Translate screen drawing to another part of the screen - c++

Is it possible to make it so that all drawing to an area "A" is translated to an area "B"?
For example drawing to the area(0,0)(100,100) and have it appear in area(200,200)(300,300).

The question is actually tagged with windows and graphics. This might have been targeted to Win32 and GDI (where I've unfortunately nearly no experience). So, the following might be seen as proof of concept:
I couldn't resist to implement the idea / concept using QWindow and QPixmap.
The concept is:
open a window fullscreen (i.e. without decoration)
make a snapshot and store it internally (in my case a )
display the internal image in window (the user cannot notice the difference)
perform a loop where pixmap is modified and re-displayed periodically (depending or not depending on user input).
And this is how I did it in Qt:
I opened a QWindow and made it fullscreen. (Maximum size may make the window full screen as well but it still will have decoration (titlebar with system menu etc.) which is unintended.)
Before painting anything, a snapshot of this window is done. That's really easy in Qt using QScreen::grabWindow(). The grabbed contents is returned as QPixmap and stored as member of my derived Window class.
The visual output just paints the stored member QPixmap.
I used a QTimer to force periodical changes of the QPixmap. To keep the sample code as short as possible, I didn't make the effort of shuffling tiles. Instead, I simply scrolled the pixmap copying a small part, moving the rest upwards, and inserting the small stripe at bottom again.
The sample code qWindowRoll.cc:
#include <QtWidgets>
class Window: public QWindow {
private:
// the Qt backing store for window
QBackingStore _qBackStore;
// background pixmap
QPixmap _qPixmap;
public:
// constructor.
Window():
QWindow(),
_qBackStore(this)
{
showFullScreen();
}
// destructor.
virtual ~Window() = default;
// disabled:
Window(const Window&) = delete;
Window& operator=(const Window&) = delete;
// do something with pixmap
void changePixmap()
{
enum { n = 4 };
if (_qPixmap.height() < n) return; // not yet initialized
const QPixmap qPixmapTmp = _qPixmap.copy(0, 0, _qPixmap.width(), n);
//_qPixmap.scroll(0, -n, 0, n, _qPixmap.width(), _qPixmap.height() - n);
{ QPainter qPainter(&_qPixmap);
qPainter.drawPixmap(
QRect(0, 0, _qPixmap.width(), _qPixmap.height() - n),
_qPixmap,
QRect(0, n, _qPixmap.width(), _qPixmap.height() - n));
qPainter.drawPixmap(0, _qPixmap.height() - n, qPixmapTmp);
}
requestUpdate();
}
protected: // overloaded events
virtual bool event(QEvent *pQEvent) override
{
if (pQEvent->type() == QEvent::UpdateRequest) {
paint();
return true;
}
return QWindow::event(pQEvent);
}
virtual void resizeEvent(QResizeEvent *pQEvent)
{
_qBackStore.resize(pQEvent->size());
paint();
}
virtual void exposeEvent(QExposeEvent*) override
{
paint();
}
// shoot screen
// inspired by http://doc.qt.io/qt-5/qtwidgets-desktop-screenshot-screenshot-cpp.html
void makeScreenShot()
{
if (QScreen *pQScr = screen()) {
_qPixmap = pQScr->grabWindow(winId());
}
}
private: // internal stuff
// paint
void paint()
{
if (!isExposed()) return;
QRect qRect(0, 0, width(), height());
if (_qPixmap.width() != width() || _qPixmap.height() != height()) {
makeScreenShot();
}
_qBackStore.beginPaint(qRect);
QPaintDevice *pQPaintDevice = _qBackStore.paintDevice();
QPainter qPainter(pQPaintDevice);
qPainter.drawPixmap(0, 0, _qPixmap);
_qBackStore.endPaint();
_qBackStore.flush(qRect);
}
};
int main(int argc, char **argv)
{
QApplication app(argc, argv);
// setup GUI
Window win;
win.setVisible(true);
// setup timer
QTimer qTimer;
qTimer.setInterval(50); // 50 ms -> 20 Hz (round about)
QObject::connect(&qTimer, &QTimer::timeout,
&win, &Window::changePixmap);
qTimer.start();
// run application
return app.exec();
}
I compiled and tested with Qt 5.9.2 on Windows 10. And this is how it looks:
Note: On my desktop, the scrolling is smooth. I manually made 4 snapshots and composed a GIF in GIMP – hence the image appears a bit stuttering.

