JDK-8350009 : [XWayland] SwingNodePlatformExitCrashTest hangs on Ubuntu 24.04
  • Type: Bug
  • Component: javafx
  • Sub-Component: other
  • Affected Version: jfx25
  • Priority: P3
  • Status: Resolved
  • Resolution: Fixed
  • OS: linux_ubuntu
  • CPU: generic
  • Submitted: 2025-02-13
  • Updated: 2025-04-15
  • Resolved: 2025-04-02
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
Other
jfx25 b11Fixed
Related Reports
Relates :  
JDK-8335468 - [XWayland] JavaFX hangs when calling java.awt.Robot.getPixelColor
Relates :  
JDK-8335469 - [XWayland] crash when an AWT ScreenCast session overlaps with an FX ScreenCast session
Relates :  
JDK-8339679 - Stabilize tests on Ubuntu 24.04
Relates :  
JDK-8340442 - Enable SwingNodePlatformExitCrashTest after JDK-8340849 is fixed
Relates :  
JDK-8353163 - [XWayland] Disable SwingNodePlatformExitCrashTest on wayland
Relates :  
JDK-8335470 - [XWayland] JavaFX tests that use AWT Robot fail on Wayland
Description
SwingNodePlatformExitCrashTest hangs when running system tests on Ubuntu 24.04.2 LTS VM. Two test windows appear when running the test, but outside of that nothing happens and they hang there forever until test times out or until you manually kill the test process (which causes gradle to not generate a test report). This does not happen on other platforms and seems to be specific to Ubuntu 24.04 (previous LTS works fine).

So far I can reliably reproduce it every single test run, no idea of what the cause might be yet.
Comments
Changeset: 056c7f52 Branch: master Author: Jayathirth D V <jdv@openjdk.org> Date: 2025-04-02 06:08:32 +0000 URL: https://git.openjdk.org/jfx/commit/056c7f52e544fa0f462a9c8f81e82df834bfd596
02-04-2025
A pull request was submitted for review. Branch: master URL: https://git.openjdk.org/jfx/pull/1753 Date: 2025-04-01 15:51:07 +0000
01-04-2025
I can reproduce the hang on my system, using either JDK 23 or JDK 24 with the latest JavaFX. I have attached a thread stack dump from the test run using JDK 24. UPDATE: There was something wrong with my system at the time I ran the test that hung with JDK 24: Wayland Screencast had gotten into a state where Robot capture wasn't working at all; after a reboot and granting screen cast access, the test hangs on JDK 23 and passes on JDK 24.
31-03-2025
It doesn't reproduce on all systems, even when running on JDK 23. Given that JDK-8335468, which is fixed in JDK 24, is a hang caused by concurrent access to Wayland screencast, it isn't surprising that it works in some cases and not in others. I recommend skipping this test on Wayland when running on a JDK < 24.
31-03-2025
It does not hang on my Ubuntu 24.04 running on Wayland. It's not a VM.
29-03-2025
I had changed the windowing system to X org in login screen and after every restart it is selecting X org(I was under the impression that it will move to default wayland). I changed it to wayland and now i can see this hang everytime. So the hang is specific to wayland windowing system.
27-03-2025
[~lkostyra] Initially when i was setting up my Ubuntu 24.04 i saw this issue of SwingNodePlatformExitCrashTest hang. At this time i had no 3D acceleration in Virtual Box enabled so tests were running using SW pipeline. But my initial setup was not stable, i have done multiple things like updating Ubuntu software, using appropriate hardware pipeline(using -Dprism.forceGPU=true), updating Virtual Box to latest version(which enabled me to increase video memory from 128MB to 256MB). With all these changes i see that SwingNodePlatformExitCrashTest doesn't hang/fail in my Ubuntu 24.04 VM. I have run the test multiple times and also ran full systemtests to check whether it fails in full run, but it doesn't. JavaFX is built using the same toolchain, so apart from above mentioned changes i have updated nothing else. On current state of my VM even if i set -Dprism.forceGPU=false(using SW pipeline) i see that this test passes. So change in behaviour is not related to -Dprism.forceGPU. I think failure observed by you might be related to some VM state and not actual FX bug.
27-03-2025