Fixes https://github.com/flutter/flutter/issues/154615
When scrolling back to the top, the position of the first visible item should have its scrollExtent subtracted. Otherwise, the accumulated position may be retained, which may result in the first visible item being hidden in some cases.
## Description
Relands https://github.com/flutter/flutter/pull/156968 wich was reverted in https://github.com/flutter/flutter/pull/157378
This PR makes `EditableText` aware of the lifecycle 'resumed' state to let the current selection unchanged when the application is resumed (on web and desktop, 'resumed' means the Flutter app window regained focus).
Before this PR, on web and desktop, the whole content of a `TextField` was selected whenever a `TextField` gained focus. This is the correct behavior when tabbing between fields but it is not when a field regains focus after the application is resumed
## Related Issue
Fixes [When switching to another browser tab or window and then going back, all text on TextField is selected automatically](https://github.com/flutter/flutter/issues/156078).
## Tests
Adds 1 test.
Reverts: flutter/flutter#156968
Initiated by: jacobsimionato
Reason for reverting: Google3 roll failing - see b/375019489
Original PR Author: bleroux
Reviewed By: {justinmc, gspencergoog}
This change reverts the following previous change:
## Description
This PR makes `EditableText` aware of the lifecycle 'resumed' state to let the current selection unchanged when the application is resumed (on web and desktop, 'resumed' means the Flutter app window regained focus).
Before this PR, on web and desktop, the whole content of a `TextField` was selected whenever a `TextField` gained focus. This is the correct behavior when tabbing between fields but it is not when a field regains focus after the application is resumed
## Related Issue
Fixes [When switching to another browser tab or window and then going back, all text on TextField is selected automatically](https://github.com/flutter/flutter/issues/156078).
## Tests
Adds 1 test.
Hit testing behavior is currently hardcoded to `opaque` in `HtmlElementView` which causes the platform view to swallow pointer events when it's a descendant of a `GestureDetector` (see this [example](https://dartpad.dev/?id=5348fbf82be5eeb7d995f953437f0ce6)).
In order to fix this, we need to make `hitTestBehavior` configurable in `HtmlElementView`. This way users can decide to make their platform views `transparent` for hit testing purposes.
**_Note_**: this specific case isn't fixable by `PointerInterceptor` because the framework is already receiving the pointer events, so there's nothing that `PointerInterceptor` can do. This issue is all happening on the framework side since it treats the rectangle occupied by the platform view as an opaque rectangle for hit testing.
**_Workaround_**: in the meantime, users can work around this limitation by surrounding the `HtmlElementView` with an `IgnorePointer` widget.
## Description
This PR makes `EditableText` aware of the lifecycle 'resumed' state to let the current selection unchanged when the application is resumed (on web and desktop, 'resumed' means the Flutter app window regained focus).
Before this PR, on web and desktop, the whole content of a `TextField` was selected whenever a `TextField` gained focus. This is the correct behavior when tabbing between fields but it is not when a field regains focus after the application is resumed
## Related Issue
Fixes to [When switching to another browser tab or window and then going back, all text on TextField is selected automatically](https://github.com/flutter/flutter/issues/156078).
## Tests
To be done
Adds `onTapUpOutside` and `onTapUpInside` callbacks to `TapRegion`, and `onTapUpOutside` to `EditableText` text fields. This allows users to detect `PointerUpEvent`s, which in turn can be used to differentiate between a tap and pan / scroll gesture (for example by comparing the distance between the down and up events).
Fixes#140271 and potentially #138190
Set the `hasSelectedState` flag in all selectable widgets. This happens automatically because only selectable widgets have `Semantics.selected` set to a non-null value. When the value is propagated to `SemanticsConfiguration.isSelected`, it sets both `hasSelectedState` to true and `isSelected` as appropriate.
More progress towards https://github.com/flutter/flutter/issues/66673
This is simply removing unnecessary parenthesis from various places. This change is because of a change to the unnecessary_parentesis lint that will trigger there. Here is the CL https://dart-review.googlesource.com/c/sdk/+/390161.
- https://github.com/dart-lang/linter/issues/4996
If anything else is needed please let me know.
I'd like to ask for this PR to wait a bit until the bots are run again on that CL so that I can be sure nothing else will trigger, I will come back here and update this whenever everything is complete. Thanks!
This test relies on the enumeration of all engine flags, and it will fail when https://github.com/flutter/engine/pull/55780 adds `SemanticsFlag.hasSelectedState`. So temporarily skip the test to allow the new flag to be added. I will later unskip the test and change it to check for the new flag as well.