Related

Changing the QPushButton region mask in its subclass to create a RoundButton

I am trying to create a round button by subclassing and setting the region mask so that I can reuse it in my project. I know we can override paintEvent method and draw a circle to show it as a round button. But the problem with this approach is that if user clicks outside the circle (but within button rect) it will be treated as a button click. This problem we don't see when set the region mask.
I tried to set the region by calling setmask method inside resizeEvent/paintEvent. In either of case, button will be blank. I am trying to figure out the place inside the subclass to set the region mask.
RoundAnimatingButton.h ->
#include <QPushButton>
namespace Ui {
class CRoundAnimatingBtn;
}
class CRoundAnimatingBtn : public QPushButton
{
Q_OBJECT
public:
explicit CRoundAnimatingBtn(QWidget *parent = nullptr);
~CRoundAnimatingBtn();
void StartAnimation(QColor r);
void StopAnimation();
public slots:
void timerEvent(QTimerEvent *e);
private:
Ui::CRoundAnimatingBtn *ui;
bool m_Spinning;
// QWidget interface
protected:
void resizeEvent(QResizeEvent *event) override;
void paintEvent(QPaintEvent * e) override;
};
#endif // ROUNDANIMATINGBTN_H
RoundAnimatingButton.cpp
CRoundAnimatingBtn::CRoundAnimatingBtn(QWidget *parent)
: QPushButton (parent)
, ui(new Ui::CRoundAnimatingBtn)
, m_Spinning(false)
{
ui->setupUi(this);
}
CRoundAnimatingBtn::~CRoundAnimatingBtn()
{
delete ui;
}
void CRoundAnimatingBtn::paintEvent(QPaintEvent *e)
{
QPushButton::paintEvent(e);
if(m_Spinning)
{
// Animating code
}
}
void CRoundAnimatingBtn::StartAnimation(QColor r)
{
m_Spinning=true;
startTimer(5);
}
void CRoundAnimatingBtn::StopAnimation()
{
m_Spinning=false;
this->update();
}
void CRoundAnimatingBtn::timerEvent(QTimerEvent *e)
{
if(m_Spinning)
this->update();
else
killTimer(e->timerId());
}
void CRoundAnimatingBtn::DrawRing()
{
}
void CRoundAnimatingBtn::resizeEvent(QResizeEvent *event)
{
// -----------------------------------
// This code didn't work
// -----------------------------------
QRect rect = this->geometry();
QRegion region(rect, QRegion::Ellipse);
qDebug() << "PaintEvent Reound button - " << region.boundingRect().size();
this->setMask(region);
// ----------------------------------
// ------------------------------------
// This code worked
// -------------------------------------
int side = qMin(width(), height());
QRegion maskedRegion(width() / 2 - side / 2, height() / 2 - side / 2, side,
side, QRegion::Ellipse);
setMask(maskedRegion);
}
Qt doc. provides a sample for “non-rectangular” widgets – Shaped Clock Example.
(Un-)Fortunately, I remembered this not before I got my own sample running.
I started in Qt doc. with
void QWidget::setMask(const QBitmap &bitmap)
Causes only the pixels of the widget for which bitmap has a corresponding 1 bit to be visible. If the region includes pixels outside the rect() of the widget, window system controls in that area may or may not be visible, depending on the platform.
Note that this effect can be slow if the region is particularly complex.
The following code shows how an image with an alpha channel can be used to generate a mask for a widget:
QLabel topLevelLabel;
QPixmap pixmap(":/images/tux.png");
topLevelLabel.setPixmap(pixmap);
topLevelLabel.setMask(pixmap.mask());
The label shown by this code is masked using the image it contains, giving the appearance that an irregularly-shaped image is being drawn directly onto the screen.
