IFileTest#testIssue2290() fails sporadically in GHA runs

[HeikoKlare]

Test was recently introduced by:

• Do not throw NPE from IFile.getContents(false) #2290 #2291

And improved in response to recent failures on Windows as reported here:

• IFileTest#testIssue2290() fails sporadically on Windows #2323

Now the test sporadically fails in GHA runs on every OS:

• This run on Windows: https://github.com/eclipse-platform/eclipse.platform/pull/2337/checks?check_run_id=59148499012
• This run on MacOS and Linux: https://github.com/eclipse-platform/eclipse.platform/pull/2336/checks?check_run_id=59148370836

Status CANCEL: unknown code=1
java.lang.AssertionError: Status CANCEL: unknown code=1 
	at org.eclipse.core.tests.resources.regression.IFileTest.testIssue2290(IFileTest.java:186)
	at java.base/java.lang.reflect.Method.invoke(Method.java:580)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1596)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1596)
Caused by: org.eclipse.core.runtime.CoreException: 
	... 4 more

It seems to be caused by createDelete.cancel() being called and result.isOK() being expected to yield true afterwards.

@basilevs fyi

[basilevs]

createDelete.cancel() being called and result.isOK() being expected to yield true afterwards

Yes. The assumption was that after job execution starts (ensured by CyclicBarrier), its result is determined by the return value. I'm not sure what's going on here.

Second link leads to a successfull build. Here is the broken one: https://github.com/eclipse-platform/eclipse.platform/actions/runs/20594683423/job/59146773828#step:5:19269

Switching to a boolean flag for cancellation.