QTransform no simple call to just set scale?

[JonB]

I am using a graphics view, and simply want to set the scale (only that, everything else untouched) to some absolute number, not a relative delta from where it is now.

QTransform::scale(qreal sx, qreal sy) scales from where the current transform is presently, e.g. each time you pass in 2.0 it doubles in size (1.0 -> 2.0 -> 4.0 -> 8.0 ...). I call that a "delta" scaling, it is a relative change in scale. So that will not work to set the scale to an absolute number like, say, 2.0.

I can read the current absolute scale from m11() & m22(), but there is no setm11/setm22(). So the only way I can see is:

t = graphicsView.transform()
t2 = QTransform(2.0, t.m12(), t.m13(), t.m21(), 2.0, t.m23(), t.m31(), t.m32(), t.m33())
graphicsView.setTransform(t2)

That seems like a lot of typing in the long line (I allowed variable to be named t in the example, it will be longer than that!) to just set two of the values! Am I missing something neater? (I am totally new to graphics/scaling/transforms, so don't expect any knowledge from me!)

[SGaist]

Hi,

Likely not ideal but based on the basic matrix operation, it looks like you should keep one transformation per type and then calculate the final one.

[JonB]

@SGaist
Thank you for replying. Sometimes your answers are cryptic to me :) I have read that doc section, and still don't know what you mean.

My goal is to reset the zoom level, in a standalone function which knows nothing else about the state the current graphics transform might or might not be in. I have a feeling you are implying I am responsible for tracking other, non-zoom-scale settings separately in code, which is not my responsibility, but I'm not sure if that's what you have in mind? :)

Anyway, if the answer does indeed seem to be "yes, there is no simple call to set just m11/m22 absolutely, you might as well do it with all the parameters to QTransform() as you showed", which is where I am at the moment, that's fine, at least I know!

[SGaist]

The example shows three QTransform objects, one per transformation type, that are combined before being applied.

What I was suggesting would be to keep this idea of one QTransform per modification type and combine them when needed so you can more freely modify their parameters.