Masked widgets receive mouse events only on their visible portions.
See also mask(), clearMask(), windowOpacity(), and Shaped Clock Example.
(When reading this, I still missed the link to example.)
At first, I prepared a suitable pixmap for my purpose – dialog-error.png:
for which I converted an SVG from one of my applications.
I tried to apply it to a QPushButton as icon and as mask. This looked very strange. I'm not quite sure what exactly was the problem:
- using the resp. QPushButton as toplevel widget (i.e. main window)
- the fact that QPushButtons icon rendering and the mask may not match concerning position or size.
Without digging deeper, I changed the code and fixed both issues in next try:
making a derived button (like described by OP)
using the button as non-toplevel widget.
This worked soon. I added some code to make the effect more obvious:
a mouse press event handler for main window to show whether shape is considered correctly
a signal handler to show whether clicks on button (in shape) are received correctly.
So, I came to the following sample – testQPushButtonMask.cc:
#include <QtWidgets>
class MainWindow: public QWidget {
public:
explicit MainWindow(QWidget *pQParent = nullptr):
QWidget(pQParent)
{ }
virtual ~MainWindow() = default;
MainWindow(const MainWindow&) = delete;
MainWindow& operator=(const MainWindow&) = delete;
protected:
virtual void mousePressEvent(QMouseEvent *pQEvent) override;
};
void MainWindow::mousePressEvent(QMouseEvent *pQEvent)
{
qDebug() << "MainWindow::mousePressEvent:" << pQEvent->pos();
QWidget::mousePressEvent(pQEvent);
}
class RoundButton: public QPushButton {
private:
QPixmap _qPixmap;
public:
RoundButton(const QPixmap &qPixmap, QWidget *pQParent = nullptr):
QPushButton(pQParent),
_qPixmap(qPixmap)
{
setMask(_qPixmap.mask());
}
virtual ~RoundButton() = default;
RoundButton(const RoundButton&) = delete;
RoundButton& operator=(const RoundButton&) = delete;
virtual QSize sizeHint() const override;
protected:
virtual void paintEvent(QPaintEvent *pQEvent) override;
};
QSize RoundButton::sizeHint() const { return _qPixmap.size(); }
void RoundButton::paintEvent(QPaintEvent*)
{
QPainter qPainter(this);
const int xy = isDown() * -2;
qPainter.drawPixmap(xy, xy, _qPixmap);
}
int main(int argc, char **argv)
{
qDebug() << "Qt Version:" << QT_VERSION_STR;
QApplication app(argc, argv);
QPixmap qPixmap("./dialog-error.png");
// setup GUI
MainWindow qWin;
qWin.setWindowTitle(QString::fromUtf8("QPushButton with Mask"));
QVBoxLayout qVBox;
RoundButton qBtn(qPixmap);
qVBox.addWidget(&qBtn);
qWin.setLayout(&qVBox);
qWin.show();
// install signal handlers
QObject::connect(&qBtn, &RoundButton::clicked,
[](bool) { qDebug() << "RoundButton::clicked()"; });
// runtime loop
return app.exec();
}
The corresponding Qt project file testQPushButtonMask.pro
SOURCES = testQPushButtonMask.cc
QT += widgets
Compiled and tested on cygwin64:
$ qmake-qt5 testQPushButtonMask.pro
$ make && ./testQPushButtonMask
Qt Version: 5.9.4
MainWindow::mousePressEvent: QPoint(23,22)
MainWindow::mousePressEvent: QPoint(62,24)
MainWindow::mousePressEvent: QPoint(62,61)
MainWindow::mousePressEvent: QPoint(22,60)
RoundButton::clicked()
Concerning the output:
I clicked into the four corners of button.
I clicked on the center of button.

How do I reuse Qt's QTextCursor in my C++ application?

I am implementing a text editor on Windows using Qt in C++ and am using a QWidget as the surface on which I am displaying text using OpenGL. So far, I have my own layout engine and document model and am able to get the text to display on the widget.
Now I am trying to implement a text cursor that will be used similar to the one Qt provides, but QTextCursor is closely tied to its QTextDocument model and I am not able to subclass it for reuse using my model. Is there any way to reuse just the cursor without the model?
