'CurrentPoint' fails in GUI when function is called by 'ButtonDownFcn'

I have a GUI with an axis. I would like to be able to click and drag on the axis and have the start and end points detected in the units of the axis.

For a blank GUI with axes1 placed:

Code in OpeningFcn

% --- Executes just before currentpointtest is made visible.
function currentpointtest_OpeningFcn(hObject, eventdata, handles, varargin)
% Choose default command line output for currentpointtest
handles.output = hObject;

% Update handles structure
guidata(hObject, handles);

I = imread('cameraman.tif'); 
h = imshow(I, 'Parent', handles.axes1);
set(h, 'ButtonDownFcn', @getClicks)

Code in Callback:

function getClicks(~,~)
point1 = get(gca,'CurrentPoint')    % button down detected
finalRect = rbbox;                  % return figure units
point2 = get(gca,'CurrentPoint')    % button up detected

point1 is correct, but point2 returns the same value as point1. I need the values relative to the axes, so 'PointerLocation' doesn't seem like an option. Also as a side note, if the getClicks function is called from a push button everything works fine.

Does anyone know what is happening here and how I can get both point1 and point2?

Update(2012-03-07): I also tried the following:

function getClicks(~,~)
 point1 = get(gca,'CurrentPoint')    % button down detected
 set(gcf, 'WindowButtonUpFcn', @upCall)
 finalRect = rbbox;                  % return figure units
 set(gcf, 'WindowButtonUpFcn', '')
 point2 = get(gca,'CurrentPoint')    % button up detected

function upCall(~,~)
 upPoint = get(gca,'CurrentPoint')

This returns point1, then point2 (which is the same as point1) and finally upPoint (which is correct), in that order.

Answer 1

[Walter Roberson]

http://www.mathworks.com/help/techdoc/ref/figure_props.html

Read the section "CurrentPoint and Cursor Motion".

axes ButtonDownFcn is not one of the events that triggers CurrentPoint update.

You might be able to improve the situation by adding a drawnow() call after the rbbox .

[Jeff]

drawnow did not change the result, but thanks for the idea.

[KAE]

The documentation seems now to be here .

Answer 2

[Dirk Engel]

Maybe the "Interruptible" property of the ancestor figure is set to "off". This was responsible for a missing update of the CurrentPoint in my case.