Reimplement of resizeEvent

[marcel]

Because you always create the label in showCB.
By creating a new one, the old one will remain in the old position.

You should create your label only once, and position it afterwards, as needed.
But my question is, what will you do when you have more labels?

[marcel]

So why not draw them on a pixmap and then just apply the pixmap to the label?

Yes, of course that would be better, but I thought maybe he wants to interact somehow with those images.

Anyway, there are many solutions, neither is that hard, and this thread seems to be one of those that gets 2-3 pages long, so I'm bailing out .

Regards

[Shawn]

I consider a label as a "canvas" for showing a image.
so when I want to show a image, I creat a label for it.
maybe it's less efficient but I thought it would work.

just as you said, by creating a new label, the old one will remain in the old position
then there should 2 image, one at the new position and the other at the previous position, is that right?
in my program, the image just appear at the previous position, that's why I said"the image is still at the previous position"

waiting for your kindly reply!

[marcel]

I consider a label as a "canvas" for showing a image.
so when I want to show a image, I creat a label for it.
maybe it's less efficient but I thought it would work.

just as you said, by creating a new label, the old one will remain in the old position
then there should 2 image, one at the new position and the other at the previous position, is that right?
in my program, the image just appear at the previous position, that's why I said"the image is still at the previous position"

waiting for your kindly reply!

But the resize events are posted very quickly. So, if you resize your window by 100 pixels, there is a good chance that the window receives 50 ( I'm just estimating here, could be less, could be more ) resize events.

Therefore you create 50 labels with the same image.
JPN suggested in an earlier post ( in this thread ) to draw the label delayed ( using a QTimer or whatever ). But this isn't to work in your case either because you're creating the labels in the resize event handler.

I consider a label as a "canvas" for showing a image.
so when I want to show a image, I creat a label for it.
maybe it's less efficient but I thought it would work.

But why do you create them in when a resize event is received?
I don't know exactly what are you trying to do by this, but shouldn't you load them as a response to some user action ( button press, action trigger )?
Or do you want this to be some kind of background for your main window?

Please elaborate on this a little.
I understand you're trying to show some images, but why like this?

[Shawn]

this is the outcome of my program:
every element is a small image.
the position for each element is given by a XML file, which I should parse at the beginning.
So, my program should parsing a XML file and generate this layout automatically.

what I want to implement now is to keep the layout at the central(horizontal) when someone resize the window.

[marcel]

In this case I approve Wysota's advice and I also strongly suggest you use a QGraphicsView.

Using the Graphics View Framework you will be able to achieve what I have seen in the screen shot much easier.

The Graphics View Framework is very well documented in Assistant and there are clarifying examples in the Qt Demos.

You should use this if you want your app to be fast and well behaved.
Once you understand the basics of this framework you will be able to do all those things very easily.

Of course, if you get stuck, you can ask here.

Regards

[Shawn]

and this one is the final goal

[marcel]

and this one is the final goal

Yes, you can do this with QGraphicsView.

The main drawbacks of the method you're currently using are:
- If the document gets bigger you will have to add a QScrollArea manually;
- You have to manage image positions manually at resize and scroll.
- Zooming would be very hard to implement

QGraphicsView offers all of the above by default, with simple to use API, so you should think about using it.
You could event interact with those components in your document( move them around, change connections, etc ).

Regards

[wysota]

QGraphicsView is certainly the way to go.

[Shawn]

Thank you very much for your suggestion!
I am gona to try it.

[wysota]

Why not use QGraphicsView?