As long as the lowerLayer is displaying static elements such as a picture, or some colored shapes, everything works fine, full transparency and functionality.
As soon as I start displaying a video, such as the described *.ts using the libvlc OR some random video using the Phonon::VideoPlayer, the parts of the upperLayer which are above the video parts of the lowerLayer are displayed in the color of the lowerLayer(default: gray), the parts of the upperLayer which are positioned above parts of the lowerLayer or others which do not contain video elements are displayed in correct behaviour.
Question:
Is there any posibility and if, how, to get the upperLayer transparent, even if there is a video playing?
Are you still fighting with this issue? I, sadly, do not have a satisfying answer for you. The best I can do is point you to reasons why it doesn't work:
http://lists.trolltech.com/qt-interest/2007-02/thread01061-0.html
See Message #4 in the link above.
I have tried many different methods to get transparent painting over a video (specifically Phonon::VideoPlayer) using Qt. The only method I've found so far, is to set the overlaying QWidget as a toolTip doing something like
pWidget->setWindowFlags(Qt::ToolTip)
Depending on what exactly you're wanting to do this may be sufficient, but (in my opinion) it's a hack at best. I'm actively struggling with this issue and if I can find some sort of solution, I'll be sure to post it here.
Best of luck.
you're using direct rendering (by passing the wid of the widget) which draws the video overtop at that geometry:
libvlc_media_player_set_x_window
you need to use offscreen rendering and draw that to your qwidget. this can be done with an opengl context (complicated) or using the callback methods available in libvlc.
if you use the display callback (libvlc_video_display_cb) libvlc will also generate lock/unlock methods also, if you need. in this method libvlc will expect some parameters to be set such as canvas geometry and pixel format.
that said, phonon has a libvlc backend which might help, but may still use direct rendering depending on some factors..
My application consists of a WebView widget. A mouse click on the widget is not handled by the mousePressEvent() of my application, but by the WebView widget. So, I installed an event filter to receive the events. Now, I get notified of all events, except the mouseReleaseEvent for the right click (Everything works fine for left clicks and mousePressEvent for the right click is also getting registered). I guess it has got something to do with context events getting generated by right clicks (a pop-up menu gets generated). But since I am using a filter, the event should first be sent to me. The following is the code for the event filter
in Jambi, but I hope I can modify an answer given in Qt for Jambi.
public boolean eventFilter(QObject o,QEvent event)
{
if (event.type()==QEvent.Type.MouseButtonPress) // One can call the mousePressEvent() functions from here,which can do this work but speed
{
if (((QMouseEvent)event).button()==Qt.MouseButton.LeftButton)
{
mousebuttontype=1;
clickedorpressed=1;
}
else
if (((QMouseEvent)event).button()==Qt.MouseButton.RightButton)
{
mousebuttontype=2;
System.out.println("right");
}
t1=QTime.currentTime();
t1.start();
}
else
if (event.type()==QEvent.Type.MouseButtonRelease)
{
if (t1.elapsed()>900)
{
switch(mousebuttontype)
{
case 1: browser.back();
break;
case 2: browser.forward();
break;
}
}System.out.println("choda");
}
return false;
}
MY BASIC AIM IS GETTING THE TIME INTERVAL FOR WHICH THE RIGHT CLICK WAS PRESSED. ANY WORKAROUND ?
Some digging around certainly seems to suggest that there may be no right-mouse release event being generated if that is the trigger for a context menu on your particular system. These are conveyed as a QContextMenuEvent.
This Qt Labs post hints about this too.
A work around may be tricky if it is system dependent. Have you tried overriding event() in the QApplication class to see if the event comes through there? Some events don't necessarily get to the children but everything goes through QApplication::event().
Another option for you is, subclass Qwebview and override mouseReleaseEvent event
I've found a fix for this that I dont think will be system dependent.
My particular case was using mouseReleaseEvent in order to catch the events, and use them myself. I didn't get these events.
On all child widgets of the widget that I want to handle the event, I added to the class definition:
protected:
void mouseReleaseEvent(QMouseEvent *event) {event->ignore();}
This overrides the default implementation for context menu's, and sends the mouseReleaseEvent back up the parent chain like it would for other mousebuttons.
http://doc.qt.io/qt-5/qevent.html#ignore
This shows that it will probably propagate to the parent widget. As the link indicates, this fixed it for me at Qt 5.9, but I think it should work for virtually all version.
(I am aware this question is literally 7 years old, but it doesn't contain the fix that I think would be the best fix, and shows up as result 2 on google(qt not getting mouse release event on rightclick). So I think it deserves an up to date answer.)
Without going in to broader implementation explanations, the core problem I needed to solve related to this issue was that I needed to know if a context menu swallowed the right-click so I could ensure some custom state was handled properly. The simplest workaround I was able to find was to implement contextMenuEvent, which gets called after mousePressEvent, to detect if the event was accepted (a context menu was opened somewhere in my QGraphicsScene/Items). Here is a minimal example in Python to demonstrate how the event state might be used:
def contextMenuEvent(self, event):
event.setAccepted(False)
super(MyGraphicsView, self).contextMenuEvent(event)
self.__context_menu_used = event.isAccepted()
Note you can also skip running the the base class contextMenuEvent entirely to block the scene/items from opening a context menu. This is handy if you want to do 3D-DCC-like alt-RMB-zooming without accidentally opening a context menu, e.g.:
def contextMenuEvent(self, event):
# Block context menus if alt is held down
if event.modifiers() & QtCore.Qt.AltModifier:
return
event.setAccepted(False)
super(MyGraphicsView, self).contextMenuEvent(event)
self.__context_menu_used = event.isAccepted()
How can I in code of the custom Qt widget know that it is currently instantiated in Qt designer?
Use case:
I build a complex custom widget that has several child widgets like QPushButton, QLabel etc.
As application logic require, when widget is created most of those sub component are not visible but in design time when I put it on a form I would like to see them.
To be able to play with style sheet at design time.
Currently what I get is a empty is only a result of constructor - minimal view (actually empty in my case).
What I am looking for is to be able to do something like
MyQWidget::(QWidget *parent)
{
....
if(isRunningInDesigner())
{
myChildWidget1->setVisible(true);
myChildWidget2->setVisible(true);
myChildWidget3->setVisible(true);
}
else
{
myChildWidget1->setVisible(false);
myChildWidget2->setVisible(false);
myChildWidget3->setVisible(false);
}
....
}
So what should I put in to this bool isRunningInDesigner() ?
From the Qt Designer manual:
To give custom widgets special behavior in Qt Designer, provide an implementation of the initialize() function to configure the widget construction process for Qt Designer specific behavior. This function will be called for the first time before any calls to createWidget() and could perhaps set an internal flag that can be tested later when Qt Designer calls the plugin’s createWidget() function.
Those are methods from the QDesignerCustomWidgetInterface plugin interface. In short: you tell the widget to behave differently when Qt Designer asks your plugin to create instances of your custom widget.