[Qt-interest] Getting Coordinates of QGraphicsItem in absolute terms of another

[Michael Jackson]

On 10/29/10 1:14 PM, in article
201010291814.21872.sean.harmer at maps-technology.com, "Sean Harmer" wrote:

> Hi,
> 
> On Friday 29 October 2010 16:13:56 Michael Jackson wrote:
>> I am displaying a QGraphicsView with a QGraphicsScene. Pretty basic stuff.
>> I load up an image (Tiff: 1286x955 pixels w x h) and then allow the user
>> to create a "selection" area using a combination of QRubberBand and a
>> QGraphicsPolygonItem derived item. This all shows up just fine. I can
>> select the "selection area" and move it around. I can resize it just fine.
>> What I would like to do is show in the user interface (a QTableView to be
>> precise) the upper left corner coordinates (in pixel coordinates) of that
>> "selection area". I have all the signals/slots hooked up. I have a custom
>> table model ready to go. The problem is that the values I get for things
>> like the boundingRect seem relative to the initial position that the
>> "selection area" was generated at? I have tried mapToScene, mapfromScene,
>> mapToParent, mapFromParent and a few others but nothing is really making
>> sense. A gently push in a direction would be greatly appreciated.
> 
> Would QGraphicsItem::mapFromItem( QGraphicsItem* item, const QPointF& ) do the
> trick? Possibly followed by a mapFromScene() as I am not quite sure exactly
> which coordinate system you wish you final answer to be in.
> 
> Cheers,
> 
> Sean
I think that is the same as mapFromParent(Qpoint) as I set the image as the
"selection area's" parent. I think this is what is giving me what I want:

QPoint p = pos().toPoint();
QRect b = boundingRect().toAlignedRect();
std::cout << "combined rect: " << b.x()+p.x() << ", " << b.y()+p.y() <<
std::endl;

I wanted the final values to be the pixel coordinates that are the
"selection area" covers in the original image.

Thanks for the heads up
Mike Jackson