Working on custom WPF control which is called "MultiSelectTreeView," which inherits from System.Windows.Controls.TreeView. Its purpose is to allow for multi-selecting and dragging and dropping.
The xaml for the view which contains this MultiSelectTreeView control binds an ObservableCollection that is exposed by the underlying view model.
A drag-and-drop operation might possibly involve many remove/add (or move) operations on the ObservableCollection, however I need to encapsulate all of the operations for a single drag-and-drop operation into a command object to support the undo/redoing of the drag-and-drop as a single atomic operation.
When I hook into the ViewModel.ObservableCollection's CollectionChanged event multiple events fire and from the perspective of the ViewModel, there is no way to know if any particular add/remove/move event will exist in isolation or whether it will be a part of a series of events, all related to a single user drag-and-drop.
I can imagine all sorts of wonky solutions, such as giving the MultiSelectTreeView control all sorts of in-depth knowledge of its underlying ViewModel's possible structure (to momentarily unhook the ObservableCollection's CollectionChanged event), but that doesn't feel right at all.
Perhaps I should create my own descendant of ObservableCollection, which supports a .MoveRange() method that only fires one event, or something along those lines.
I'm sure someone with more than just a few weeks of WPF experience could probably suggest a profoundly better solution than these.
