scaling a QPixmap, devides Pixmap in different parts

Hello everyone,

I've written a application which resizes a QPixmap (which is added in a QGraphicsPixmapItem which is added to a QGraphicsScene). When I grab a corner, the QPixmap resizes (with a QTransform object). My Problem is, that the scaling isn't doing what I want. It's deviding the picture in 4 parts (see image).
Has someone an idea why this happens? Or is this a performance issue?

thank you

airglide


Attachment 8455

Quote:

---

Has someone an idea why this happens? Or is this a performance issue?

---

Sure, you are doing something odd in your code. I only we could see the code...

I've set some Dots to grab and set the derived QPixmap class as a parentitem, then I've installed a sceneEventFilter

Code:

---

bool PixmapItem::sceneEventFilter(QGraphicsItem *watched, QEvent *event)
{
    GrabDot *dot = dynamic_cast<GrabDot *> (watched);
    if(dot == NULL)
        return false;
 
    int expandX = 0;
    int expandY = 0;
 
    switch(event->type())
    {
 
    case QEvent::GraphicsSceneMousePress:
        {
 
        }break;
 
    case QEvent::GraphicsSceneMouseRelease:
        {
 
        }break;
 
 
    case QEvent::GraphicsSceneMouseMove:
        {
        QGraphicsSceneMouseEvent *e = dynamic_cast<QGraphicsSceneMouseEvent *>(event);
        switch(dot->getDotNumber())
        {
            case 0:
            {
                expandX = -1;
                expandY = -1;
                //this->setPos(this->cursor().pos());
            }break;
 
            case 1:
            {
                expandX = 0;
                expandY = -1;
            }break;
 
            case 2:
            {
                expandX = 1;
                expandY = -1;
            }break;
 
            case 3:
            {
                expandX = 1;
                expandY = 0;
            }break;
 
            case 4:
            {
                expandX = 1;
                expandY = 1;
            }break;
 
            case 5:
            {
                expandX = 0;
                expandY = 1;
            }break;
 
            case 6:
            {
                expandX = -1;
                expandY = 1;
            }break;
 
            case 7:
            {
                expandX = -1;
                expandY = 0;
            }break;
        }
 
        int width = 0;
        int height = 0;
 
        //if side is dragged, save the new Bounding rect
        bool sideDragged = false;
 
        switch(dot->getDotNumber())
        {
        case 0:
        case 2:
        case 4:
        case 6:
        {
            //if a corner is grabbed, the proportion has to stay the same
            //only the change of the x value is interesting
            width = boundingRect().width() + (expandX * e->pos().x());
            height = width * propRect.height()/propRect.width();
            sideDragged = false;
        }break;
 
        default:
            width = boundingRect().width() +(expandX * e->pos().x());
            height = boundingRect().height() +(expandY * e->pos().y());
            sideDragged = true;
            break;
 
        }
 
        QTransform trans = QTransform();
        trans.scale(width/boundingRect().width(), height/boundingRect().height());
        prepareGeometryChange();
        setPixmap(pixmap().transformed(trans));
 
        if(sideDragged)
        {
            saveBoundingRect();
        }
        switch(dot->getDotNumber())
        {
 
            case 0:
            {
                setPos(pos().x() + e->pos().x(), pos().y() + e->pos().y());
            }break;
 
            case 1:
            {
                setPos(pos().x(), pos().y() + e->pos().y());
            }break;
 
            case 2:
            {
                setPos(pos().x(), pos().y() + e->pos().y());
            }break;
 
            case 6:
            {
                setPos(pos().x() + e->pos().x(), pos().y());
            }break;
 
            case 7:
            {
                setPos(this->pos().x() + e->pos().x(), this->pos().y());
            }break;
        }
 
        setGrabDotPos();
        }break;
 
    default:
        {
            return false;
        }break;
 
    }

---

thank you for your help ;)

Why not save the original pixmap as an object ( in PyQt: self.originalPixmap = self.pixmap() ) and then when you scale the image set the scale from the original pixmap ( in PyQt: self.setPixmap(self.originalPixmap.scaled(width/boundingRect().width(), height/boundingRect().height()) )?

I wanted to implement that later, but now it works. I've added to my Pixmap class a QPixmap object called originalPixmap. I scale this image and display it.

great
thank you