/////////////////////////////////////////////////////////////////////////////
// Name:        dragimag.h
// Purpose:     documentation for wxDragImage class
// Author:      wxWidgets team
// RCS-ID:      $Id$
// Licence:     wxWindows license
/////////////////////////////////////////////////////////////////////////////

/**
    @class wxDragImage
    @wxheader{dragimag.h}

    This class is used when you wish to drag an object on the screen,
    and a simple cursor is not enough.

    On Windows, the Win32 API is used to achieve smooth dragging. On other
    platforms,
    wxGenericDragImage is used. Applications may also prefer to use
    wxGenericDragImage on Windows, too.

    @b wxPython note: wxPython uses wxGenericDragImage on all platforms, but
    uses the wxDragImage name.

    To use this class, when you wish to start dragging an image, create a
    wxDragImage
    object and store it somewhere you can access it as the drag progresses.
    Call BeginDrag to start, and EndDrag to stop the drag. To move the image,
    initially call Show and then Move. If you wish to update the screen contents
    during the drag (for example, highlight an item as in the dragimag sample),
    first call Hide,
    update the screen, call Move, and then call Show.

    You can drag within one window, or you can use full-screen dragging
    either across the whole screen, or just restricted to one area
    of the screen to save resources. If you want the user to drag between
    two windows, then you will need to use full-screen dragging.

    If you wish to draw the image yourself, use wxGenericDragImage and
    override wxDragImage::DoDrawImage and
    wxDragImage::GetImageRect.

    Please see @c samples/dragimag for an example.

    @library{wxcore}
    @category{FIXME}
*/
class wxDragImage : public wxObject
{
public:
    //@{
    /**
        )
        Constructs a drag image an optional cursor. This constructor is only available
        for
        wxGenericDragImage, and can be used when the application
        supplies DoDrawImage() and GetImageRect().
        
        @param image
            Icon or bitmap to be used as the drag image. The bitmap can
            have a mask.
        @param text
            Text used to construct a drag image.
        @param cursor
            Optional cursor to combine with the image.
        @param hotspot
            This parameter is deprecated.
        @param treeCtrl
            Tree control for constructing a tree drag image.
        @param listCtrl
            List control for constructing a list drag image.
        @param id
            Tree or list control item id.
    */
    wxDragImage();
    wxDragImage(const wxBitmap& image,
                const wxCursor& cursor = wxNullCursor);
    wxDragImage(const wxIcon& image,
                const wxCursor& cursor = wxNullCursor);
    wxDragImage(const wxString& text,
                const wxCursor& cursor = wxNullCursor);
    wxDragImage(const wxTreeCtrl& treeCtrl, wxTreeItemId& id);
    wxDragImage(const wxListCtrl& treeCtrl, long id);
    wxDragImage(const wxCursor& cursor = wxNullCursor);
    //@}

    //@{
    /**
        Start dragging the image, using the first window to capture the mouse and the
        second
        to specify the bounding area. This form is equivalent to using the first form,
        but more convenient than working out the bounding rectangle explicitly.
        You need to then call Show()
        and Move() to show the image on the screen.
        Call EndDrag() when the drag has finished.
        Note that this call automatically calls CaptureMouse.
        
        @param hotspot
            The location of the drag position relative to the upper-left corner
            of the image.
        @param window
            The window that captures the mouse, and within which the dragging
            is limited unless fullScreen is @true.
        @param boundingWindow
            In the second form of the function, specifies the
            area within which the drag occurs.
        @param fullScreen
            If @true, specifies that the drag will be visible over the full
            screen, or over as much of the screen as is specified by rect. Note that
        the mouse will
            still be captured in window.
        @param rect
            If non-@NULL, specifies the rectangle (in screen coordinates) that
            bounds the dragging operation. Specifying this can make the operation more
        efficient
            by cutting down on the area under consideration, and it can also make a
        visual difference
            since the drag is clipped to this area.
    */
    bool BeginDrag(const wxPoint& hotspot, wxWindow* window,
                   bool fullScreen = false,
                   wxRect* rect = NULL);
    bool BeginDrag(const wxPoint& hotspot, wxWindow* window,
                   wxWindow* boundingWindow);
    //@}

    /**
        Draws the image on the device context with top-left corner at the given
        position.
        This function is only available with wxGenericDragImage, to allow applications
        to
        draw their own image instead of using an actual bitmap. If you override this
        function,
        you must also override GetImageRect().
    */
    virtual bool DoDrawImage(wxDC& dc, const wxPoint& pos);

    /**
        Call this when the drag has finished.
        Note that this call automatically calls ReleaseMouse.
    */
    bool EndDrag();

    /**
        Returns the rectangle enclosing the image, assuming that the image is drawn
        with its
        top-left corner at the given point.
        This function is available in wxGenericDragImage only, and may be overridden
        (together with
        wxDragImage::DoDrawImage) to provide a virtual drawing capability.
    */
    virtual wxRect GetImageRect(const wxPoint& pos) const;

    /**
        Hides the image. You may wish to call this before updating the window
        contents (perhaps highlighting an item). Then call Move()
        and Show().
    */
    bool Hide();

    /**
        Call this to move the image to a new position. The image will only be shown if
        Show() has been called previously (for example
        at the start of the drag).
        @a pt is the position in client coordinates (relative to the window specified
        in BeginDrag).
        You can move the image either when the image is hidden or shown, but in general
        dragging
        will be smoother if you move the image when it is shown.
    */
    bool Move(const wxPoint& pt);

    /**
        Shows the image. Call this at least once when dragging.
    */
    bool Show();

    /**
        Override this if you wish to draw the window contents to the backing bitmap
        yourself. This can be desirable if you wish to avoid flicker by not having to
        redraw the updated window itself just before dragging, which can cause a
        flicker just
        as the drag starts. Instead, paint the drag image's backing bitmap to show the
        appropriate
        graphic @e minus the objects to be dragged, and leave the window itself to be
        updated
        by the drag image. This can provide eerily smooth, flicker-free drag behaviour.
        The default implementation copies the window contents to the backing bitmap. A
        new
        implementation will normally copy information from another source, such as from
        its
        own backing bitmap if it has one, or directly from internal data structures.
        This function is available in wxGenericDragImage only.
    */
    bool UpdateBackingFromWindow(wxDC& windowDC, wxMemoryDC& destDC,
                                 const wxRect& sourceRect,
                                 const wxRect& destRect) const;
};