If not, how do I go about implementing a text cursor using Qt?
Note: I did go through the Caret methods that Windows provides here, but am hoping to avoid using them directly.
TL;DR: You can't. It's not a caret.
QTextCursor is an iterator, it has nothing to do whatsoever with the on-screen cursor. You can certainly reuse it if it is useful as an iterator, and iff your own text representation is built on top of QTextDocument. But it's not a caret.
The visible cursor control is provided by the QTextEdit implementation. Recall that QTextEdit is a view for a QTextDocument - it is entirely devoted to the graphical rendering of the text, controlling the visible cursor, etc.
There is no public Qt API that you can use for the caret. The WINAPI caret methods are completely useless when you're using Qt for rendering. You need your own caret implementation. Given that you already have a text representation and a renderer, you presumably have an iterator that works on the text representation, so implementing the caret should be a trivial affair.
The WINAPI caret is very simple and trivial to reimplement with the power of Qt:
class Caret : public QWidget {
Q_OBJECT
Q_PROPERTY(int period READ period WRITE setPeriod)
QPicture m_shape;
BasicTimer m_timer;
int m_period;
void updateSize() {
auto size = m_shape.boundingRect().size();
setFixedSize(size);
resize(size);
}
void timerEvent(QTimerEvent * ev) {
if (ev->timerId() != m_timer.timerId()) return;
if (isVisible()) hide(); else show();
}
public:
QPicture defaultShape(int w, int h) {
QPicture pic;
QPainter p(&pic);
p.fillRect(0, 0, w, h, Qt::black);
return pic;
}
Caret(QWidget * parent = 0, const QPicture & pic = defaultShape()) :
QWidget(parent), m_shape(pic), m_period(250) {
setAttribute(Qt::WA_TransparentForMouseEvents);
setAttribute(Qt::WA_TranslucentBackground);
m_timer.start(m_period);
updateSize();
}
void setShape(const QPicture & pic) {
m_shape = pic;
updateSize();
update();
}
void setPeriod(int period) {
if (period < 1) {
m_timer.stop();
if (m_period > 0) show();
} else
m_timer.start(period);
m_period = period;
}
int period() const { return m_period; }
void paintEvent(QPaintEvent *) Q_DECL_OVERRIDE {
auto painter = QPainter(this);
painter.fillRect(rect(), Qt::transparent);
painter.drawPicture(QPoint(), m_shape);
}
};

Programmatically invoke Snap/Aero maximize

Is there a way to programmatically invoke the Aera maximize effect using C or C++ for a specific window/window ID?
For example:
or
(source: thebuzzmedia.com)
I am using a border-less Qt window and Qt has an API for getting the window ID. I want to programmatically trigger the windows effects without the known triggers.
I don't want to talk about every single detail involved in achieving this effect, not only there's a lot that goes on but you also mentioned you understand the logic to place the windows at their specific locations. In this answer I'll address what I believe are the 2 main challenges:
How to receive and handle a maximize event?
How to create an approximation of the aero snap effect?
In order to answer the first question, we must analyze which event handlers are triggered when the window is maximized:
void resizeEvent(QResizeEvent* evt); // Invoked first,
void paintEvent(QPaintEvent* event); // then second,
void changeEvent(QEvent* evt); // and at last.
A Qt application is first notified of a resizeEvent(), which is followed by a paintEvent() to draw the window (or widget), and only after everything has been displayed, changeEvent() is invoked to let you know the widget was maximized (maybe it's a little bit late to receive such notification, I don't know).
Of all these, the only one we care about is resizeEvent(). This event handler informs the new window/widget size that can be used for comparison with the desktop size, thus allowing us to know if the event was actually a maximize request. Once we identify a maximize request, we can figure out whether the application should be maximized (and anchored) to right, left or to the center of the screen.
This would be the time to create the aero snap widget and place it on the screen as a visual clue to the user.
To answer the second question, I don't think is possible to call the native Windows API and ask it politely to perform this effect on your window. The only other logical choice is to write a code that approximates this effect ourselves.
