This fixes propagation of needsCompositing from child widgets.
When needsCompositing is turned on by a child widget, it necessarily sets the needsCompositing bit of its parent widget, but RenderPointerListener was ignoring that piece of information and only turning on compositing if it thought it needed it for itself.
This corrects that, and adds a test for the condition, and updates a test that was affected by the change.
Fixes#32525 (again)
This fixes#32525, because it now marks the compositing bits as needing to be recalculated if the mouse tracker changes its idea of whether or not a mouse is attached.
This bug occurred because the test framework was leaking state from one test to the next (the state about whether a mouse pointer was active), and so even though there was a "passing" test when run in order with the other tests in the file, when the test was run individually (or first), it would have failed and caught the bug.
This adds an assert to make sure that after each test there are no simulated mouse pointers connected, and now calls removePointer in all of the tests where this was a problem.
This fixes a benchmark regression introduced in #32350. The performance is improved by just reverting the Layer.find<S> routines to use their old definitions, instead of defining them in terms of the findAll<S> lazy iterators.
Fixes#32387
This changes Listener to trigger enter/move/exit in all Listeners below the pointer, not just the leaf region (the first region hit). This is because we need to allow listeners to be nested so that, say, a widget that handles changing color on hover, but also is wrapped in a Tooltip (that handles hover) can trigger both actions, not just one.
To that end, I added a findAll to Layer, similar to the existing find method that was previously used. It returns an iterator over annotated layers which match the given data type.
Since the findAll is implemented as returning an Iterable (and is sync*), I re-implemented the find routines as just returning the first result from findAll, since that should be just as efficient, and would then prevent duplication in the implementation.
I recently added some code to keep hover events from being propagated when a mouse wasn't attached. While that works, there are times when it can fire callbacks during the building of other components, since they can now be called from detach/attach. This is not ideal, since it will assert then. This changes the code so that it won't update the annotations during attach/detach, but also won't push the annotation layer unless a mouse is connected, achieving the same result as before, but with better semantics.
The basic problem is that in the detach for RenderPointerListener, it would detach the annotation, which could cause onExit to be called on the annotation, since the widget was disappearing under the mouse, and thus needs to receive an onExit, but that onExit might be (and probably will be) calling setState, which marks the owning widget as needing to be built, sometimes when it already has been.
The fix creates a new _ListenerElement that overrides activate and deactivate in order to tell the render object ahead of the detach that it might be detached, and so the onExit gets called before the detach instead of during it.
In addition, I now avoid scheduling more than one check for mouse positions per frame.
This is a re-land of #31561, after fixing performance regressions.
Added change listening to the MouseTracker so that the Listener and tooltip can react to whether or not a mouse is connected at all. Added a change check to make sure Listener only repaints when something changed.
Fixes#22817
## Description
Make `_RenderCustomClip`'s `clipBehavior` non-final so we can update it during `updateRenderObject`. This will support `clipBehavior` changes in hot reload.
## Related Issues
Fixes#30863
## Tests
I added the following tests:
* ClipRect updates clipBehavior in updateRenderObject
* ClipRRect updates clipBehavior in updateRenderObject
* ClipOval updates clipBehavior in updateRenderObject
* ClipPath updates clipBehavior in updateRenderObject
* PhysicalModel updates clipBehavior in updateRenderObject
* PhysicalShape updates clipBehavior in updateRenderObject
This re-lands the Focus changes in #30040. Correctness changes in routes.dart, and removes the automatic requesting of focus on reparent when there is no current focus, which caused undesirable selections.
Addresses #11344, #1608, #13264, and #1678Fixes#30084Fixes#26704
Implements focus traversal for desktop platforms, including re-implementing the existing focus manager and focus tree.
This implements a Focus widget that can be put into a widget tree to allow input focus to be given to a particular part of a widget tree.
It incorporates with the existing FocusScope and FocusNode infrastructure, and has minimal breakage to the API, although FocusScope.reparentIfNeeded is removed, replaced by a call to FocusAttachment.reparent(), so this is a breaking change:
FocusScopeNodes must now be attached to the focus tree using FocusScopeNode.attach, which takes a context and an optional onKey callback, and returns a FocusAttachment that should be kept by the widget that hosts the FocusScopeNode. This is necessary because of the need to make sure that the focus tree reflects the widget hierarchy.
Callers that used to call FocusScope(context).reparentIfNeeded in their build method will call reparent on a FocusAttachment instead, which they will obtain by calling FocusScopeNode.attach in their initState method. Widgets that own FocusNodes will need to call dispose on the focus node in their dispose method.
Addresses #11344, #1608, #13264, and #1678Fixes#30084Fixes#26704
Adds a paint after detaching/attaching hover annotations to keep the annotation layers in sync with the annotations attached to the mouse detector.
Fixes#30744
RenderTable._computeColumnWidths() had a logic error that caused flex columns to be collapsed to their minimum widths in certain situations dependent on the layout width constraint and the number of flex columns.
