Re: Zoomer always zooms to absolute maximum zoom

I have figured out, that the new ZoomBase seems to be indeed unified 
intentionally with the current ScaleRect. The question remains then, 
however, why this mechanism is not effective when there hasn't been any 
zoom action, but becomes relevant after that initial zoom happens...

Cheers,
Matthias

Matthias Reich wrote:
> Uwe Rathmann wrote:
> 
>> If you tell us, what your intentions are chances are better to get a useful 
>> answer.
> 
> Ok. Here we go.
> 
> I'm trying to code a generic qwt_plot that has an inherent zoomer 
> implemented. After receiving a call to setZoomEnable (similar to Bode) 
> the plot can be zoomed as needed.
> 
> At the same time the plot is supposed to stick around while curves with 
> different bounding boxes are added and removed to/from it. On every 
> removeCurve and addCurve, the plot should adjust its zoomBase so that it 
> is the envelope of all boundingRects of the attached curves.
> 
> Currently, it seems the adding and removing of curves works ok. The 
> scales are adjusted and the (auto?)replot sends the full glory to the 
> screen. I use setZoomBase(boundingRect) and then zoom(0) to achieve this.
> 
> So far so good.
> 
> But this is before I actually use the zoomMode.
> 
> Enabling zoom mode and disabling it straight without doing a zoom is 
> fine. Everything still works as advertised above. But if I do a zoom and 
> then switch around curves by removing and reattaching curves with 
> smaller y-bounds, the zoom base somehow resists being reset to a smaller 
> rectangle than it had at some point before. The x-axis likely would do 
> the same, but I have identical x-axis lengths, so the issue doesn't surface.
> 
> That's the problem:
> -------------------
> Setting zoomBase with a smaller Rectangle than was set before fails for 
> the y-direction, if the zoomStack was used before. All my x-axis are 
> identical, so I can't relly tell whether x-axis does the same.
> 
> Further observations:
> Before I actually zoom, the zoomBase() reports (using cerr<<) integer 
> bounds, although the data does not have integer bounds (1.9504 instead 
> of 2). As soon as the first zoomRect has been added to the zoomStack, 
> the zoomBase reports double values identical to the data and doesn't 
> shrink anymore, unless I zoom and replot while zoomed.
> 
> Esentially, it looks like the setZoomBase() doesn't allow smaller 
> y-values than are present in its current zoomRect. Is this intentional ? 
> If so, how can I elgantly get around it (I don't want to have to delete 
> the zoomer and allocate a new one, for instance).
> 
> I can supply code, if that helps to sort out the problem. A nudge in the 
> right direction could already help, though.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008. 
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/