The visual appearance can be replicated by drawing a transparent window with a shadow-ish border. The approach demonstrated in the source code below, creates and customizes a QWidget to make it behave and look like a aero snap window:
It's not the most beautiful thing in the world, I know. This demo creates a regular window for the user to interact with, and once it's maximized, it places itself to the left of the screen. To the right size of the screen it displays something that resembles an aero snap window (shown above).
The idea behind the aero snap widget is very simple: a QWidget with transparent background and a custom painting procedure. In other words, it's a transparent window which draws a rounded rectangle with a shadow and that's it.
To make it a bit more realistic, you should add some animation to resize the widget little by little. A for loop might do the trick, but if you need something fancy you'll end up using timers. If you take a look here, you can see the quickest & dirtiest method to perform animation with Qt in action, and better ways to deal with animation. However, for simple tasks like this, stick with frame-based animation.
main.cpp:
#include "window.h"
#include <QApplication>
int main(int argc, char* argv[])
{
QApplication app(argc, argv);
Window window;
window.show();
return app.exec();
}
window.h:
#pragma once
#include "snapwindow.h"
#include <QMainWindow>
#include <QEvent>
class Window : public QMainWindow
{
public:
Window();
void resizeEvent(QResizeEvent* evt);
//void paintEvent(QPaintEvent* event);
void changeEvent(QEvent* evt);
private:
SnapWindow* _sw;
};
window.cpp:
#include "window.h"
#include "snapwindow.h"
#include <QDebug>
#include <QWindowStateChangeEvent>
#include <QApplication>
#include <QDesktopWidget>
Window::Window()
{
setWindowTitle("AeroSnap");
resize(300, 300);
_sw = new SnapWindow(this);
_sw->hide();
}
void Window::changeEvent(QEvent* evt)
{
if (evt->type() == QEvent::WindowStateChange)
{
QWindowStateChangeEvent* event = static_cast<QWindowStateChangeEvent*>(evt);
if (event->oldState() == Qt::WindowNoState &&
windowState() == Qt::WindowMaximized)
{
qDebug() << "changeEvent: window is now maximized!";
}
}
}
// resizeEvent is triggered before window_maximized event
void Window::resizeEvent(QResizeEvent* evt)
{
qDebug() << "resizeEvent: request to resize window to: " << evt->size();
QSize desktop_sz = QApplication::desktop()->size();
//qDebug() << "resizeEvent: desktop sz " << desktop_sz.width() << "x" << desktop_sz.height();
// Apparently, the maximum size a window can have in my system (1920x1080)
// is actually 1920x990. I suspect this happens because the taskbar has 90px of height:
desktop_sz.setHeight(desktop_sz.height() - 90);
// If this not a request to maximize the window, don't do anything crazy.
if (desktop_sz.width() != evt->size().width() ||
desktop_sz.height() != evt->size().height())
return;
// Alright, now we known it's a maximize request:
qDebug() << "resizeEvent: maximize this window to the left";
// so we update the window geometry (i.e. size and position)
// to what we think it's appropriate: half width to the left
int new_width = evt->size().width();
int new_height = evt->size().height();
int x_offset = 10;
setGeometry(x_offset, 45, new_width/2, new_height-45); // y 45 and height -45 are due to the 90px problem
/* Draw aero snap widget */
_sw->setGeometry(new_width/2-x_offset, 0, new_width/2, new_height);
_sw->show();
// paintEvent() will be called automatically after this method ends,
// and will draw this window with the appropriate geometry.
