Annotating types for function expression parameters is usually
unnecessary because the parameter types can almost always be
inferred from the context, thus making the practice redundant.
Declaring variables as final when possible is a good practice
because it helps avoid accidental reassignments and allows the
compiler to do optimizations.
* Upgrade upgradable packages
* Apply default null-safety migrations
* Remove hashnode as a dependency and localize its functionality to the package.
Maintenance was done long time ago hence no need to wait for the update
* Localize ui package to reduce maintenance burdens
* Replace universal html with dart:html
* Remove unnecessary checks
* Fix formatting
* Migrate app to null safety
* Enable methods to be nullable
* Fix non-nullable issue with node methods
* Cast as Node
* Use universal html
* Use universal html package to bring in the ImageElement class
* Remove unused imports
* Fix imports on the editor file
* Add key to quill editor
* Remove final from the GlobalKey
* Remove final on GlobalKey
* Remove custom util implementation in favor of quiver
* Fix issue with null on token attrivute
* Remove final hashcode that is replaced by quiver functionality
* Fix merge request
* Fix hit test position in text_line.dart
* Fix null safety errors on text_selection.dart
* Fix sound null safe errors in toolbar.dart
* Import null safe file picker
The code change allows to override the default behaviour of clicking on
images and thus solves #90.
Now the only thing left to do is to use a `RawGestureDetector` in the
`EmbedBuilder` to deactivate the "old" behaviour (by entering the
GestureArena and win). The `embedBuilder` could look like this:
```dart
Widget embedBuilder(BuildContext context, Embed node, bool readOnly) {
return FutureBuilder<File>(
future: File('image_path.png'),
builder: (context, snapshot) {
final gestures = {
ReadDependentGestureRecognizer: GestureRecognizerFactoryWithHandlers<
ReadDependentGestureRecognizer>(
() => ReadDependentGestureRecognizer(readOnly: readOnly),
(instance) => instance
..onTapUp = ((_) => print('Test'))
)
};
return RawGestureDetector(
gestures: gestures,
child: Image.file(snapshot.data),
);
}
);
}
class ReadDependentGestureRecognizer extends TapGestureRecognizer {
ReadDependentGestureRecognizer({@required this.readOnly});
final bool readOnly;
@override
bool isPointerAllowed(PointerDownEvent event) {
if (!readOnly) {
return false;
}
return super.isPointerAllowed(event);
}
}
```
**Explanation of the code:**
When a `PointerDownEvent` is received by the `GestureBinding` a hit
test is performed to determine which `HitTestTarget` nodes are affected.
For this the (I think) element tree is traversed downwards to find
out which elements has been hit. To find out which elements were
hit, the method `hitTestSelf` and `hitTestChildren` are used. Since
the `hitTestChildren` method of `TextLine` has not been implemented
yet, `false` was always returned, which is why the hit test
never arrived at the `embedBuilder`. With the code change, the hit is
passed on and can be processed correctly in the embed.
Additionally I removed the class `_TransparentTapGestureRecognizer`,
because this class always accepted the gesture, even if the recognizer
lost in the arena.
X Code
Michael Allen
Andy Trand
li3317
Cheryl Zhang
gtyhn
Till Friebe
Xun Gong
Jeremie Corpinot
Miller Adulu
rish07