RFR: 8359061: Update and ProblemList manual test java/awt/Cursor/CursorDragTest/ListDragCursor.java
Damon Nguyen
dnguyen at openjdk.org
Tue Jun 10 19:28:28 UTC 2025
On Tue, 10 Jun 2025 13:21:14 GMT, Alexey Ivanov <aivanov at openjdk.org> wrote:
>> This change is to restore the original intent of the test by updating the instructions to check that the type of Cursor is preserved when clicked and dragged. Now the test correctly has instructions to check that an I-beam cursor stays an I-beam until released over a List with its cursor being updated to a Hand cursor.
>>
>> There is a bug where this does not correctly update in macOS (found in [JDK-7177297](https://bugs.openjdk.org/browse/JDK-7177297)). So, this test needs to be problem-listed.
>>
>> I have confirmed that preserving the cursor image when dragging is native behavior across macOS, Windows, and Ubuntu. And I have checked that the test passes on both Windows and Ubuntu, while macOS fails and immediately updates the cursor as it leaves the TextArea.
>
> test/jdk/java/awt/Cursor/CursorDragTest/ListDragCursor.java line 48:
>
>> 46:
>> 47: If the mouse cursor starts as a Text Line Cursor (I-beam),
>> 48: and does not change until you reached the List and
>
> Is it really how native applications behave on macOS?
>
> At least on Windows, the cursor changes as soon as drag operation is started. The cursor shouldn't remain the I-beam when you drag from one component to another.
In Windows 11, I tried using the `file explorer's` search text field. It turns into an I-beam and stays one until I let go of the left mouse button. I also tried with `Cisco's VPN` settings button (cog wheel) that is a hand cursor and it behaves the same way. I also tried dragging from the `file explorer's` search field to `Cisco's` cog wheel and it stays an I-beam throughout.
In macOS, I used the search bar in `Finder` to get an I-beam and drag. It also stays an I-beam until released.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/25705#discussion_r2138625236
More information about the client-libs-dev
mailing list