}
snapwindow.h:
#pragma once
#include <QWidget>
class SnapWindow : public QWidget
{
public:
SnapWindow(QWidget* parent = 0);
void paintEvent(QPaintEvent *event);
};
snapwindow.cpp:
#include "snapwindow.h"
#include <QPainter>
#include <QGraphicsDropShadowEffect>
SnapWindow::SnapWindow(QWidget* parent)
: QWidget(parent)
{
// Set this widget as top-level (i.e. owned by user)
setParent(0);
/* Behold: the magic of creating transparent windows */
setWindowFlags(Qt::Widget | Qt::FramelessWindowHint);
setStyleSheet("background:transparent;");
setAttribute(Qt::WA_NoSystemBackground, true); // speed up drawing by removing unnecessary background initialization
setAttribute(Qt::WA_TranslucentBackground);
//setAutoFillBackground(true);
/* Use Qt tricks to paint stuff with shadows */
QGraphicsDropShadowEffect* effect = new QGraphicsDropShadowEffect();
effect->setBlurRadius(12);
effect->setOffset(0);
effect->setColor(QColor(0, 0, 0, 255));
setGraphicsEffect(effect);
}
void SnapWindow::paintEvent(QPaintEvent *event)
{
QWidget::paintEvent(event);
/* Lazy way of painting a shadow */
QPainter painter(this);
QPen pen(QColor(180, 180, 180, 200));
pen.setWidth(3);
painter.setPen(pen);
// Offset 6 and 9 pixels so the shadow shows up properly
painter.drawRoundedRect(QRect(6, 6, (width()-1)-9, (height()-1)-9), 18, 18);
}
This is just a quick demo to point you to the right direction. It is by no means a complete implementation of the effect you are looking for.
Maybe it is not what you need, but this effect is just resizing and moving window then try use Qt methods to do this.
bool left = false;
QSize size = QApplication::desktop()->size();//resolution of current screen
if(left)
{//left side
this->setGeometry(0, 0, size.width()/2, size.height());//(maybe need do some changes)
}
else
{//right side
this->setGeometry(size.width()/2, 0, size.width()/2, size.height());
}
With QApplication::desktop() it will work properly on screen with different resolutions.
In web I found something similar in winapi, but it didn't work properly:
HWND act = GetForegroundWindow();
PostMessage((HWND)act,WM_NCLBUTTONDBLCLK, HTTOP, 0);
The best way
Combine this approaches. For example:
HWND act = GetForegroundWindow();
bool left = false;
QSize size = QApplication::desktop()->size();
if(left)
{
this->move(0,0);
PostMessage((HWND)act,WM_NCLBUTTONDBLCLK, HTTOP, 0);
this->resize(size.width()/2,QApplication::desktop()->height());
}
else
{
this->move(size.width()/2,0);
PostMessage((HWND)act,WM_NCLBUTTONDBLCLK, HTTOP, 0);
this->resize(size.width()/2,QApplication::desktop()->height());
}
Why? Because move() regulate left and right sides, but PostMessage (winapi) set window's height properly on every screen (window will not locate lower then taskbar, as in your example)
EDIT
I changed code a little and now it is better. Yes, it is resizing again, but now it hasn't winapi code (PostMessage etc), so Photoshop doesn't catch it, there is one interesting method in Qt which called availableGeometry. It return normal height of screen which we need, with this method borderless windows perfectly simulates Aero Snap effects in different directions. It is works, maybe don't so good, but as I can see, there isn't API for Aero effects. Maybe this approach will be normal for yoo.
There is Aero Peek in Qt : http://qt-project.org/doc/qt-5/qtwinextras-overview.html , but it is can't solve this problem too.
Code:
bool left = true;
bool upper = true;
if(upper)
{
QRect rect = QApplication::desktop()->availableGeometry(-1);
this->setGeometry(rect);
}
else if(left)
{
QRect rect = QApplication::desktop()->availableGeometry(-1);
rect.setWidth(rect.width()/2);
this->setGeometry(rect);
}
else
{
QRect rect = QApplication::desktop()->availableGeometry(-1);
int half = rect.width()/2;
rect.setX(half);
rect.setWidth(half);
this->setGeometry(rect);
}
Try it with frameless window! You should choose one direction or let user choose it.

Dark transparent layer over a QMainWindow in Qt

I need to implement a "Loading..." window in my application but I do prefer to cover the whole QMainWindow with a dark transparent layer with a text above. Does anybody know how to do that? I am not sure how to overlap widgets/layouts in Qt. Any help will be appreciated.
This answer is in a series of my overlay-related answers: first, second, third.
