>>>> You might just be able to use the renderer as the drag representation (not 
>>>> necessarily the one that is attached to the TreeView, but a new instance 
>>>> populated with the drag content).
>>> That would mean I have to call render(Object node, Path path, int rowIndex, 
>>> TreeView treeView, boolean expanded, boolean selected, 
>>> TreeView.NodeCheckState checkState, boolean highlighted, boolean disabled), 
>>> right? I think that will take some conciderable effort to set up.
>> There's nothing really to set up. You just need to call render(), passing 
>> the appropriate values (most of which you can either hard-code or get from 
>> the tree view itself).
> I understand. It seems like more work/code than the first alternative, but 
> nice to know :)

Well, of course I can't force you to use the renderer approach.  ;-)  But I'm 
curious to know why you think calling a single method is more work than 
creating an offscreen buffer and painting a clipped, translated view of a tree 
view into it.

The renderer is what actually paints the node content, so you're basically 
indirectly doing the same thing with your current approach, only with more 
steps.

G

Reply via email to