Hello,
we (several users) are facing strange behaviour of SQuirreL SQL (5.1.0+) running on Java 25. Once SQuirreL SQL (and or OS?) falls asleep on Java 25 due its inactivity, it cannot be woken up and must be terminated (killed). We tried to investigate more and it looks like it is caused by Java version itself, not SQuirreL SQL. When we switch the same version, the same installation of SQuirreL SQL back to Java 17, the problem disappears.
Google and its AI said: In Windows, when you upgraded your Swing application from Java 17 to Java 25, Windows started going to sleep because modern Java versions have completely cleaned up legacy, CPU-intensive event-loop tasks. Under Java 17, the internal event queue infrastructure kept Windows continuously "awake" by trickling tiny bits of execution down to the OS kernel. In Java 25, the Swing Event Dispatch Thread (EDT) and basic timers sleep flawlessly using native, highly optimized Windows primitives. Windows sees 0% micro-activity, assumes your computer is idle, and goes to sleep.
The same behaviour we observed in our own Java GUI software and we solved it now by invisible activity, no JNI. I think, that any easiest and reliable solution would be satisfactory.
Thank You for Your attention, Stepan
Hello,
we (several users) are facing strange behaviour of SQuirreL SQL (5.1.0+) running on Java 25. Once SQuirreL SQL (and or OS?) falls asleep on Java 25 due its inactivity, it cannot be woken up and must be terminated (killed). We tried to investigate more and it looks like it is caused by Java version itself, not SQuirreL SQL. When we switch the same version, the same installation of SQuirreL SQL back to Java 17, the problem disappears.
Google and its AI said: In Windows, when you upgraded your Swing application from Java 17 to Java 25, Windows started going to sleep because modern Java versions have completely cleaned up legacy, CPU-intensive event-loop tasks. Under Java 17, the internal event queue infrastructure kept Windows continuously "awake" by trickling tiny bits of execution down to the OS kernel. In Java 25, the Swing Event Dispatch Thread (EDT) and basic timers sleep flawlessly using native, highly optimized Windows primitives. Windows sees 0% micro-activity, assumes your computer is idle, and goes to sleep.
The same behaviour we observed in our own Java GUI software and we solved it now by invisible activity, no JNI. I think, that any easiest and reliable solution would be satisfactory.
Thank You for Your attention, Stepan