A step towards https://github.com/flutter/flutter/issues/66673
This class looks like it _should_ give valid equality checks:
```dart
class _WidgetStateColorMapper extends _WidgetStateColorTransparent {
const _WidgetStateColorMapper(this.map);
final WidgetStateMap<Color> map;
_WidgetStateMapper<Color> get _mapper => _WidgetStateMapper<Color>(map);
@override
Color resolve(Set<WidgetState> states) => _mapper.resolve(states);
@override
bool operator ==(Object other) {
return other is _WidgetStateColorMapper && other.map == map;
}
@override
int get hashCode => map.hashCode;
}
```
But I made a mistake: I should have used `other._mapper == _mapper` and `_mapper.hashCode`.
<br>
It'd be pretty nice if no copy-pasting was needed in the first place:
```dart
class _WidgetStateColorMapper extends WidgetStateMapper<Color> implements WidgetStateColor {
const _WidgetStateColorMapper(super.map);
}
```
<br><br>
This PR makes `WidgetStateMapper` a public class, similar to `WidgetStatePropertyAll`. The new API, when used incorrectly, will `throw` with a detailed error message, rather than silently returning irrelevant values.
This PR adds `TapOutsideConfiguration` where you can define a custom default logic for `onTapOutside`.
```dart
TapOutsideConfiguration(
behavior: AlwaysUnfocusTapOutsideBehavior(),
// behavior: const NeverUnfocusTapOutsideBehavior(),
// behavior: const CustomTapOutsideBehavior(),
child: Column(
children: [
// This TextField will use onTapOutside from TapOutsideConfiguration
TextField(),
// Of course you can still define onTapOutside
TextField(onTapOutside: (event) => print('Tapped outside')),
],
),
)
```
Custom logic can be define like this:
```dart
class CustomTapOutsideBehavior extends TapOutsideBehavior {
const CustomTapOutsideBehavior();
@override
void defaultOnTapOutside(PointerDownEvent event, FocusNode focusNode) {
// any custom logic here
}
}
```
This PR implements also two simple `AlwaysUnfocusTapOutsideBehavior` and `NeverUnfocusTapOutsideBehavior`.
Resolves#150123
NavigatorPopHandler now includes the return value from Route. Recently some navigation infrastructure was updated to support passing through these return values, but NavigatorPopHandler was missed until now.
Following https://github.com/flutter/flutter/pull/151914, this PR is to normalize `ThemeData.cardTheme`; change the `CardTheme cardTheme` property to `CardThemeData cardTheme` in `ThemeData`. In `ThemeData()` and `ThemeData.copyWith()`, the `cardTheme` parameter type is changed to `Object?` to accept both `CardTheme` and `CardThemeData` so that we won't cause immediate breaking change and make sure rolling is smooth. Once all component themes are normalized, these `Object?` types should be changed to `xxxThemeData`.
There's no way to create a dart fix because we can't add a "@deprecated" label for `CardTheme` because `CardTheme` is a new InheritedWidget subclass now.
Addresses the "theme normalization" sub project within https://github.com/flutter/flutter/issues/91772
> ### Write Test, Find Bug
>
> When you fix a bug, first write a test that fails, then fix the bug and verify the test passes.
<br>
When `Theme.of(context)` is called in a `build()` method, the widget is rebuilt each frame during an `AnimatedTheme` transition.
I wanted to create a way for `RenderObject`s to be updated directly, so I wrote a test:
```dart
testWidgets('InheritedWidgets can trigger RenderObject updates', (WidgetTester tester) async {
// ...
});
```
â¦and it passed.
<br><br>
As it turns out, no change is needed at all!
This PR resolves#155852 by adding the "InheritedWidgets can trigger RenderObject updates" test, to ensure that this awesome capability doesn't break in the future.
This pull request aims to improve code readability, based on feedback gathered in a recent design doc.
<br>
There are two factors that hugely impact how easy it is to understand a piece of code: **verbosity** and **complexity**.
Reducing **verbosity** is important, because boilerplate makes a project more difficult to navigate. It also has a tendency to make one's eyes gloss over, and subtle typos/bugs become more likely to slip through.
Reducing **complexity** makes the code more accessible to more people. This is especially important for open-source projects like Flutter, where the code is read by those who make contributions, as well as others who read through source code as they debug their own projects.
