Hi. I have experienced a problem when moving a landmark when zoom is not 1:1.
I have based the following fix on some other code present in v3d_core.cpp. Please check it to see if there are any inconsistencies.

thanks.

Index: v3d_core.cpp
===================================================================
--- v3d_core.cpp        (revision 1962)
+++ v3d_core.cpp        (working copy)
@@ -1748,7 +1748,10 @@
        else if (b_moveCurrentLandmark==true && ind_landmarkToBeChanged>=0 && QApplication::keyboardModifiers()==Qt::ShiftModifier)
        {
                QPoint cp = e->pos(); ///(disp_scale*m_scale); //100909
-
+
+               cp.setX(double((e->x()-curDisplayCenter.x()))/(disp_scale*m_scale)+pixmap.width()/2.0+0.5);
+               cp.setY(double((e->y()-curDisplayCenter.y()))/(disp_scale*m_scale)+pixmap.height()/2.0+0.5);
+
                int sx,sy,sz; //current selection location's x,y,z
 
                switch(Ptype)

furthermore, when a landmark is moved the comments and other information was not kept. This diff fixes that as well:

@@ -1770,9 +1773,12 @@
                                break;
                }
 
-               LocationSimple tmpLocation(sx,sy,sz);
-               tmpLocation.inputProperty = imgData->listLandmarks.at(ind_landmarkToBeChanged).inputProperty;
-               tmpLocation.radius = imgData->listLandmarks.at(ind_landmarkToBeChanged).radius;
+               // make a copy before replacing
+               LocationSimple tmpLocation = imgData->listLandmarks.at(ind_landmarkToBeChanged);
+               tmpLocation.x = sx;
+               tmpLocation.y = sy;
+               tmpLocation.z = sz;
+
                imgData->listLandmarks.replace(ind_landmarkToBeChanged, tmpLocation);
                imgData->setFocusFeatureViewText();

fixed in the version v2.700 (svn -r 1963). Thanks for the suggestions!

-Hanchuan



Originally posted by Carlos Becker:

You are welcome, thank you for such a nice software ;)

I have run into other bugs related to landmarks and I realized that, as before, they have to do with coordinate transformations.
In the attached diff I have written a simple function based on your code to compute a transformation between screen coordinates and image coordinates, so that it can be reused.
I also removed the old code that hard-coded that transformation, calling this new function instead.

Best regards.

excellent revision! I have committed the change your made to the version v2.701. Thanks for the contribution.

-hanchuan


Originally posted by Carlos Becker: