I vote #3 (I actually thought we had already decided this a while back).
On Mar 13, 2009, at 7:20 PM, Todd Volkert wrote:
If a developer put an object in a pivot.collections.List more than once, then used that List as the data model for a TreeView, we'll get unpredictable behavior. This is because TreeView's internal data handler uses indexOf() to construct the path that it fires in its events (such as TreeViewNodeListener#nodeInserted()). Thus, even though listeners should be notified that N paths were affected, they'll only be notified of one path. Possibly solutions: 1) Have TreeView account for this by not just using indexOf() but actually looking for multiple occurrences of the data object. This would handle the case correctly but would be significantly less efficient. 2) Have TreeView not support data models with duplicate items by way of checking for that case and throwing an error. This might not be easily doable. 3) Have TreeView not support data models with duplicate items by way of documentation, and call such occurrences "developer error" Thoughts? -T