<hr>
<br>
The following examples show how pattern-matching might affect these two factors:
<details> <summary><h3>Example 1 (GOOD)</h3> [click to expand]</summary>
```dart
if (ancestor case InheritedElement(:final InheritedTheme widget)) {
themes.add(widget);
}
```
Without using patterns, this might expand to
```dart
if (ancestor is InheritedElement) {
final InheritedWidget widget = ancestor.widget;
if (widget is InheritedTheme) {
themes.add(widget);
}
}
```
Had `ancestor` been a non-local variable, it would need to be "converted" as well:
```dart
final Element ancestor = this.ancestor;
if (ancestor is InheritedElement) {
final InheritedWidget inheritedWidget = ancestor.widget;
if (widget is InheritedTheme) {
themes.add(theme);
}
}
```
</details>
<details> <summary><h3>Example 2 (BAD) </h3> [click to expand]</summary>
```dart
if (widget case PreferredSizeWidget(preferredSize: Size(:final double height))) {
return height;
}
```
Assuming `widget` is a non-local variable, this would expand to:
```dart
final Widget widget = this.widget;
if (widget is PreferredSizeWidget) {
return widget.preferredSize.height;
}
```
<br>
</details>
In both of the examples above, an `if-case` statement simultaneously verifies that an object meets the specified criteria and performs a variable assignment accordingly.
But there are some differences: Example 2 uses a more deeply-nested pattern than Example 1 but makes fewer useful checks.
**Example 1:**
- checks that `ancestor` is an `InheritedElement`
- checks that the inherited element's `widget` is an `InheritedTheme`
**Example 2:**
- checks that `widget` is a `PreferredSizeWidget`
(every `PreferredSizeWidget` has a `size` field, and every `Size` has a `height` field)
<br>
<hr>
I feel hesitant to try presenting a set of cut-and-dry rules as to which scenarios should/shouldn't use pattern-matching, since there are an abundance of different types of patterns, and an abundance of different places where they might be used.
But hopefully the conversations we've had recently will help us converge toward a common intuition of how pattern-matching can best be utilized for improved readability.
<br><br>
- resolves https://github.com/flutter/flutter/issues/152313
- Design Doc: [flutter.dev/go/dart-patterns](https://flutter.dev/go/dart-patterns)
Fixes#33799
Allows for a route to inform the route below it in the navigation stack how to animate when the topmost route enters are leaves the stack.
It does this by making a `DelegatedTransition` available for the previous route to look up and use. If available, the route lower in the stack will wrap it's transition builders with that delegated transition and use it instead of it's default secondary transition.
This is what the sample code in this PR shows an app that is able to use both a Material zoom transition and a Cupertino slide transition in one app. It also includes a custom vertical transition. Every page animates off the screen in a way to match up with the incoming page's transition. When popped, the correct transitions play in reverse.
https://github.com/user-attachments/assets/1fc910fa-8cde-4e05-898e-daad8ff4a697
The below video shows this logic making a pseudo iOS styled sheet transition.
https://github.com/flutter/flutter/assets/58190796/207163d8-d87f-48b1-aad9-7e770d1d96c5
All existing page transitions in Flutter will be overwritten by the incoming route if a `delegatedTransition` is provided. This can be opted out of through `canTransitionTo` for a new route widget. Of Flutter's existing page transitions, this PR only adds a `DelegatedTransition` for the Zoom and Cupertino transitions. The other transitions possible in Material will get delegated transitions in a later PR.
This PR fixes the `ReorderableList` in `flutter/widgets.dart` not passing the item extent builder to the underlying SliverReorderableList. I double checked and the material equivalent is working as intended.
Fixes https://github.com/flutter/flutter/issues/155936
*If you had to change anything in the [flutter/tests] repo, include a link to the migration guide as per the [breaking change policy].*
Fixes#133637
This change updates the `_SelectableFragment.boundingBoxes` logic to consider the max line height. Previously we were using boxes that were tightly bound around each glyph, so you would have to click within the bounds of the glyph for double tap to select word to work. This is different than `SelectableText` which considers the max line height, as well as the native web behavior that also considers the max line height.
## Web
https://github.com/user-attachments/assets/4ce8c0ca-ec6f-4969-88b1-baa356be8278
## Flutter SelectableText
https://github.com/user-attachments/assets/54c22ad3-75d7-475b-856b-e9b2dbe09d54
## Flutter Text widget under SelectionArea
https://github.com/user-attachments/assets/27db0e2b-1d19-43cc-8ab3-16050e3a5bc7
After this change, Flutter's Text widget under a SelectionArea now matches the SelectableText and native web behavior.
