* fix bug when tap ListTitle when tristate enable #63846
* fix bug when tap ListTitle when tristate enable #63846
* code style
* improve the unit test case
* bow to convention
On a local benchmark, removing compute is faster than compute in both debug and release mode on a Pixel 4. On a MotoG4, a much larger text sample (800 Kb) takes 50 ms, so we cannot simple remove the limit. Increase to 50 Kb which should take at most a ms or two on an older device, and only microseconds on newerones.
The remaining uses of this pattern were all due to wanting to have
the child's slot when `Element.forgetChild()` was called. However,
when that method is called, the child's `slot` value is still valid
in the context of the parent, so the uses can just use `child.slot`.
This is the final round of cleanup from the fallout of #63269
* Redesigns the interface between MouseTracker and RendererBinding&RenderView.
* Simplifies the structure of RenderMouseRegion.
* Extracts the common utility code between mouse_tracker_test and mouse_tracker_cursor_test.
* (insert|move|remove)ChildRenderObject Deprecation: Step 1
This deprecates the following methods:
* RenderObjectElement.insertChildRenderObject
* RenderObjectElement.moveChildRenderObject
* RenderObjectElement.removeChildRenderObject
...and replaces them with the following methods:
* RenderObjectElement.insertRenderObjectChild
* RenderObjectElement.moveRenderObjectChild
* RenderObjectElement.removeRenderObjectChild
The reason for the deprecation is to provide the `oldSlot` argument to
the `moveRenderObjectChild` method (such an argument was missing from
the now-deprecated `moveChildRenderObject` method) and the `slot`
argument to the `removeRenderObjectChild` method (such an argument was
missing from the now-deprecated `removeChildRenderObject` method). While
no argument was added to `insertRenderObjectChild`, the name change (and
corresponding deprecation) was made to maintain naming parity with the
other two methods.
This initial step does not update or remove any of the `slotToChild`
patterns that exist in the framework. This work is being separated
into two commits in case something needs to be reverted to minimize
the scope of each commit.
See https://github.com/flutter/flutter/issues/63269 for more info
This standardizes the handling of nullability for the LogicalKeyboardKey.keyLabel and RawKeyEventData.keyLabel accessors so that they are non-nullable, but can be empty.
Before this change, the keyLabel could be either null or an empty string to indicate that there wasn't a label, which makes it harder to test for, since both need to be checked for. Since an empty string is sufficient, there is no need for it to be nullable.
Also, in raw_keyboard.dart, the web and Windows implementations wouldn't accept null values for parameters in the Map coming from the message, but tests were supplying null for some of them. This makes web and Windows creation of events match the other platforms, and makes the migration of tests to non-nullability easier.
