Fix Windows Defender status retrieval under Java-25 or later

[HannesWell]

In recent Java versions Process.isAlive() seems to report the termination of a process a little bit later then before, relative to the time when the process output reader is closed, which leads to false errors about process time-out.

Fixes errors like:

!MESSAGE Failed to obtain 'WinDefend' service state
!STACK 0
java.io.IOException: Process timed-out and it was attempted to forcefully terminate it
	at org.eclipse.ui.internal.WindowsDefenderConfigurator.runProcess(WindowsDefenderConfigurator.java:395)
	at org.eclipse.ui.internal.WindowsDefenderConfigurator.runPowershell(WindowsDefenderConfigurator.java:369)
	at org.eclipse.ui.internal.WindowsDefenderConfigurator.isWindowsDefenderServiceRunning(WindowsDefenderConfigurator.java:302)
	at org.eclipse.ui.internal.WindowsDefenderConfigurator.runExclusionCheck(WindowsDefenderConfigurator.java:159)
	at org.eclipse.ui.internal.WindowsDefenderConfigurator.lambda$0(WindowsDefenderConfigurator.java:106)
	at org.eclipse.core.runtime.jobs.Job$2.run(Job.java:187)
	at org.eclipse.core.internal.jobs.Worker.run(Worker.java:63)

I cannot yet tell why, because I haven’t investigated the details yet.

CC @merks and @azoitl
Does this also help for

• Error while running the Windows Defender exclusion check #3591

[merks]

It's not so easy to test without a build, I see zero risk in this producing worse behavior given If the process has already terminated then this method returns immediately with the value {@code true}. So it should exactly deal with the case of it taking a while for isAlive to return false, and in the worst case taking 5 second longer to pop up a failure message...

[github-actions]

Test Results

   867 files  ±0     867 suites  ±0   55m 8s ⏱️ + 2m 23s
 7 988 tests ±0   7 745 ✅ ±0  243 💤 ±0  0 ❌ ±0 
20 466 runs  ±0  19 811 ✅ ±0  655 💤 ±0  0 ❌ ±0 

Results for commit ac4d0af. ± Comparison against base commit 195df72.

[merks]

@HannesWell

If we merge this today, I can more easily test tomorrow's I-build...

[HannesWell]

It's not so easy to test without a build,

The check is performed even in a secondary Eclipse, if you remove the inDevelopmentMode() test at

eclipse.platform.ui/bundles/org.eclipse.ui.workbench/eclipseui/org/eclipse/ui/internal/WindowsDefenderConfigurator.java

Line 114 in 195df72

if (Platform.OS.isWindows() && !Platform.inDevelopmentMode()) {

I see zero risk in this producing worse behavior given If the process has already terminated then this method returns immediately with the value {@code true}. So it should exactly deal with the case of it taking a while for isAlive to return false, and in the worst case taking 5 second longer to pop up a failure message...

Absolutely agree. And since it runs asynchronously in a job, its almost free.
So yes lets submit this now so that more I-build users can test this tomorrow.