On Sat, 24 Apr 2021 20:55:52 GMT, mstr2 <github.com+43553916+ms...@openjdk.org> wrote:
>> This PR fixes the implementation of `ControlUtils.reducingChange`, which >> incorrectly computed adjacent removed indices, thus resulting in incorrect >> removal notifications. >> >> Since there were no unit tests for this method, I also added a bunch of >> tests. >> >> After applying this fix, I can no longer reproduce >> [JDK-8189354](https://bugs.openjdk.java.net/browse/JDK-8189354) and >> [JDK-8189228](https://bugs.openjdk.java.net/browse/JDK-8189228). > > mstr2 has updated the pull request incrementally with two additional commits > since the last revision: > > - Formatting > - Revert change to private ctor Hmm .. don't quite understand why there _is_ this method: it is (only?) used in tree/TableView treeModificationListeners to sync (aka: remove and fire the appropriate changes) selected indices. In that particular case, removed is a sublist of selectedIndices (all selections in the subtree), that is they are already adjacent (in coordinates of selectedIndices). A single nextRemove should be okay, f.i. in TreeTableView int firstRemoved = selectedIndices.indexOf(removed.get(0)); selectedIndices._nextRemove(firstRemoved, removed); // orig: not needed, all elements in removed _are_ adjacent (in coordinates of selectedIndices) // ControlUtils.reducingChange(selectedIndices, removed); ------------- PR: https://git.openjdk.java.net/jfx/pull/480