This change also:
* Invalidates the cached bounding boxes when the paragraph layout changes.
* Updates `textOffsetToPosition` to consider `preferredLineHeight`. In cases when the text wraps, it was sometimes inaccurate.
This PR tweaks `wrapWithDefaultView` (used by `runApp`) to raise a StateError with a legible error message when the `platformDispatcher.implicitView` is missing (for example, when enabling multi-view embedding on the web), instead of crashing with an unexpected `nullCheck` a few lines below.
* Before:
<img width="619" alt="Screenshot 2024-09-25 at 7 33 47â¯PM" src="https://github.com/user-attachments/assets/4897dd3c-bdd0-4217-9f23-7eee9fab4999">
* After:
<img width="613" alt="Screenshot 2024-09-26 at 5 01 49â¯PM" src="https://github.com/user-attachments/assets/3febb91d-a8c3-41b6-bf34-c2c8743b637c">
## Issues
* Fixes https://github.com/flutter/flutter/issues/153198
## Tests
Added a test to ensure the assertion is thrown when the `implicitView` is missing. Had to hack a little because I couldn't find any clean way of overriding the `implicitView`. The problem is that the flutter_test bindings [use `runApp` internally](8925e1ffdf/packages/flutter_test/lib/src/binding.dart (L1020)) a couple of times, so I can only disable the implicitView inside the test body (and must re-enable it before returning). Not sure if it's the best way, but it seems to do the trick for this simple test case!
This change updates the behavior of `SelectableText`, to clear its selection when it loses focus and the application is currently running. This fixes the behavior where you may have multiple active highlights if you have `SelectableText` along with other "selectable" widgets such as `TextField`, or `Text` widgets under a `SelectionArea`.
If the application is in the background, for example when another window is focused, the selection should be retained so when a user returns to the application it is still there.
This change also updates the behavior of selection on macOS, single tap up, previously it was selecting the word edge closest to the tapped position, the correct behavior on native is to select the precise position. This was causing `onSelectionChanged` to be called twice, once for tap down (sets the precise tapped position, handled by logic in `TextSelectionGestureDetector`), and a second time for single tap up (moves the cursor to closest word edge, handled by logic in `_SelectableTextSelectionGestureDetectorBuilder`). This type of selection inconsistency is related to this issue https://github.com/flutter/flutter/issues/129726, I plan to look into this further in a separate PR.
Fixes#117573Fixes#103725
fixes#154515, #150048 and other similar issues where user non-draggable scrolls (mouse wheel, two-fingers) should behave same as draggable ones.
In this PR, scrollUpdateNotification.dragDetails check is removed and it has a supporting test which simulates a scroll which does not produce a drag.
Fix: Flicker when no update in index of dragged item
Resolves#150843
This PR ensures that even if we move dragged item anywhere in list and comeback to initial position, we smoothly adjust to that position.
## Description
This PR fixes the default selection on desktop when a text field is gaining focus.
Before this PR, when a text field is focused, the selection was collapsed at the end for all platforms except on Web where the entire content was selected.
After this PR, when a text field is focused, the entire content is selected on desktop and Web, and the selection is collapsed at the end on mobile platforms.
The implementation extends the work done in https://github.com/flutter/flutter/pull/119583 which implemented this feature for web.
## Related Issue
Fixes https://github.com/flutter/flutter/issues/150339.
## Tests
Updates 1 test.
Fixes 2 tests.
Fixes https://github.com/flutter/flutter/issues/154060
The error message doesn't make sense to me since one can call `setState` during the idle phase, and I'm not sure what this is guarding against without the right error message.
In the case of #154060 the layout builder was never laid out:
```
ââchild 1: _RenderLayoutBuilder#7c319 NEEDS-LAYOUT NEEDS-PAINT
â creator: LayoutBuilder â _BodyBuilder â MediaQuery â
â LayoutId-[<_ScaffoldSlot.body>] â CustomMultiChildLayout â
â _ActionsScope â Actions â AnimatedBuilder â DefaultTextStyle â
â AnimatedDefaultTextStyle â _InkFeatures-[GlobalKey#1f6eb ink
â renderer] â NotificationListener<LayoutChangedNotification> â â¯
â parentData: offset=Offset(0.0, 0.0); id=_ScaffoldSlot.body
â constraints: MISSING
â size: MISSING
```
So https://github.com/flutter/flutter/pull/154681 doesn't really fix#154060 since the layout callback cannot be run without a set of valid constraints.