The most trivial solution is to simply add a child transparent widget to QMainWindow. That widget must merely track the size of its parent window. It is important to properly handle changes of widget parentage, and the z-order with siblings. Below is a correct example of how to do it.
If you want to stack overlays, subsequent overlays should be the children of OverlayWidget, in the z-order. If they were to be siblings of the OverlayWidget, their stacking order is undefined.
This solution has the benefit of providing minimal coupling to other code. It doesn't require any knowledge from the widget you apply the overlay to. You can apply the overlay to a QMainWindow or any other widget, the widget can also be in a layout.
Reimplementing QMainWindow's paint event would not be considered the best design. It makes it tied to a particular class. If you really think that a QWidget instance is too much overhead, you better had measurements to show that being the case.
It is possible, of course, to make the overlay merely a QObject and to put the painting code into an event filter. That'd be an alternative solution. It's harder to do since you have to also properly deal with the parent widget's Qt::WA_StaticContents attribute, and with the widget potentially calling its scroll() method. Dealing with a separate widget is the simplest.
// https://github.com/KubaO/stackoverflown/tree/master/questions/overlay-widget-19362455
#include <QtGui>
#if QT_VERSION >= QT_VERSION_CHECK(5,0,0)
#include <QtWidgets>
#endif
class OverlayWidget : public QWidget
{
void newParent() {
if (!parent()) return;
parent()->installEventFilter(this);
raise();
}
public:
explicit OverlayWidget(QWidget * parent = {}) : QWidget{parent} {
setAttribute(Qt::WA_NoSystemBackground);
setAttribute(Qt::WA_TransparentForMouseEvents);
newParent();
}
protected:
//! Catches resize and child events from the parent widget
bool eventFilter(QObject * obj, QEvent * ev) override {
if (obj == parent()) {
if (ev->type() == QEvent::Resize)
resize(static_cast<QResizeEvent*>(ev)->size());
else if (ev->type() == QEvent::ChildAdded)
raise();
}
return QWidget::eventFilter(obj, ev);
}
//! Tracks parent widget changes
bool event(QEvent* ev) override {
if (ev->type() == QEvent::ParentAboutToChange) {
if (parent()) parent()->removeEventFilter(this);
}
else if (ev->type() == QEvent::ParentChange)
newParent();
return QWidget::event(ev);
}
};
class LoadingOverlay : public OverlayWidget
{
public:
LoadingOverlay(QWidget * parent = {}) : OverlayWidget{parent} {
setAttribute(Qt::WA_TranslucentBackground);
}
protected:
void paintEvent(QPaintEvent *) override {
QPainter p{this};
p.fillRect(rect(), {100, 100, 100, 128});
p.setPen({200, 200, 255});
p.setFont({"arial,helvetica", 48});
p.drawText(rect(), "Loading...", Qt::AlignHCenter | Qt::AlignVCenter);
}
};
int main(int argc, char * argv[])
{
QApplication a{argc, argv};
QMainWindow window;
QLabel central{"Hello"};
central.setAlignment(Qt::AlignHCenter | Qt::AlignTop);
central.setMinimumSize(400, 300);
LoadingOverlay overlay{&central};
QTimer::singleShot(5000, &overlay, SLOT(hide()));
window.setCentralWidget(&central);
window.show();
return a.exec();
}
I would suggest execute a modal, frameless dialog on top and add a graphics effect on the background widget.
This is IMHO a very flexible and short solution without touching the event system directly.
The performance might be bad - one could improve that by calling drawSource(), but I haven't a reliable solution here yet.
