Hi Eric,
just a brief follow-up on this.
As per your suggestion I declared the field in the controller this way:
FXML private TreeView<Path> treeView
And now everything is working again as expected without cast.
*>> There have been recent changes in the area of raw types (TreeView field
in this case). Instead of holding T as "?" or "? extends Object", it is
now resolved to the erasure type "Object"*
Good to know for the future.
Thank you,
Mirco
Il Mar 3 Dic 2024, 15:49 Milles, Eric (TR Technology) <
eric.mil...@thomsonreuters.com> ha scritto:
> There have been recent changes in the area of raw types (TreeView field in
> this case). Instead of holding T as "?" or "? extends Object", it is now
> resolved to the erasure type "Object". So, TreeCell<Path> is not
> compatible with TreeCell<Object> and an error is emitted.
>
> If you specified your field with a type argument, the return type checking
> should be okay:
>
> @FXML private TreeView<Path> treeView
>
>
> ------------------------------
> *From:* Mirco Colletta <mirch...@gmail.com>
> *Sent:* Monday, December 2, 2024 4:42 PM
> *To:* dev@groovy.apache.org <dev@groovy.apache.org>
> *Cc:* Milles, Eric (TR Technology) <eric.mil...@thomsonreuters.com>
> *Subject:* Re: [EXT] Static type checking error from groovy version 4.0.22
>
> Hi Eric,
> actually there isn't an explicit instantiation of the treeView in the code
> since it's defined inside an .fxml file:
>
> <TreeView fx:id="treeView" styleClass="folder-tree" maxWidth="Infinity"
> maxHeight="Infinity" VBox.vgrow="ALWAYS"/>
>
> that is eventually loaded inside the controller:
>
> FXMLLoader fxmlLoader = new
> FXMLLoader(getClass().getResource(getClass().getSimpleName() + ".fxml"))
> ...
> fxmlLoader.load()
>
> But in the controller declaration I didn't specify the type of the generics
>
> @FXML private TreeView treeView
>
> so maybe a recent modification made the groovy type checker more strict.
>
> Thanks
>
> Il Lun 2 Dic 2024, 21:40 Milles, Eric (TR Technology) via dev <
> dev@groovy.apache.org> ha scritto:
>
> The missing info from your description for me is how treeView is
> declared. When I ensure that treeView is an instance of TreeView<Path>,
> the checks work out fine.
>
>
> interface Callback<P,R> {
> R call(P param)
> }
> class TreeCell<T> {
> }
> class TreeView<T> {
> void setCellFactory(Callback<TreeView<T>,TreeCell<T>> value) {
> }
> }
> class Path { }
> class PathTreeCell extends TreeCell<Path> { }
>
> *@groovy.transform.TypeChecked* test() {
> def treeView = new TreeView<Path>() // not TreeView or TreeView<?>
> treeView.setCellFactory({ /*TreeView<Path>*/ p ->
> def treeCell = new PathTreeCell()
> return treeCell
> })
> }
>
>
> ------------------------------
> *From:* Milles, Eric (TR Technology) via dev <dev@groovy.apache.org>
> *Sent:* Monday, December 2, 2024 9:45 AM
> *To:* dev@groovy.apache.org <dev@groovy.apache.org>
> *Cc:* Milles, Eric (TR Technology) <eric.mil...@thomsonreuters.com>
> *Subject:* Re: [EXT] Static type checking error from groovy version 4.0.22
>
> Micro,
> Please create a new bug report at issues.apache.org
> <https://urldefense.com/v3/__http://issues.apache.org__;!!GFN0sa3rsbfR8OLyAw!ZOfYR2D6zT7F52tmV8qCubBdUFRDQNPfX56LZWLNMCdJMavUdhHk16eT0jgS-b3W4VkAxN5J0w3lU_GMAmvD7m8$>
> and
> I will have a look at what changed.
> ------------------------------
> *From:* Mirco Colletta <mirch...@gmail.com>
> *Sent:* Saturday, November 30, 2024 12:20 PM
> *To:* dev@groovy.apache.org <dev@groovy.apache.org>
> *Subject:* [EXT] Static type checking error from groovy version 4.0.22
>
> *External Email:* Use caution with links and attachments.
>
> Hello,
> recently I've updated the groovy version of my project (STC with
> compiler-configuration) from 4.0.21 to 4.0.24 and in the process I've
> encountered the following unexpected error:
> *[Static type checking] - Cannot return value of type ...PathTreeCell for
> closure expecting javafx.scene.control.TreeCell<java.lang.Object>*
> (after further investigation it seems introduced from v4.0.22)
>
> The error came up calling this method of the JavaFx *TreeView *(with a
> closure)
>
> https://openjfx.io/javadoc/21/javafx.controls/javafx/scene/control/TreeView.html#setCellFactory(javafx.util.Callback)
> <https://urldefense.com/v3/__https://openjfx.io/javadoc/21/javafx.controls/javafx/scene/control/TreeView.html*setCellFactory(javafx.util.Callback)__;Iw!!GFN0sa3rsbfR8OLyAw!bhkxAtKCcCoBP6_yq_UXSTJtpdyaaP3fx207yBuyP6IrCOCFSCZBuachWG4pzvoKw3JK31poWMrMii3T5yhIza0$>
> public final void setCellFactory(Callback<TreeView<T>,TreeCell<T>> value)
>
> My code:
> public class PathTreeCell extends TreeCell<Path> { … }
> ...
> treeView.setCellFactory({ p ->
> var treeCell = new PathTreeCell(this)
> makeDraggable(treeCell)
>
> *// This doesn't work anymore // return treeCell*
> *return (TreeCell<Path>)treeCell // OK*
> })
>
> This portion of code has always worked for every version of groovy STC
> I've used up until now.
> As above, the workaround is effortless, just an explicit cast to
> *TreeCell<Path>*
> Nevertheless I'm worried that it could be some sort of side effect of a
> regression.
>
> Is this an intentional change of behaviour or a bug?
>
> Cheers,
> Mirco
>
>