Before the `BuildScope` change all `_inDirtyList` flags were unset after the `BuildOwner` finishes rebuilding the widget tree, so `LayoutBuilder._inDirtyLst` is always set to false after a frame even for layout builders that were never laid out.
With the `BuildScope` change, `LayoutBuilder` has its own `BuildScope` which is only flushed after LayoutBuilder gets its constraints.
Fixes https://github.com/flutter/flutter/issues/154156
Some iOS keyboard implementations change the selection in the text field if dismissed with active composing regions. The framework should not call `requestKeyboard` in such cases since that would bring up the keyboard again.
In general the `TextInput.show` call is not needed for selection only changes. For working around https://github.com/flutter/flutter/issues/68571 the show call is needed only if we restarted the input on Android (and we don't restart on selection-only changes any way).
This migrates the last failing test for https://github.com/flutter/engine/pull/54981. In order to effectively resolve that test I had to make `equalsIgnoringHashCodes` more usable by printing out the line that differs instead of just a huge blob of "expected" vs "actual.
## example
Here's the output after the change.
### test
```
test('equalsIgnoringHashCodes - wrong line', () {
expect(
'1\n2\n3\n4\n5\n6\n7\n8\n9\n10',
equalsIgnoringHashCodes('1\n2\n3\n4\n5\n6\na\n8\n9\n10'),
);
});
```
### output
```
Expected: normalized value matches
'1\n'
'2\n'
'3\n'
'4\n'
'5\n'
'6\n'
'a\n'
'8\n'
'9\n'
'10'
Actual: '1\n'
'2\n'
'3\n'
'4\n'
'5\n'
'6\n'
'7\n'
'8\n'
'9\n'
'10'
Which: Lines 7 differed, expected:
'a'
but got
'7'
```
Fixes#153889 an issue where nodes were being removed incorrectly when using `AnimationStyle.noAnimation `or the animation duration was zero seconds, which previously caused the application to freeze due to hidden state updates. By skipping the animation and updating active nodes immediately in these cases, we avoid these issues and ensure smooth and accurate management of node states.
Reverts: flutter/flutter#154048
Initiated by: nate-thegrate
Reason for reverting: using `AnimationStyle` instead would allow for complete animation duration and curve customization.
Original PR Author: nploi
Reviewed By: {piedcipher, nate-thegrate}
This change reverts the following previous change:
Currently we don't support custom transition duration for `DialogRoute`, `CupertinoDialogRoute` and show dialog methods , This PR will to support that.
Currently we don't support custom transition duration for `DialogRoute`, `CupertinoDialogRoute` and show dialog methods , This PR will to support that.
Added function and result identifiers into the RawAutocomplete onChangedField method, now the options value will not be updated if the data is from an optionsBuilder call older than the last call which updated the options
*List which issues are fixed by this PR. You must list at least one issue. An issue is not required if the PR fixes something trivial like a typo.*
Issue #150774 is fixed by this PR
*If you had to change anything in the [flutter/tests] repo, include a link to the migration guide as per the [breaking change policy].*
Fixes a bug where the app would exit on a back gesture when there were still routes in the navigation stack after returning to the app from a notification.
So, uhm, 2 years ago in https://github.com/flutter/flutter/pull/116429 we introduced this concept of a lookup boundary to hide certain InheritedWidgets in widget subtrees who belong to a different render tree than the InheritedWidget itself. This is for example needed for the Material widget: Buttons reach out to their Material ancestor to draw ink splashes on its associated render object. This only produces the desired effect if the button render object is a descendant of the Material render object (the two need to be in the same render tree). Overlay widgets have a similar problem.
Lookup boundaries were specifically designed for multi view support, where a sub view would be powered by a separate and independent render tree. Ergo, widgets in the sub view shouldn't see these InheritedWidgets if they are part of the parent view. After all, it would be strange if clicking on a button in a subview draws the ink splash effect into the parent view. Unfortunately, we (and by that I really mean I) forgot to put a LayoutBoundary into the relevant multi view widgets when those were introduced in https://github.com/flutter/flutter/pull/125003. This PR addresses that by wrapping the `ViewAnchor.view` child in a LookupBoundary so that its subtree (which bootstraps a separate render tree) cannot see these InheritedWidgets in the parent view.