class DarkenEffect : public QGraphicsEffect
{
public:
void draw( QPainter* painter ) override
{
QPixmap pixmap;
QPoint offset;
if( sourceIsPixmap() ) // No point in drawing in device coordinates (pixmap will be scaled anyways)
pixmap = sourcePixmap( Qt::LogicalCoordinates, &offset );
else // Draw pixmap in device coordinates to avoid pixmap scaling;
{
pixmap = sourcePixmap( Qt::DeviceCoordinates, &offset );
painter->setWorldTransform( QTransform() );
}
painter->setBrush( QColor( 0, 0, 0, 255 ) ); // black bg
painter->drawRect( pixmap.rect() );
painter->setOpacity( 0.5 );
painter->drawPixmap( offset, pixmap );
}
};
// prepare overlay widget
overlayWidget->setWindowFlags( Qt::FramelessWindowHint | Qt::Dialog | Qt::WindowStaysOnTopHint );
// usage
parentWidget->setGraphicsEffect( new DarkenEffect );
overlayWidget->exec();
parentWidget->setGraphicsEffect( nullptr );
If you are talking about using a separate layout/widget over your 'Main Window', you could just make the "Loading..." window modal either through the UI editor or in the constructor for your UI.

Using QWidget::render() to draw widgets that draw other widgets (custom paintEvent)

I'm using Qt and C++ to create a custom widget I call ThumbnailView, which allows for left/right scrolling of thumbnail images:
class ThumbnailView : public QWidget {
public:
virtual void paintEvent(QPaintEvent *);
private:
QList<Thumbnail*> thumbList;
};
The ThumbnailView keeps an internal list of Thumbnail objects, which are also QWidget objects:
class Thumbnail : public QWidget
{
public:
virtual void paintEvent(QPaintEvent *);
};
I embed the ThumbnailView into a PreviewPane object I created:
class PreviewPane : public QWidget {
public:
virtual void paintEvent(QPaintEvent *);
private:
ThumbnailView thumbnailView;
};
When the main application loads, I create a dock widget and add the PreviewPane:
previewPaneDock = new QDockWidget(QString("PREVIEW"), this);
previewPane = new PreviewPane;
previewPaneDock->setWidget(previewPane);
this->addDockWidget(Qt::RightDockWidgetArea, previewPaneDock, Qt::Vertical);
So the idea is this: the dock widget has its widget set to previewPane, which in turn handles custom painting via paintEvent() and all mouse events (which I have omitted here). The previewPane's paintEvent does this:
void PreviewPane::paintEvent(QPaintEvent *)
{
QPainter painter(this);
...
thumbnailView.render(&painter);
}
The render() method is inherited from QWidget; this causes ThumbnailView::paintEvent() to be called:
void ThumbnailView::paintEvent(QPaintEvent *)
{
QPainter painter(this);
QList<Thumbnail*>::iterator itr;
int curX = 0;
for (itr = thumbList.begin(); itr != thumbList.end(); ++itr) {
curX += (*itr)->width();
if (curX < xScrollOffset) continue;
(*itr)->render(&painter, QPoint(curX - xScrollOffset - (*itr)->width(), 0));
if (curX - xScrollOffset >= this->width() ) break;
}
}
As you can see, the render() method is called again on each instance of Thumbnail.
Up to this point, there are no problems, and everything works as I expect it to. ThumbnailView allows the user to scroll left/right through a list of images (Thumbnail objects) by using a timer and kinetic scrolling using mouse flicks (or touch flicks).
I've embedded a ThumbnailView onto the PreviewPane, but the ThumbnailView isn't the only thing I want on the PreviewPane, and I want to be able to specify the origin at which the widget should start drawing--this is what I tried:
void PreviewPane::paintEvent(QPaintEvent *)
{
QPainter painter(this);
// specify a target offset of 10 pixels in y direction
thumbnailView.render(&painter, QPoint(0, 10));
}
which seems to have the same effect as this:
void PreviewPane::paintEvent(QPaintEvent *)
{
QPainter painter(this);
painter.translate(0, 10);
thumbnailView.render(&painter);
}
What I am expecting to happen is that the 10 pixel y-offset that I've specified to the paint transform will be passed to ThumbnailView::paintEvent(). Instead, it seems that this 10 pixel offset is propgated to each Thumbnail object, but instead of translating the Thumbnail widget, it crops it!
I've tried printing things like painter.combinedTransform().dy() and painter.worldTransform().dy() but they are always 0. Does anyone have some insight as to what happens when painter.translate() is called, or what the targetOffset parameter does in the function QWidget::render()?
In QtWidgets, all painting is clipped. It would seem that you need to apply an offset to both the painter and the clip rectangle.